[
{
"name": "GneissWeb.7B_ablation_model_on_350B_FineWeb.Edu.seed1",
"arch": "LlamaForCausalLM",
"mt": "llama",
"params": 6847467520,
"layers": 32,
"hidden": 4096,
"heads": 32,
"kv": 32,
"inter": 10928,
"vocab": 49152,
"ctx": 8192,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 189,
"mod": "2025-02-28",
"tags": [
"transformers",
"safetensors",
"llama",
"text-generation",
"en",
"arxiv:2502.14907",
"license:apache-2.0",
"text-generation-inference",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/GneissWeb.7B_ablation_model_on_350B_FineWeb.Edu.seed1",
"impl": {
"rope_theta": 10000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 8192,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": false,
"rope": {
"theta": 10000.0
},
"class": "LlamaForCausalLM",
"model_type": "llama"
}
},
{
"name": "GneissWeb.7B_ablation_model_on_350B_FineWeb.Edu.seed2",
"arch": "LlamaForCausalLM",
"mt": "llama",
"params": 6847467520,
"layers": 32,
"hidden": 4096,
"heads": 32,
"kv": 32,
"inter": 10928,
"vocab": 49152,
"ctx": 8192,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 206,
"mod": "2025-02-28",
"tags": [
"transformers",
"safetensors",
"llama",
"text-generation",
"en",
"arxiv:2502.14907",
"license:apache-2.0",
"text-generation-inference",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/GneissWeb.7B_ablation_model_on_350B_FineWeb.Edu.seed2",
"impl": {
"rope_theta": 10000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 8192,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": false,
"rope": {
"theta": 10000.0
},
"class": "LlamaForCausalLM",
"model_type": "llama"
}
},
{
"name": "GneissWeb.7B_ablation_model_on_350B_FineWeb.Edu.seed3",
"arch": "LlamaForCausalLM",
"mt": "llama",
"params": 6847467520,
"layers": 32,
"hidden": 4096,
"heads": 32,
"kv": 32,
"inter": 10928,
"vocab": 49152,
"ctx": 8192,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 196,
"mod": "2025-02-28",
"tags": [
"transformers",
"safetensors",
"llama",
"text-generation",
"en",
"arxiv:2502.14907",
"license:apache-2.0",
"text-generation-inference",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/GneissWeb.7B_ablation_model_on_350B_FineWeb.Edu.seed3",
"impl": {
"rope_theta": 10000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 8192,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": false,
"rope": {
"theta": 10000.0
},
"class": "LlamaForCausalLM",
"model_type": "llama"
}
},
{
"name": "GneissWeb.7B_ablation_model_on_350B_FineWeb.seed1",
"arch": "LlamaForCausalLM",
"mt": "llama",
"params": 6847467520,
"layers": 32,
"hidden": 4096,
"heads": 32,
"kv": 32,
"inter": 10928,
"vocab": 49152,
"ctx": 8192,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 197,
"mod": "2025-02-28",
"tags": [
"transformers",
"safetensors",
"llama",
"text-generation",
"en",
"arxiv:2502.14907",
"license:apache-2.0",
"text-generation-inference",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/GneissWeb.7B_ablation_model_on_350B_FineWeb.seed1",
"impl": {
"rope_theta": 10000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 8192,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": false,
"rope": {
"theta": 10000.0
},
"class": "LlamaForCausalLM",
"model_type": "llama"
}
},
{
"name": "GneissWeb.7B_ablation_model_on_350B_FineWeb.seed2",
"arch": "LlamaForCausalLM",
"mt": "llama",
"params": 6847467520,
"layers": 32,
"hidden": 4096,
"heads": 32,
"kv": 32,
"inter": 10928,
"vocab": 49152,
"ctx": 8192,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 190,
"mod": "2025-02-28",
"tags": [
"transformers",
"safetensors",
"llama",
"text-generation",
"en",
"arxiv:2502.14907",
"license:apache-2.0",
"text-generation-inference",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/GneissWeb.7B_ablation_model_on_350B_FineWeb.seed2",
"impl": {
"rope_theta": 10000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 8192,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": false,
"rope": {
"theta": 10000.0
},
"class": "LlamaForCausalLM",
"model_type": "llama"
}
},
{
"name": "GneissWeb.7B_ablation_model_on_350B_FineWeb.seed3",
"arch": "LlamaForCausalLM",
"mt": "llama",
"params": 6847467520,
"layers": 32,
"hidden": 4096,
"heads": 32,
"kv": 32,
"inter": 10928,
"vocab": 49152,
"ctx": 8192,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 198,
"mod": "2025-02-28",
"tags": [
"transformers",
"safetensors",
"llama",
"text-generation",
"en",
"arxiv:2502.14907",
"license:apache-2.0",
"text-generation-inference",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/GneissWeb.7B_ablation_model_on_350B_FineWeb.seed3",
"impl": {
"rope_theta": 10000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 8192,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": false,
"rope": {
"theta": 10000.0
},
"class": "LlamaForCausalLM",
"model_type": "llama"
}
},
{
"name": "GneissWeb.7B_ablation_model_on_350B_GneissWeb.seed1",
"arch": "LlamaForCausalLM",
"mt": "llama",
"params": 6847467520,
"layers": 32,
"hidden": 4096,
"heads": 32,
"kv": 32,
"inter": 10928,
"vocab": 49152,
"ctx": 8192,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 185,
"mod": "2025-02-28",
"tags": [
"transformers",
"safetensors",
"llama",
"text-generation",
"en",
"arxiv:2502.14907",
"license:apache-2.0",
"text-generation-inference",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/GneissWeb.7B_ablation_model_on_350B_GneissWeb.seed1",
"impl": {
"rope_theta": 10000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 8192,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": false,
"rope": {
"theta": 10000.0
},
"class": "LlamaForCausalLM",
"model_type": "llama"
}
},
{
"name": "GneissWeb.7B_ablation_model_on_350B_GneissWeb.seed2",
"arch": "LlamaForCausalLM",
"mt": "llama",
"params": 6847467520,
"layers": 32,
"hidden": 4096,
"heads": 32,
"kv": 32,
"inter": 10928,
"vocab": 49152,
"ctx": 8192,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 200,
"mod": "2025-02-28",
"tags": [
"transformers",
"safetensors",
"llama",
"text-generation",
"en",
"arxiv:2502.14907",
"license:apache-2.0",
"text-generation-inference",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/GneissWeb.7B_ablation_model_on_350B_GneissWeb.seed2",
"impl": {
"rope_theta": 10000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 8192,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": false,
"rope": {
"theta": 10000.0
},
"class": "LlamaForCausalLM",
"model_type": "llama"
}
},
{
"name": "GneissWeb.7B_ablation_model_on_350B_GneissWeb.seed3",
"arch": "LlamaForCausalLM",
"mt": "llama",
"params": 6847467520,
"layers": 32,
"hidden": 4096,
"heads": 32,
"kv": 32,
"inter": 10928,
"vocab": 49152,
"ctx": 8192,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 190,
"mod": "2025-02-28",
"tags": [
"transformers",
"safetensors",
"llama",
"text-generation",
"en",
"arxiv:2502.14907",
"license:apache-2.0",
"text-generation-inference",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/GneissWeb.7B_ablation_model_on_350B_GneissWeb.seed3",
"impl": {
"rope_theta": 10000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 8192,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": false,
"rope": {
"theta": 10000.0
},
"class": "LlamaForCausalLM",
"model_type": "llama"
}
},
{
"name": "GneissWeb.Edu_classifier",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": [
"apache-2.0"
],
"pipe": null,
"dl": 0,
"mod": "2025-02-21",
"tags": [
"en",
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/GneissWeb.Edu_classifier"
},
{
"name": "GneissWeb.Med_classifier",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": [
"apache-2.0"
],
"pipe": null,
"dl": 0,
"mod": "2025-02-21",
"tags": [
"en",
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/GneissWeb.Med_classifier"
},
{
"name": "GneissWeb.Quality_annotator",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": null,
"dl": 0,
"mod": "2025-02-21",
"tags": [
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/GneissWeb.Quality_annotator"
},
{
"name": "GneissWeb.Sci_classifier",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": [
"apache-2.0"
],
"pipe": null,
"dl": 0,
"mod": "2025-02-21",
"tags": [
"en",
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/GneissWeb.Sci_classifier"
},
{
"name": "GneissWeb.Tech_classifier",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": [
"apache-2.0"
],
"pipe": null,
"dl": 0,
"mod": "2025-02-21",
"tags": [
"en",
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/GneissWeb.Tech_classifier"
},
{
"name": "GneissWeb.bloom",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": [
"apache-2.0"
],
"pipe": null,
"dl": 0,
"mod": "2025-02-21",
"tags": [
"en",
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/GneissWeb.bloom"
},
{
"name": "granite-20b-code-base-8k",
"arch": "GPTBigCodeForCausalLM",
"mt": "gpt_bigcode",
"params": 20066993152,
"layers": 52,
"hidden": 6144,
"heads": 48,
"kv": 1,
"inter": 24576,
"vocab": 49152,
"ctx": 8192,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 1819,
"mod": "2026-04-25",
"tags": [
"transformers",
"safetensors",
"gpt_bigcode",
"text-generation",
"code",
"granite",
"dataset:codeparrot/github-code-clean",
"dataset:bigcode/starcoderdata",
"dataset:open-web-math/open-web-math",
"dataset:math-ai/StackMathQA",
"arxiv:2405.04324",
"license:apache-2.0",
"model-index",
"text-generation-inference",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-20b-code-base-8k",
"impl": {
"layer_norm_epsilon": 1e-05,
"n_positions": 8192,
"activation_function": "gelu",
"multi_query": true,
"pad_token_id": 0,
"class": "GPTBigCodeForCausalLM",
"model_type": "gpt_bigcode"
}
},
{
"name": "granite-20b-code-base-r1.1",
"arch": "GPTBigCodeForCausalLM",
"mt": "gpt_bigcode",
"params": 20066993152,
"layers": 52,
"hidden": 6144,
"heads": 48,
"kv": 1,
"inter": 24576,
"vocab": 49152,
"ctx": 8192,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 232,
"mod": "2024-07-16",
"tags": [
"transformers",
"safetensors",
"gpt_bigcode",
"text-generation",
"code",
"granite",
"dataset:codeparrot/github-code-clean",
"dataset:bigcode/starcoderdata",
"dataset:open-web-math/open-web-math",
"dataset:math-ai/StackMathQA",
"arxiv:2405.04324",
"license:apache-2.0",
"model-index",
"text-generation-inference",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-20b-code-base-r1.1",
"impl": {
"layer_norm_epsilon": 1e-05,
"n_positions": 8192,
"activation_function": "gelu",
"multi_query": true,
"pad_token_id": 0,
"class": "GPTBigCodeForCausalLM",
"model_type": "gpt_bigcode"
}
},
{
"name": "granite-20b-code-instruct-8k",
"arch": "GPTBigCodeForCausalLM",
"mt": "gpt_bigcode",
"params": 20066993152,
"layers": 52,
"hidden": 6144,
"heads": 48,
"kv": 1,
"inter": 24576,
"vocab": 49152,
"ctx": 8192,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 2775,
"mod": "2026-04-25",
"tags": [
"transformers",
"safetensors",
"gpt_bigcode",
"text-generation",
"code",
"granite",
"conversational",
"dataset:bigcode/commitpackft",
"dataset:TIGER-Lab/MathInstruct",
"dataset:meta-math/MetaMathQA",
"dataset:glaiveai/glaive-code-assistant-v3",
"dataset:glaive-function-calling-v2",
"dataset:bugdaryan/sql-create-context-instruction",
"dataset:garage-bAInd/Open-Platypus",
"dataset:nvidia/HelpSteer",
"arxiv:2405.04324",
"base_model:ibm-granite/granite-20b-code-base-8k",
"base_model:finetune:ibm-granite/granite-20b-code-base-8k",
"license:apache-2.0",
"model-index",
"text-generation-inference",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-20b-code-instruct-8k",
"impl": {
"layer_norm_epsilon": 1e-05,
"n_positions": 8192,
"activation_function": "gelu",
"multi_query": true,
"pad_token_id": 0,
"class": "GPTBigCodeForCausalLM",
"model_type": "gpt_bigcode"
}
},
{
"name": "granite-20b-code-instruct-accelerator",
"arch": "MLPSpeculatorPreTrainedModel",
"mt": "mlp_speculator",
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": 49152,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": null,
"dl": 125,
"mod": "2024-10-07",
"tags": [
"transformers",
"pytorch",
"mlp_speculator",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-20b-code-instruct-accelerator",
"impl": {
"emb_dim": 6144,
"inner_dim": 4096,
"n_predict": 4,
"n_candidates": 4,
"scale_input": false,
"tie_weights": false,
"top_k_tokens_per_head": [
4,
3,
2,
2
],
"class": "MLPSpeculatorPreTrainedModel",
"model_type": "mlp_speculator"
}
},
{
"name": "granite-20b-code-instruct-r1.1",
"arch": "GPTBigCodeForCausalLM",
"mt": "gpt_bigcode",
"params": 20066993152,
"layers": 52,
"hidden": 6144,
"heads": 48,
"kv": 1,
"inter": 24576,
"vocab": 49152,
"ctx": 8192,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 438,
"mod": "2024-07-16",
"tags": [
"transformers",
"safetensors",
"gpt_bigcode",
"text-generation",
"code",
"granite",
"conversational",
"dataset:bigcode/commitpackft",
"dataset:TIGER-Lab/MathInstruct",
"dataset:meta-math/MetaMathQA",
"dataset:glaiveai/glaive-code-assistant-v3",
"dataset:glaive-function-calling-v2",
"dataset:bugdaryan/sql-create-context-instruction",
"dataset:garage-bAInd/Open-Platypus",
"dataset:nvidia/HelpSteer",
"dataset:bigcode/self-oss-instruct-sc2-exec-filter-50k",
"arxiv:2405.04324",
"base_model:ibm-granite/granite-20b-code-instruct-r1.1",
"base_model:finetune:ibm-granite/granite-20b-code-instruct-r1.1",
"license:apache-2.0",
"model-index",
"text-generation-inference",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-20b-code-instruct-r1.1",
"impl": {
"layer_norm_epsilon": 1e-05,
"n_positions": 8192,
"activation_function": "gelu",
"multi_query": true,
"pad_token_id": 0,
"class": "GPTBigCodeForCausalLM",
"model_type": "gpt_bigcode"
}
},
{
"name": "granite-20b-functioncalling",
"arch": "GPTBigCodeForCausalLM",
"mt": "gpt_bigcode",
"params": 20066993152,
"layers": 52,
"hidden": 6144,
"heads": 48,
"kv": 1,
"inter": 24576,
"vocab": 49152,
"ctx": 8192,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 316,
"mod": "2024-07-10",
"tags": [
"transformers",
"safetensors",
"gpt_bigcode",
"text-generation",
"conversational",
"arxiv:2407.00121",
"license:apache-2.0",
"text-generation-inference",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-20b-functioncalling",
"impl": {
"layer_norm_epsilon": 1e-05,
"n_positions": 8192,
"activation_function": "gelu_pytorch_tanh",
"multi_query": true,
"pad_token_id": 0,
"class": "GPTBigCodeForCausalLM",
"model_type": "gpt_bigcode"
}
},
{
"name": "granite-3.0-1b-a400m-base",
"arch": "GraniteMoeForCausalLM",
"mt": "granitemoe",
"params": 1384956928,
"layers": 24,
"hidden": 1024,
"heads": 16,
"kv": 8,
"inter": 512,
"vocab": 49152,
"ctx": 4096,
"experts": 32,
"topk": 8,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 20407,
"mod": "2024-12-19",
"tags": [
"transformers",
"safetensors",
"granitemoe",
"text-generation",
"language",
"granite-3.0",
"arxiv:0000.00000",
"license:apache-2.0",
"model-index",
"region:us"
],
"id": "ibm-granite/granite-3.0-1b-a400m-base",
"impl": {
"rope_theta": 10000,
"rms_norm_eps": 1e-06,
"max_position_embeddings": 4096,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 6.0,
"num_local_experts": 32,
"num_experts_per_tok": 8,
"output_router_logits": false,
"pad_token_id": 0,
"rope": {
"theta": 10000
},
"class": "GraniteMoeForCausalLM",
"model_type": "granitemoe"
}
},
{
"name": "granite-3.0-1b-a400m-instruct",
"arch": "GraniteMoeForCausalLM",
"mt": "granitemoe",
"params": 1334628352,
"layers": 24,
"hidden": 1024,
"heads": 16,
"kv": 8,
"inter": 512,
"vocab": 49155,
"ctx": 4096,
"experts": 32,
"topk": 8,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 57828,
"mod": "2024-12-19",
"tags": [
"transformers",
"safetensors",
"granitemoe",
"text-generation",
"language",
"granite-3.0",
"conversational",
"arxiv:0000.00000",
"base_model:ibm-granite/granite-3.0-1b-a400m-base",
"base_model:finetune:ibm-granite/granite-3.0-1b-a400m-base",
"license:apache-2.0",
"model-index",
"region:us"
],
"id": "ibm-granite/granite-3.0-1b-a400m-instruct",
"impl": {
"rope_theta": 10000,
"rms_norm_eps": 1e-06,
"max_position_embeddings": 4096,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 6.0,
"num_local_experts": 32,
"num_experts_per_tok": 8,
"output_router_logits": false,
"pad_token_id": 0,
"rope": {
"theta": 10000
},
"class": "GraniteMoeForCausalLM",
"model_type": "granitemoe"
}
},
{
"name": "granite-3.0-2b-base",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 2533525504,
"layers": 40,
"hidden": 2048,
"heads": 32,
"kv": 8,
"inter": 8192,
"vocab": 49152,
"ctx": 4096,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 2516,
"mod": "2024-12-19",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"language",
"granite-3.0",
"arxiv:0000.00000",
"license:apache-2.0",
"model-index",
"eval-results",
"region:us"
],
"id": "ibm-granite/granite-3.0-2b-base",
"impl": {
"rope_theta": 10000,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 4096,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.1,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 8.0,
"pad_token_id": 0,
"rope": {
"theta": 10000
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-3.0-2b-instruct",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 2634201088,
"layers": 40,
"hidden": 2048,
"heads": 32,
"kv": 8,
"inter": 8192,
"vocab": 49155,
"ctx": 4096,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 5670,
"mod": "2024-12-19",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"language",
"granite-3.0",
"conversational",
"arxiv:0000.00000",
"base_model:ibm-granite/granite-3.0-2b-base",
"base_model:finetune:ibm-granite/granite-3.0-2b-base",
"license:apache-2.0",
"model-index",
"region:us"
],
"id": "ibm-granite/granite-3.0-2b-instruct",
"impl": {
"rope_theta": 10000,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 4096,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.1,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 8.0,
"pad_token_id": 0,
"rope": {
"theta": 10000
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-3.0-3b-a800m-base",
"arch": "GraniteMoeForCausalLM",
"mt": "granitemoe",
"params": 3374286336,
"layers": 32,
"hidden": 1536,
"heads": 24,
"kv": 8,
"inter": 512,
"vocab": 49152,
"ctx": 4096,
"experts": 40,
"topk": 8,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 703,
"mod": "2024-12-19",
"tags": [
"transformers",
"safetensors",
"granitemoe",
"text-generation",
"language",
"granite-3.0",
"arxiv:0000.00000",
"license:apache-2.0",
"model-index",
"region:us"
],
"id": "ibm-granite/granite-3.0-3b-a800m-base",
"impl": {
"rope_theta": 10000,
"rms_norm_eps": 1e-06,
"max_position_embeddings": 4096,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 6.0,
"num_local_experts": 40,
"num_experts_per_tok": 8,
"output_router_logits": false,
"pad_token_id": 0,
"rope": {
"theta": 10000
},
"class": "GraniteMoeForCausalLM",
"model_type": "granitemoe"
}
},
{
"name": "granite-3.0-3b-a800m-instruct",
"arch": "GraniteMoeForCausalLM",
"mt": "granitemoe",
"params": 3374295552,
"layers": 32,
"hidden": 1536,
"heads": 24,
"kv": 8,
"inter": 512,
"vocab": 49155,
"ctx": 4096,
"experts": 40,
"topk": 8,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 1285,
"mod": "2024-12-19",
"tags": [
"transformers",
"safetensors",
"granitemoe",
"text-generation",
"language",
"granite-3.0",
"conversational",
"arxiv:0000.00000",
"base_model:ibm-granite/granite-3.0-3b-a800m-base",
"base_model:finetune:ibm-granite/granite-3.0-3b-a800m-base",
"license:apache-2.0",
"model-index",
"region:us"
],
"id": "ibm-granite/granite-3.0-3b-a800m-instruct",
"impl": {
"rope_theta": 10000,
"rms_norm_eps": 1e-06,
"max_position_embeddings": 4096,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 6.0,
"num_local_experts": 40,
"num_experts_per_tok": 8,
"output_router_logits": false,
"pad_token_id": 0,
"rope": {
"theta": 10000
},
"class": "GraniteMoeForCausalLM",
"model_type": "granitemoe"
}
},
{
"name": "granite-3.0-8b-base",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 8170835968,
"layers": 40,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 12800,
"vocab": 49152,
"ctx": 4096,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 8734,
"mod": "2024-12-19",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"language",
"granite-3.0",
"arxiv:0000.00000",
"license:apache-2.0",
"model-index",
"eval-results",
"region:us"
],
"id": "ibm-granite/granite-3.0-8b-base",
"impl": {
"rope_theta": 10000,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 4096,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.1,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 16.0,
"pad_token_id": 0,
"rope": {
"theta": 10000
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-3.0-8b-instruct",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 8170848256,
"layers": 40,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 12800,
"vocab": 49155,
"ctx": 4096,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 41921,
"mod": "2024-12-19",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"language",
"granite-3.0",
"conversational",
"arxiv:0000.00000",
"base_model:ibm-granite/granite-3.0-8b-base",
"base_model:finetune:ibm-granite/granite-3.0-8b-base",
"license:apache-2.0",
"model-index",
"region:us"
],
"id": "ibm-granite/granite-3.0-8b-instruct",
"impl": {
"rope_theta": 10000,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 4096,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.1,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 16.0,
"pad_token_id": 0,
"rope": {
"theta": 10000
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-3.0-8b-instruct-accelerator",
"arch": "MLPSpeculatorPreTrainedModel",
"mt": "mlp_speculator",
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": 49155,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": null,
"dl": 92,
"mod": "2024-10-16",
"tags": [
"pytorch",
"mlp_speculator",
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/granite-3.0-8b-instruct-accelerator",
"impl": {
"emb_dim": 4096,
"inner_dim": 4096,
"n_predict": 4,
"n_candidates": 4,
"scale_input": true,
"tie_weights": true,
"top_k_tokens_per_head": [
4,
3,
2,
2
],
"class": "MLPSpeculatorPreTrainedModel",
"model_type": "mlp_speculator"
}
},
{
"name": "granite-3.0-8b-lora-intrinsics-v0.1",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 8170848256,
"layers": 40,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 12800,
"vocab": 49155,
"ctx": 4096,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 113,
"mod": "2024-12-24",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"conversational",
"en",
"arxiv:2403.08819",
"arxiv:2403.06009",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-3.0-8b-lora-intrinsics-v0.1",
"impl": {
"rope_theta": 10000,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 4096,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.1,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 16.0,
"pad_token_id": 0,
"rope": {
"theta": 10000
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-3.1-1b-a400m-base",
"arch": "GraniteMoeForCausalLM",
"mt": "granitemoe",
"params": 1334625280,
"layers": 24,
"hidden": 1024,
"heads": 16,
"kv": 8,
"inter": 512,
"vocab": 49152,
"ctx": 131072,
"experts": 32,
"topk": 8,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 4514,
"mod": "2025-01-31",
"tags": [
"transformers",
"safetensors",
"granitemoe",
"text-generation",
"language",
"granite-3.1",
"arxiv:0000.00000",
"license:apache-2.0",
"eval-results",
"region:us"
],
"id": "ibm-granite/granite-3.1-1b-a400m-base",
"impl": {
"rope_theta": 1500000.0,
"rms_norm_eps": 1e-06,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.1,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 6.0,
"num_local_experts": 32,
"num_experts_per_tok": 8,
"output_router_logits": false,
"pad_token_id": 0,
"rope": {
"theta": 1500000.0
},
"class": "GraniteMoeForCausalLM",
"model_type": "granitemoe"
}
},
{
"name": "granite-3.1-1b-a400m-instruct",
"arch": "GraniteMoeForCausalLM",
"mt": "granitemoe",
"params": 1334628352,
"layers": 24,
"hidden": 1024,
"heads": 16,
"kv": 8,
"inter": 512,
"vocab": 49155,
"ctx": 131072,
"experts": 32,
"topk": 8,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 8121,
"mod": "2026-07-07",
"tags": [
"transformers",
"safetensors",
"granitemoe",
"text-generation",
"language",
"granite-3.1",
"conversational",
"arxiv:0000.00000",
"base_model:ibm-granite/granite-3.1-1b-a400m-base",
"base_model:finetune:ibm-granite/granite-3.1-1b-a400m-base",
"license:apache-2.0",
"eval-results",
"region:us"
],
"id": "ibm-granite/granite-3.1-1b-a400m-instruct",
"impl": {
"rope_theta": 1500000.0,
"rms_norm_eps": 1e-06,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 6.0,
"num_local_experts": 32,
"num_experts_per_tok": 8,
"output_router_logits": false,
"pad_token_id": 0,
"rope": {
"theta": 1500000.0
},
"class": "GraniteMoeForCausalLM",
"model_type": "granitemoe"
}
},
{
"name": "granite-3.1-2b-base",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 2533525504,
"layers": 40,
"hidden": 2048,
"heads": 32,
"kv": 8,
"inter": 8192,
"vocab": 49152,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 795,
"mod": "2025-04-16",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"language",
"granite-3.1",
"license:apache-2.0",
"eval-results",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-3.1-2b-base",
"impl": {
"rope_theta": 5000000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.1,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 8.0,
"pad_token_id": 0,
"rope": {
"theta": 5000000.0
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-3.1-2b-instruct",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 2533531648,
"layers": 40,
"hidden": 2048,
"heads": 32,
"kv": 8,
"inter": 8192,
"vocab": 49155,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 13018,
"mod": "2025-04-16",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"language",
"granite-3.1",
"conversational",
"arxiv:0000.00000",
"base_model:ibm-granite/granite-3.1-2b-base",
"base_model:finetune:ibm-granite/granite-3.1-2b-base",
"license:apache-2.0",
"eval-results",
"region:us"
],
"id": "ibm-granite/granite-3.1-2b-instruct",
"impl": {
"rope_theta": 5000000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.1,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 8.0,
"pad_token_id": 0,
"rope": {
"theta": 5000000.0
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-3.1-3b-a800m-base",
"arch": "GraniteMoeForCausalLM",
"mt": "granitemoe",
"params": 3298788864,
"layers": 32,
"hidden": 1536,
"heads": 24,
"kv": 8,
"inter": 512,
"vocab": 49152,
"ctx": 131072,
"experts": 40,
"topk": 8,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 3901,
"mod": "2025-01-30",
"tags": [
"transformers",
"safetensors",
"granitemoe",
"text-generation",
"language",
"granite-3.1",
"arxiv:0000.00000",
"license:apache-2.0",
"eval-results",
"region:us"
],
"id": "ibm-granite/granite-3.1-3b-a800m-base",
"impl": {
"rope_theta": 10000000.0,
"rms_norm_eps": 1e-06,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.1,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 6.0,
"num_local_experts": 40,
"num_experts_per_tok": 8,
"output_router_logits": false,
"pad_token_id": 0,
"rope": {
"theta": 10000000.0
},
"class": "GraniteMoeForCausalLM",
"model_type": "granitemoe"
}
},
{
"name": "granite-3.1-3b-a800m-instruct",
"arch": "GraniteMoeForCausalLM",
"mt": "granitemoe",
"params": 3298793472,
"layers": 32,
"hidden": 1536,
"heads": 24,
"kv": 8,
"inter": 512,
"vocab": 49155,
"ctx": 131072,
"experts": 40,
"topk": 8,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 6449,
"mod": "2025-01-31",
"tags": [
"transformers",
"safetensors",
"granitemoe",
"text-generation",
"language",
"granite-3.1",
"conversational",
"arxiv:0000.00000",
"base_model:ibm-granite/granite-3.1-3b-a800m-base",
"base_model:finetune:ibm-granite/granite-3.1-3b-a800m-base",
"license:apache-2.0",
"eval-results",
"region:us"
],
"id": "ibm-granite/granite-3.1-3b-a800m-instruct",
"impl": {
"rope_theta": 10000000.0,
"rms_norm_eps": 1e-06,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 6.0,
"num_local_experts": 40,
"num_experts_per_tok": 8,
"output_router_logits": false,
"pad_token_id": 0,
"rope": {
"theta": 10000000.0
},
"class": "GraniteMoeForCausalLM",
"model_type": "granitemoe"
}
},
{
"name": "granite-3.1-8b-base",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 8170835968,
"layers": 40,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 12800,
"vocab": 49152,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 834,
"mod": "2025-04-16",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"language",
"granite-3.1",
"license:apache-2.0",
"eval-results",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-3.1-8b-base",
"impl": {
"rope_theta": 10000000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.1,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 16.0,
"pad_token_id": 0,
"rope": {
"theta": 10000000.0
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-3.1-8b-instruct",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 8170848256,
"layers": 40,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 12800,
"vocab": 49155,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 40604,
"mod": "2025-04-16",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"language",
"granite-3.1",
"conversational",
"arxiv:0000.00000",
"base_model:ibm-granite/granite-3.1-8b-base",
"base_model:finetune:ibm-granite/granite-3.1-8b-base",
"license:apache-2.0",
"eval-results",
"region:us"
],
"id": "ibm-granite/granite-3.1-8b-instruct",
"impl": {
"rope_theta": 10000000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.1,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 16.0,
"pad_token_id": 0,
"rope": {
"theta": 10000000.0
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-3.1-8b-lora-intrinsics-v0.1",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 8170848256,
"layers": 40,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 12800,
"vocab": 49155,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 110,
"mod": "2024-12-18",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"conversational",
"en",
"arxiv:2403.08819",
"arxiv:2403.06009",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-3.1-8b-lora-intrinsics-v0.1",
"impl": {
"rope_theta": 10000000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.1,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 16.0,
"pad_token_id": 0,
"rope": {
"theta": 10000000.0
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-3.2-2b-instruct",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 2533531648,
"layers": 40,
"hidden": 2048,
"heads": 32,
"kv": 8,
"inter": 8192,
"vocab": 49155,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 5601,
"mod": "2025-04-17",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"language",
"granite-3.2",
"conversational",
"arxiv:0000.00000",
"base_model:ibm-granite/granite-3.1-2b-instruct",
"base_model:finetune:ibm-granite/granite-3.1-2b-instruct",
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/granite-3.2-2b-instruct",
"impl": {
"rope_theta": 5000000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.1,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 8.0,
"pad_token_id": 0,
"rope": {
"theta": 5000000.0
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-3.2-8b-alora-jailbreak",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 0,
"mod": "2025-04-15",
"tags": [
"transformers",
"safetensors",
"text-generation",
"en",
"arxiv:2409.15398",
"arxiv:2408.01605",
"arxiv:2308.03825",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-3.2-8b-alora-jailbreak"
},
{
"name": "granite-3.2-8b-alora-rag-query-rewrite",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 0,
"mod": "2025-08-08",
"tags": [
"transformers",
"safetensors",
"text-generation",
"en",
"arxiv:2501.03468",
"arxiv:2309.15217",
"arxiv:2409.12558",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-3.2-8b-alora-rag-query-rewrite"
},
{
"name": "granite-3.2-8b-alora-requirement-check",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 0,
"mod": "2025-08-08",
"tags": [
"transformers",
"safetensors",
"text-generation",
"en",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-3.2-8b-alora-requirement-check"
},
{
"name": "granite-3.2-8b-instruct",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 8170848256,
"layers": 40,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 12800,
"vocab": 49155,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 32638,
"mod": "2025-04-17",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"language",
"granite-3.2",
"conversational",
"arxiv:0000.00000",
"base_model:ibm-granite/granite-3.1-8b-instruct",
"base_model:finetune:ibm-granite/granite-3.1-8b-instruct",
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/granite-3.2-8b-instruct",
"impl": {
"rope_theta": 10000000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.1,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 16.0,
"pad_token_id": 0,
"rope": {
"theta": 10000000.0
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-3.2-8b-instruct-preview",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 8170848256,
"layers": 40,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 12800,
"vocab": 49155,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 257,
"mod": "2025-02-26",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"language",
"granite-3.2",
"conversational",
"arxiv:0000.00000",
"base_model:ibm-granite/granite-3.1-8b-instruct",
"base_model:finetune:ibm-granite/granite-3.1-8b-instruct",
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/granite-3.2-8b-instruct-preview",
"impl": {
"rope_theta": 10000000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 16.0,
"pad_token_id": 0,
"rope": {
"theta": 10000000.0
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-3.2-8b-lora-jailbreak",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 0,
"mod": "2025-04-14",
"tags": [
"transformers",
"safetensors",
"text-generation",
"en",
"arxiv:2409.15398",
"arxiv:2408.01605",
"arxiv:2308.03825",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-3.2-8b-lora-jailbreak"
},
{
"name": "granite-3.2-8b-lora-rag-citation-generation",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 75,
"mod": "2025-04-17",
"tags": [
"peft",
"safetensors",
"text-generation",
"conversational",
"en",
"arxiv:2409.11500",
"arxiv:2409.02897",
"base_model:ibm-granite/granite-3.2-8b-instruct",
"base_model:adapter:ibm-granite/granite-3.2-8b-instruct",
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/granite-3.2-8b-lora-rag-citation-generation"
},
{
"name": "granite-3.2-8b-lora-uncertainty",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 0,
"mod": "2025-05-01",
"tags": [
"transformers",
"safetensors",
"text-generation",
"en",
"arxiv:2403.08819",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-3.2-8b-lora-uncertainty"
},
{
"name": "granite-3.3-2b-base",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 2533525504,
"layers": 40,
"hidden": 2048,
"heads": 32,
"kv": 8,
"inter": 8192,
"vocab": 49152,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 1388,
"mod": "2025-05-08",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"language",
"granite-3.3",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-3.3-2b-base",
"impl": {
"rope_theta": 10000000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 8.0,
"pad_token_id": 0,
"rope": {
"theta": 10000000.0
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-3.3-2b-instruct",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 2533539840,
"layers": 40,
"hidden": 2048,
"heads": 32,
"kv": 8,
"inter": 8192,
"vocab": 49159,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 54883,
"mod": "2025-05-08",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"language",
"granite-3.3",
"conversational",
"base_model:ibm-granite/granite-3.3-2b-base",
"base_model:finetune:ibm-granite/granite-3.3-2b-base",
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/granite-3.3-2b-instruct",
"impl": {
"rope_theta": 10000000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 8.0,
"pad_token_id": 0,
"rope": {
"theta": 10000000.0
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-3.3-8b-alora-requirement-check",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 0,
"mod": "2025-08-07",
"tags": [
"transformers",
"safetensors",
"text-generation",
"en",
"arxiv:2504.12397",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-3.3-8b-alora-requirement-check"
},
{
"name": "granite-3.3-8b-alora-uncertainty",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 0,
"mod": "2025-08-08",
"tags": [
"transformers",
"safetensors",
"text-generation",
"en",
"arxiv:2504.12397",
"arxiv:2403.08819",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-3.3-8b-alora-uncertainty"
},
{
"name": "granite-3.3-8b-base",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 8170835968,
"layers": 40,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 12800,
"vocab": 49152,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 3202,
"mod": "2025-05-08",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"language",
"granite-3.3",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-3.3-8b-base",
"impl": {
"rope_theta": 10000000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 16.0,
"pad_token_id": 0,
"rope": {
"theta": 10000000.0
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-3.3-8b-instruct",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 8170864640,
"layers": 40,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 12800,
"vocab": 49159,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 59892,
"mod": "2025-05-12",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"language",
"granite-3.3",
"conversational",
"arxiv:0000.00000",
"base_model:ibm-granite/granite-3.3-8b-base",
"base_model:finetune:ibm-granite/granite-3.3-8b-base",
"license:apache-2.0",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-3.3-8b-instruct",
"impl": {
"rope_theta": 10000000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 16.0,
"pad_token_id": 0,
"rope": {
"theta": 10000000.0
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-3.3-8b-lora-math-prm",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 0,
"mod": "2025-06-17",
"tags": [
"transformers",
"safetensors",
"reward model",
"text-generation",
"conversational",
"en",
"arxiv:2412.06559",
"arxiv:2501.03124",
"arxiv:2406.06592",
"base_model:ibm-granite/granite-3.3-8b-instruct",
"base_model:finetune:ibm-granite/granite-3.3-8b-instruct",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-3.3-8b-lora-math-prm"
},
{
"name": "granite-3.3-8b-math-prm-v2",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 8170864640,
"layers": 40,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 12800,
"vocab": 49159,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 596,
"mod": "2026-01-07",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"reward model",
"conversational",
"en",
"arxiv:2412.06559",
"arxiv:2501.03124",
"arxiv:2406.06592",
"base_model:ibm-granite/granite-3.3-8b-instruct",
"base_model:finetune:ibm-granite/granite-3.3-8b-instruct",
"license:apache-2.0",
"endpoints_compatible",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-3.3-8b-math-prm-v2",
"impl": {
"rope_theta": 10000000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 16.0,
"pad_token_id": 0,
"rope": {
"theta": 10000000.0
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-3.3-8b-rag-agent-lib",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": null,
"dl": 0,
"mod": "2025-06-28",
"tags": [
"transformers",
"safetensors",
"en",
"base_model:ibm-granite/granite-3.3-8b-instruct",
"base_model:finetune:ibm-granite/granite-3.3-8b-instruct",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-3.3-8b-rag-agent-lib"
},
{
"name": "granite-3.3-8b-security-lib",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": null,
"dl": 0,
"mod": "2025-10-28",
"tags": [
"transformers",
"safetensors",
"en",
"base_model:ibm-granite/granite-3.3-8b-instruct",
"base_model:finetune:ibm-granite/granite-3.3-8b-instruct",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-3.3-8b-security-lib"
},
{
"name": "granite-34b-code-base-8k",
"arch": "GPTBigCodeForCausalLM",
"mt": "gpt_bigcode",
"params": 33715603456,
"layers": 88,
"hidden": 6144,
"heads": 48,
"kv": 1,
"inter": 24576,
"vocab": 49152,
"ctx": 8192,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 503,
"mod": "2026-04-25",
"tags": [
"transformers",
"safetensors",
"gpt_bigcode",
"text-generation",
"code",
"granite",
"dataset:codeparrot/github-code-clean",
"dataset:bigcode/starcoderdata",
"dataset:open-web-math/open-web-math",
"dataset:math-ai/StackMathQA",
"arxiv:2405.04324",
"license:apache-2.0",
"model-index",
"text-generation-inference",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-34b-code-base-8k",
"impl": {
"layer_norm_epsilon": 1e-05,
"n_positions": 8192,
"activation_function": "gelu_pytorch_tanh",
"multi_query": true,
"pad_token_id": 0,
"class": "GPTBigCodeForCausalLM",
"model_type": "gpt_bigcode"
}
},
{
"name": "granite-34b-code-instruct-8k",
"arch": "GPTBigCodeForCausalLM",
"mt": "gpt_bigcode",
"params": 33715603456,
"layers": 88,
"hidden": 6144,
"heads": 48,
"kv": 1,
"inter": 24576,
"vocab": 49152,
"ctx": 8192,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 830,
"mod": "2026-04-25",
"tags": [
"transformers",
"safetensors",
"gpt_bigcode",
"text-generation",
"code",
"granite",
"conversational",
"dataset:bigcode/commitpackft",
"dataset:TIGER-Lab/MathInstruct",
"dataset:meta-math/MetaMathQA",
"dataset:glaiveai/glaive-code-assistant-v3",
"dataset:glaive-function-calling-v2",
"dataset:bugdaryan/sql-create-context-instruction",
"dataset:garage-bAInd/Open-Platypus",
"dataset:nvidia/HelpSteer",
"arxiv:2405.04324",
"base_model:ibm-granite/granite-34b-code-base-8k",
"base_model:finetune:ibm-granite/granite-34b-code-base-8k",
"license:apache-2.0",
"model-index",
"text-generation-inference",
"endpoints_compatible",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-34b-code-instruct-8k",
"impl": {
"layer_norm_epsilon": 1e-05,
"n_positions": 8192,
"activation_function": "gelu_pytorch_tanh",
"multi_query": true,
"pad_token_id": 0,
"class": "GPTBigCodeForCausalLM",
"model_type": "gpt_bigcode"
}
},
{
"name": "granite-34b-code-instruct-accelerator",
"arch": "MLPSpeculatorPreTrainedModel",
"mt": "mlp_speculator",
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": 49152,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": null,
"dl": 103,
"mod": "2024-07-24",
"tags": [
"transformers",
"pytorch",
"mlp_speculator",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-34b-code-instruct-accelerator",
"impl": {
"emb_dim": 6144,
"inner_dim": 6144,
"n_predict": 5,
"n_candidates": 5,
"scale_input": true,
"tie_weights": true,
"top_k_tokens_per_head": [
4,
3,
2,
2,
2
],
"class": "MLPSpeculatorPreTrainedModel",
"model_type": "mlp_speculator"
}
},
{
"name": "granite-3b-code-base-128k",
"arch": "LlamaForCausalLM",
"mt": "llama",
"params": 3482503680,
"layers": 32,
"hidden": 2560,
"heads": 32,
"kv": 32,
"inter": 10240,
"vocab": 49152,
"ctx": 128000,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 375,
"mod": "2026-04-25",
"tags": [
"transformers",
"safetensors",
"llama",
"text-generation",
"code",
"granite",
"dataset:codeparrot/github-code-clean",
"dataset:bigcode/starcoderdata",
"dataset:open-web-math/open-web-math",
"dataset:math-ai/StackMathQA",
"arxiv:2405.04324",
"license:apache-2.0",
"model-index",
"text-generation-inference",
"region:us"
],
"id": "ibm-granite/granite-3b-code-base-128k",
"impl": {
"rope_theta": 10000000,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 128000,
"hidden_act": "silu",
"attention_bias": true,
"attention_dropout": 0.1,
"mlp_bias": true,
"tie_word_embeddings": true,
"pad_token_id": 0,
"rope": {
"theta": 10000000
},
"class": "LlamaForCausalLM",
"model_type": "llama"
}
},
{
"name": "granite-3b-code-base-2k",
"arch": "LlamaForCausalLM",
"mt": "llama",
"params": 3482503680,
"layers": 32,
"hidden": 2560,
"heads": 32,
"kv": 32,
"inter": 10240,
"vocab": 49152,
"ctx": 2048,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 40814,
"mod": "2026-04-25",
"tags": [
"transformers",
"safetensors",
"llama",
"text-generation",
"code",
"granite",
"dataset:codeparrot/github-code-clean",
"dataset:bigcode/starcoderdata",
"dataset:open-web-math/open-web-math",
"dataset:math-ai/StackMathQA",
"arxiv:2405.04324",
"license:apache-2.0",
"model-index",
"text-generation-inference",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-3b-code-base-2k",
"impl": {
"rope_theta": 10000,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 2048,
"hidden_act": "silu",
"attention_bias": true,
"attention_dropout": 0.1,
"mlp_bias": true,
"tie_word_embeddings": true,
"pad_token_id": 0,
"rope": {
"theta": 10000
},
"class": "LlamaForCausalLM",
"model_type": "llama"
}
},
{
"name": "granite-3b-code-instruct-128k",
"arch": "LlamaForCausalLM",
"mt": "llama",
"params": 3482503680,
"layers": 32,
"hidden": 2560,
"heads": 32,
"kv": 32,
"inter": 10240,
"vocab": 49152,
"ctx": 128000,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 838,
"mod": "2026-04-25",
"tags": [
"transformers",
"safetensors",
"llama",
"text-generation",
"code",
"granite",
"conversational",
"dataset:bigcode/commitpackft",
"dataset:TIGER-Lab/MathInstruct",
"dataset:meta-math/MetaMathQA",
"dataset:glaiveai/glaive-code-assistant-v3",
"dataset:glaive-function-calling-v2",
"dataset:bugdaryan/sql-create-context-instruction",
"dataset:garage-bAInd/Open-Platypus",
"dataset:nvidia/HelpSteer",
"dataset:bigcode/self-oss-instruct-sc2-exec-filter-50k",
"arxiv:2405.04324",
"license:apache-2.0",
"model-index",
"text-generation-inference",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-3b-code-instruct-128k",
"impl": {
"rope_theta": 10000000,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 128000,
"hidden_act": "silu",
"attention_bias": true,
"attention_dropout": 0.1,
"mlp_bias": true,
"tie_word_embeddings": true,
"pad_token_id": 0,
"rope": {
"theta": 10000000
},
"class": "LlamaForCausalLM",
"model_type": "llama"
}
},
{
"name": "granite-3b-code-instruct-2k",
"arch": "LlamaForCausalLM",
"mt": "llama",
"params": 3482503680,
"layers": 32,
"hidden": 2560,
"heads": 32,
"kv": 32,
"inter": 10240,
"vocab": 49152,
"ctx": 2048,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 4033,
"mod": "2026-04-25",
"tags": [
"transformers",
"safetensors",
"llama",
"text-generation",
"code",
"granite",
"conversational",
"dataset:bigcode/commitpackft",
"dataset:TIGER-Lab/MathInstruct",
"dataset:meta-math/MetaMathQA",
"dataset:glaiveai/glaive-code-assistant-v3",
"dataset:glaive-function-calling-v2",
"dataset:bugdaryan/sql-create-context-instruction",
"dataset:garage-bAInd/Open-Platypus",
"dataset:nvidia/HelpSteer",
"arxiv:2405.04324",
"base_model:ibm-granite/granite-3b-code-base-2k",
"base_model:finetune:ibm-granite/granite-3b-code-base-2k",
"license:apache-2.0",
"model-index",
"text-generation-inference",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-3b-code-instruct-2k",
"impl": {
"rope_theta": 10000,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 2048,
"hidden_act": "silu",
"attention_bias": true,
"attention_dropout": 0.1,
"mlp_bias": true,
"tie_word_embeddings": true,
"pad_token_id": 0,
"rope": {
"theta": 10000
},
"class": "LlamaForCausalLM",
"model_type": "llama"
}
},
{
"name": "granite-3b-code-instruct-accelerator",
"arch": "MLPSpeculatorPreTrainedModel",
"mt": "mlp_speculator",
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": 49152,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": null,
"dl": 116,
"mod": "2024-07-10",
"tags": [
"transformers",
"pytorch",
"mlp_speculator",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-3b-code-instruct-accelerator",
"impl": {
"emb_dim": 2560,
"inner_dim": 4096,
"n_predict": 5,
"n_candidates": 5,
"scale_input": false,
"tie_weights": true,
"top_k_tokens_per_head": [
4,
3,
2,
2,
2
],
"class": "MLPSpeculatorPreTrainedModel",
"model_type": "mlp_speculator"
}
},
{
"name": "granite-4.0-1b",
"arch": "GraniteMoeHybridForCausalLM",
"mt": "granitemoehybrid",
"params": 1631750144,
"layers": 40,
"hidden": 2048,
"heads": 16,
"kv": 4,
"inter": 4096,
"vocab": 100352,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": [
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention"
],
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 21862,
"mod": "2025-10-23",
"tags": [
"transformers",
"safetensors",
"granitemoehybrid",
"text-generation",
"language",
"granite-4.0",
"conversational",
"arxiv:0000.00000",
"base_model:ibm-granite/granite-4.0-1b-base",
"base_model:finetune:ibm-granite/granite-4.0-1b-base",
"license:apache-2.0",
"endpoints_compatible",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-4.0-1b",
"impl": {
"rope_theta": 10000000,
"rms_norm_eps": 1e-05,
"position_embedding_type": "rope",
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": true,
"embedding_multiplier": 12,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 8,
"num_local_experts": 0,
"num_experts_per_tok": 0,
"shared_intermediate_size": 4096,
"output_router_logits": false,
"mamba_d_state": 256,
"mamba_d_conv": 4,
"mamba_expand": 2,
"mamba_n_heads": 128,
"mamba_n_groups": 1,
"mamba_chunk_size": 256,
"mamba_d_head": 32,
"mamba_conv_bias": true,
"mamba_proj_bias": false,
"pad_token_id": 100256,
"rope": {
"theta": 10000000
},
"class": "GraniteMoeHybridForCausalLM",
"model_type": "granitemoehybrid"
}
},
{
"name": "granite-4.0-1b-base",
"arch": "GraniteMoeHybridForCausalLM",
"mt": "granitemoehybrid",
"params": 1631750144,
"layers": 40,
"hidden": 2048,
"heads": 16,
"kv": 4,
"inter": 4096,
"vocab": 100352,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": [
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention"
],
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 37116,
"mod": "2026-06-12",
"tags": [
"transformers",
"safetensors",
"granitemoehybrid",
"text-generation",
"language",
"granite-4.0",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-4.0-1b-base",
"impl": {
"rope_theta": 10000000,
"rms_norm_eps": 1e-05,
"position_embedding_type": "rope",
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": true,
"embedding_multiplier": 12,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 8,
"num_local_experts": 0,
"num_experts_per_tok": 0,
"shared_intermediate_size": 4096,
"output_router_logits": false,
"mamba_d_state": 256,
"mamba_d_conv": 4,
"mamba_expand": 2,
"mamba_n_heads": 128,
"mamba_n_groups": 1,
"mamba_chunk_size": 256,
"mamba_d_head": 32,
"mamba_conv_bias": true,
"mamba_proj_bias": false,
"pad_token_id": 100256,
"rope": {
"theta": 10000000
},
"class": "GraniteMoeHybridForCausalLM",
"model_type": "granitemoehybrid"
}
},
{
"name": "granite-4.0-1b-speech",
"arch": "GraniteSpeechForConditionalGeneration",
"mt": "granite",
"params": 2313207132,
"layers": 40,
"hidden": 2048,
"heads": 16,
"kv": 4,
"inter": 4096,
"vocab": 100353,
"ctx": 4096,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "automatic-speech-recognition",
"dl": 29170,
"mod": "2026-04-02",
"tags": [
"transformers",
"safetensors",
"granite_speech",
"automatic-speech-recognition",
"multilingual",
"en",
"fr",
"de",
"es",
"pt",
"ja",
"arxiv:2505.08699",
"arxiv:2603.11243",
"base_model:ibm-granite/granite-4.0-1b-base",
"base_model:finetune:ibm-granite/granite-4.0-1b-base",
"license:apache-2.0",
"eval-results",
"endpoints_compatible",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-4.0-1b-speech",
"impl": {
"tie_word_embeddings": false,
"downsample_rate": 5,
"window_size": 15,
"class": "GraniteSpeechForConditionalGeneration",
"model_type": "granite_speech",
"text": {
"rope_theta": 10000.0,
"rope_parameters": {
"rope_theta": 10000,
"rope_type": "default"
},
"rms_norm_eps": 1e-05,
"max_position_embeddings": 4096,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"mlp_bias": false,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 8.0,
"pad_token_id": 100256,
"num_hidden_layers": 40,
"hidden_size": 2048,
"num_attention_heads": 16,
"num_key_value_heads": 4,
"intermediate_size": 4096,
"vocab_size": 100353,
"model_type": "granite",
"rope": {
"from": "rope_parameters",
"theta": 10000.0,
"rope_type": "default"
}
},
"encoder": {
"context_size": 200,
"num_layers": 16,
"hidden_dim": 1024,
"num_heads": 8,
"dim_head": 128,
"feedforward_mult": 4,
"conv_kernel_size": 15,
"output_dim": 348,
"model_type": "granite_speech_encoder"
},
"projector": {
"layer_norm_eps": 1e-12,
"position_embedding_type": "absolute",
"max_position_embeddings": 2048,
"hidden_act": "gelu",
"num_hidden_layers": 2,
"hidden_size": 1024,
"num_attention_heads": 16,
"intermediate_size": 4096,
"vocab_size": 30522,
"model_type": "blip_2_qformer"
}
}
},
{
"name": "granite-4.0-350m",
"arch": "GraniteMoeHybridForCausalLM",
"mt": "granitemoehybrid",
"params": 352379904,
"layers": 28,
"hidden": 1024,
"heads": 16,
"kv": 4,
"inter": 2048,
"vocab": 100352,
"ctx": 32768,
"experts": null,
"topk": null,
"window": null,
"ltypes": [
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention"
],
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 23003,
"mod": "2025-10-23",
"tags": [
"transformers",
"safetensors",
"granitemoehybrid",
"text-generation",
"language",
"granite-4.0",
"conversational",
"arxiv:0000.00000",
"base_model:ibm-granite/granite-4.0-350m-base",
"base_model:finetune:ibm-granite/granite-4.0-350m-base",
"license:apache-2.0",
"endpoints_compatible",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-4.0-350m",
"impl": {
"rope_theta": 10000000,
"rms_norm_eps": 1e-05,
"position_embedding_type": "rope",
"max_position_embeddings": 32768,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": true,
"embedding_multiplier": 12,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.263,
"logits_scaling": 4,
"num_local_experts": 0,
"num_experts_per_tok": 0,
"shared_intermediate_size": 2048,
"output_router_logits": false,
"mamba_d_state": 256,
"mamba_d_conv": 4,
"mamba_expand": 2,
"mamba_n_heads": 128,
"mamba_n_groups": 1,
"mamba_chunk_size": 256,
"mamba_d_head": 16,
"mamba_conv_bias": true,
"mamba_proj_bias": false,
"pad_token_id": 100256,
"rope": {
"theta": 10000000
},
"class": "GraniteMoeHybridForCausalLM",
"model_type": "granitemoehybrid"
}
},
{
"name": "granite-4.0-350m-base",
"arch": "GraniteMoeHybridForCausalLM",
"mt": "granitemoehybrid",
"params": 352379904,
"layers": 28,
"hidden": 1024,
"heads": 16,
"kv": 4,
"inter": 2048,
"vocab": 100352,
"ctx": 32768,
"experts": null,
"topk": null,
"window": null,
"ltypes": [
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention"
],
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 2390,
"mod": "2025-10-23",
"tags": [
"transformers",
"safetensors",
"granitemoehybrid",
"text-generation",
"language",
"granite-4.0",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-4.0-350m-base",
"impl": {
"rope_theta": 10000000,
"rms_norm_eps": 1e-05,
"position_embedding_type": "rope",
"max_position_embeddings": 32768,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": true,
"embedding_multiplier": 12,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.263,
"logits_scaling": 4,
"num_local_experts": 0,
"num_experts_per_tok": 0,
"shared_intermediate_size": 2048,
"output_router_logits": false,
"mamba_d_state": 256,
"mamba_d_conv": 4,
"mamba_expand": 2,
"mamba_n_heads": 128,
"mamba_n_groups": 1,
"mamba_chunk_size": 256,
"mamba_d_head": 16,
"mamba_conv_bias": true,
"mamba_proj_bias": false,
"pad_token_id": 100256,
"rope": {
"theta": 10000000
},
"class": "GraniteMoeHybridForCausalLM",
"model_type": "granitemoehybrid"
}
},
{
"name": "granite-4.0-3b-vision",
"arch": "Granite4VisionForConditionalGeneration",
"mt": "granitemoehybrid",
"params": 3997206464,
"layers": 40,
"hidden": 2560,
"heads": 40,
"kv": 8,
"inter": 8192,
"vocab": 100353,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": [
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention"
],
"lic": "apache-2.0",
"pipe": "image-text-to-text",
"dl": 622,
"mod": "2026-04-30",
"tags": [
"transformers",
"safetensors",
"granite4_vision",
"image-text-to-text",
"conversational",
"custom_code",
"en",
"arxiv:2603.27064",
"arxiv:2404.19205",
"arxiv:2412.07626",
"arxiv:2512.10888",
"arxiv:2208.00385",
"arxiv:2502.09927",
"arxiv:2406.04334",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-4.0-3b-vision",
"impl": {
"tie_word_embeddings": true,
"downsample_rate": "4/8",
"class": "Granite4VisionForConditionalGeneration",
"model_type": "granite4_vision",
"text": {
"rope_theta": 10000000,
"rms_norm_eps": 1e-05,
"position_embedding_type": "rope",
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": true,
"embedding_multiplier": 12,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 10,
"num_local_experts": 0,
"num_experts_per_tok": 0,
"shared_intermediate_size": 8192,
"output_router_logits": false,
"mamba_d_state": 256,
"mamba_d_conv": 4,
"mamba_expand": 2,
"mamba_n_heads": 128,
"mamba_n_groups": 1,
"mamba_chunk_size": 256,
"mamba_d_head": 40,
"mamba_conv_bias": true,
"mamba_proj_bias": false,
"pad_token_id": 100256,
"num_hidden_layers": 40,
"hidden_size": 2560,
"num_attention_heads": 40,
"num_key_value_heads": 8,
"intermediate_size": 8192,
"vocab_size": 100353,
"model_type": "granitemoehybrid",
"rope": {
"theta": 10000000
}
},
"vision": {
"layer_norm_eps": 1e-06,
"hidden_act": "gelu_pytorch_tanh",
"attention_dropout": 0.0,
"num_hidden_layers": 27,
"hidden_size": 1152,
"num_attention_heads": 16,
"intermediate_size": 4304,
"model_type": "siglip_vision_model"
}
}
},
{
"name": "granite-4.0-h-1b",
"arch": "GraniteMoeHybridForCausalLM",
"mt": "granitemoehybrid",
"params": 1461538368,
"layers": 40,
"hidden": 1536,
"heads": 12,
"kv": 4,
"inter": 4096,
"vocab": 100352,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": [
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba"
],
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 4768,
"mod": "2025-10-23",
"tags": [
"transformers",
"safetensors",
"granitemoehybrid",
"text-generation",
"language",
"granite-4.0",
"conversational",
"arxiv:0000.00000",
"base_model:ibm-granite/granite-4.0-h-1b-base",
"base_model:finetune:ibm-granite/granite-4.0-h-1b-base",
"license:apache-2.0",
"endpoints_compatible",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-4.0-h-1b",
"impl": {
"rope_theta": 10000,
"rms_norm_eps": 1e-05,
"position_embedding_type": "nope",
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": true,
"embedding_multiplier": 12,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 6,
"num_local_experts": 0,
"num_experts_per_tok": 0,
"shared_intermediate_size": 4096,
"output_router_logits": false,
"mamba_d_state": 128,
"mamba_d_conv": 4,
"mamba_expand": 2,
"mamba_n_heads": 48,
"mamba_n_groups": 1,
"mamba_chunk_size": 256,
"mamba_d_head": 64,
"mamba_conv_bias": true,
"mamba_proj_bias": false,
"pad_token_id": 100256,
"rope": {
"theta": 10000
},
"class": "GraniteMoeHybridForCausalLM",
"model_type": "granitemoehybrid"
}
},
{
"name": "granite-4.0-h-1b-base",
"arch": "GraniteMoeHybridForCausalLM",
"mt": "granitemoehybrid",
"params": 1461538368,
"layers": 40,
"hidden": 1536,
"heads": 12,
"kv": 4,
"inter": 4096,
"vocab": 100352,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": [
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba"
],
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 769,
"mod": "2026-06-12",
"tags": [
"transformers",
"safetensors",
"granitemoehybrid",
"text-generation",
"language",
"granite-4.0",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-4.0-h-1b-base",
"impl": {
"rope_theta": 10000,
"rms_norm_eps": 1e-05,
"position_embedding_type": "nope",
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": true,
"embedding_multiplier": 12,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 6,
"num_local_experts": 0,
"num_experts_per_tok": 0,
"shared_intermediate_size": 4096,
"output_router_logits": false,
"mamba_d_state": 128,
"mamba_d_conv": 4,
"mamba_expand": 2,
"mamba_n_heads": 48,
"mamba_n_groups": 1,
"mamba_chunk_size": 256,
"mamba_d_head": 64,
"mamba_conv_bias": true,
"mamba_proj_bias": false,
"pad_token_id": 100256,
"rope": {
"theta": 10000
},
"class": "GraniteMoeHybridForCausalLM",
"model_type": "granitemoehybrid"
}
},
{
"name": "granite-4.0-h-350m",
"arch": "GraniteMoeHybridForCausalLM",
"mt": "granitemoehybrid",
"params": 340332224,
"layers": 32,
"hidden": 768,
"heads": 12,
"kv": 4,
"inter": 2048,
"vocab": 100352,
"ctx": 32768,
"experts": null,
"topk": null,
"window": null,
"ltypes": [
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba"
],
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 4318,
"mod": "2025-10-23",
"tags": [
"transformers",
"safetensors",
"granitemoehybrid",
"text-generation",
"language",
"granite-4.0",
"conversational",
"arxiv:0000.00000",
"base_model:ibm-granite/granite-4.0-h-350m-base",
"base_model:finetune:ibm-granite/granite-4.0-h-350m-base",
"license:apache-2.0",
"eval-results",
"endpoints_compatible",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-4.0-h-350m",
"impl": {
"rope_theta": 10000,
"rms_norm_eps": 1e-05,
"position_embedding_type": "nope",
"max_position_embeddings": 32768,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": true,
"embedding_multiplier": 12,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.246,
"logits_scaling": 3,
"num_local_experts": 0,
"num_experts_per_tok": 0,
"shared_intermediate_size": 2048,
"output_router_logits": false,
"mamba_d_state": 128,
"mamba_d_conv": 4,
"mamba_expand": 2,
"mamba_n_heads": 48,
"mamba_n_groups": 1,
"mamba_chunk_size": 256,
"mamba_d_head": 32,
"mamba_conv_bias": true,
"mamba_proj_bias": false,
"pad_token_id": 100256,
"rope": {
"theta": 10000
},
"class": "GraniteMoeHybridForCausalLM",
"model_type": "granitemoehybrid"
}
},
{
"name": "granite-4.0-h-350m-base",
"arch": "GraniteMoeHybridForCausalLM",
"mt": "granitemoehybrid",
"params": 340332224,
"layers": 32,
"hidden": 768,
"heads": 12,
"kv": 4,
"inter": 2048,
"vocab": 100352,
"ctx": 32768,
"experts": null,
"topk": null,
"window": null,
"ltypes": [
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba"
],
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 3490,
"mod": "2025-10-23",
"tags": [
"transformers",
"safetensors",
"granitemoehybrid",
"text-generation",
"language",
"granite-4.0",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-4.0-h-350m-base",
"impl": {
"rope_theta": 10000,
"rms_norm_eps": 1e-05,
"position_embedding_type": "nope",
"max_position_embeddings": 32768,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": true,
"embedding_multiplier": 12,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.246,
"logits_scaling": 3,
"num_local_experts": 0,
"num_experts_per_tok": 0,
"shared_intermediate_size": 2048,
"output_router_logits": false,
"mamba_d_state": 128,
"mamba_d_conv": 4,
"mamba_expand": 2,
"mamba_n_heads": 48,
"mamba_n_groups": 1,
"mamba_chunk_size": 256,
"mamba_d_head": 32,
"mamba_conv_bias": true,
"mamba_proj_bias": false,
"pad_token_id": 100256,
"rope": {
"theta": 10000
},
"class": "GraniteMoeHybridForCausalLM",
"model_type": "granitemoehybrid"
}
},
{
"name": "granite-4.0-h-micro",
"arch": "GraniteMoeHybridForCausalLM",
"mt": "granitemoehybrid",
"params": 3191396096,
"layers": 40,
"hidden": 2048,
"heads": 32,
"kv": 8,
"inter": 8192,
"vocab": 100352,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": [
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba"
],
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 10936,
"mod": "2025-11-03",
"tags": [
"transformers",
"safetensors",
"granitemoehybrid",
"text-generation",
"language",
"granite-4.0",
"conversational",
"arxiv:0000.00000",
"license:apache-2.0",
"endpoints_compatible",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-4.0-h-micro",
"impl": {
"rope_theta": 10000,
"rms_norm_eps": 1e-05,
"position_embedding_type": "nope",
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": true,
"embedding_multiplier": 12,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 8,
"num_local_experts": 0,
"num_experts_per_tok": 0,
"shared_intermediate_size": 8192,
"output_router_logits": false,
"mamba_d_state": 128,
"mamba_d_conv": 4,
"mamba_expand": 2,
"mamba_n_heads": 64,
"mamba_n_groups": 1,
"mamba_chunk_size": 256,
"mamba_d_head": 64,
"mamba_conv_bias": true,
"mamba_proj_bias": false,
"pad_token_id": 100256,
"rope": {
"theta": 10000
},
"class": "GraniteMoeHybridForCausalLM",
"model_type": "granitemoehybrid"
}
},
{
"name": "granite-4.0-h-micro-base",
"arch": "GraniteMoeHybridForCausalLM",
"mt": "granitemoehybrid",
"params": 3191396096,
"layers": 40,
"hidden": 2048,
"heads": 32,
"kv": 8,
"inter": 8192,
"vocab": 100352,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": [
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba"
],
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 636,
"mod": "2025-11-03",
"tags": [
"transformers",
"safetensors",
"granitemoehybrid",
"text-generation",
"language",
"granite-4.0",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-4.0-h-micro-base",
"impl": {
"rope_theta": 10000,
"rms_norm_eps": 1e-05,
"position_embedding_type": "nope",
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": true,
"embedding_multiplier": 12,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 8,
"num_local_experts": 0,
"num_experts_per_tok": 0,
"shared_intermediate_size": 8192,
"output_router_logits": false,
"mamba_d_state": 128,
"mamba_d_conv": 4,
"mamba_expand": 2,
"mamba_n_heads": 64,
"mamba_n_groups": 1,
"mamba_chunk_size": 256,
"mamba_d_head": 64,
"mamba_conv_bias": true,
"mamba_proj_bias": false,
"pad_token_id": 100256,
"rope": {
"theta": 10000
},
"class": "GraniteMoeHybridForCausalLM",
"model_type": "granitemoehybrid"
}
},
{
"name": "granite-4.0-h-small",
"arch": "GraniteMoeHybridForCausalLM",
"mt": "granitemoehybrid",
"params": 32207337984,
"layers": 40,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 768,
"vocab": 100352,
"ctx": 131072,
"experts": 72,
"topk": 10,
"window": null,
"ltypes": [
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba"
],
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 18776,
"mod": "2025-11-03",
"tags": [
"transformers",
"safetensors",
"granitemoehybrid",
"text-generation",
"language",
"granite-4.0",
"conversational",
"arxiv:0000.00000",
"license:apache-2.0",
"endpoints_compatible",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-4.0-h-small",
"impl": {
"rope_theta": 10000,
"rms_norm_eps": 1e-05,
"position_embedding_type": "nope",
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": true,
"embedding_multiplier": 12,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 16,
"num_local_experts": 72,
"num_experts_per_tok": 10,
"shared_intermediate_size": 1536,
"output_router_logits": false,
"mamba_d_state": 128,
"mamba_d_conv": 4,
"mamba_expand": 2,
"mamba_n_heads": 128,
"mamba_n_groups": 1,
"mamba_chunk_size": 256,
"mamba_d_head": 64,
"mamba_conv_bias": true,
"mamba_proj_bias": false,
"pad_token_id": 100256,
"rope": {
"theta": 10000
},
"class": "GraniteMoeHybridForCausalLM",
"model_type": "granitemoehybrid"
}
},
{
"name": "granite-4.0-h-small-base",
"arch": "GraniteMoeHybridForCausalLM",
"mt": "granitemoehybrid",
"params": 32207337984,
"layers": 40,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 768,
"vocab": 100352,
"ctx": 131072,
"experts": 72,
"topk": 10,
"window": null,
"ltypes": [
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba"
],
"lic": "apache-2.0",
"pipe": null,
"dl": 275,
"mod": "2025-11-03",
"tags": [
"transformersd",
"safetensors",
"granitemoehybrid",
"language",
"granite-4.0",
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/granite-4.0-h-small-base",
"impl": {
"rope_theta": 10000,
"rms_norm_eps": 1e-05,
"position_embedding_type": "nope",
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": true,
"embedding_multiplier": 12,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 16,
"num_local_experts": 72,
"num_experts_per_tok": 10,
"shared_intermediate_size": 1536,
"output_router_logits": false,
"mamba_d_state": 128,
"mamba_d_conv": 4,
"mamba_expand": 2,
"mamba_n_heads": 128,
"mamba_n_groups": 1,
"mamba_chunk_size": 256,
"mamba_d_head": 64,
"mamba_conv_bias": true,
"mamba_proj_bias": false,
"pad_token_id": 100256,
"rope": {
"theta": 10000
},
"class": "GraniteMoeHybridForCausalLM",
"model_type": "granitemoehybrid"
}
},
{
"name": "granite-4.0-h-tiny",
"arch": "GraniteMoeHybridForCausalLM",
"mt": "granitemoehybrid",
"params": 6939037248,
"layers": 40,
"hidden": 1536,
"heads": 12,
"kv": 4,
"inter": 512,
"vocab": 100352,
"ctx": 131072,
"experts": 64,
"topk": 6,
"window": null,
"ltypes": [
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba"
],
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 57724,
"mod": "2025-11-03",
"tags": [
"transformers",
"safetensors",
"granitemoehybrid",
"text-generation",
"language",
"granite-4.0",
"conversational",
"arxiv:0000.00000",
"license:apache-2.0",
"eval-results",
"endpoints_compatible",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-4.0-h-tiny",
"impl": {
"rope_theta": 10000,
"rms_norm_eps": 1e-05,
"position_embedding_type": "nope",
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": true,
"embedding_multiplier": 12,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 6,
"num_local_experts": 64,
"num_experts_per_tok": 6,
"shared_intermediate_size": 1024,
"output_router_logits": false,
"mamba_d_state": 128,
"mamba_d_conv": 4,
"mamba_expand": 2,
"mamba_n_heads": 48,
"mamba_n_groups": 1,
"mamba_chunk_size": 256,
"mamba_d_head": 64,
"mamba_conv_bias": true,
"mamba_proj_bias": false,
"pad_token_id": 100256,
"rope": {
"theta": 10000
},
"class": "GraniteMoeHybridForCausalLM",
"model_type": "granitemoehybrid"
}
},
{
"name": "granite-4.0-h-tiny-base",
"arch": "GraniteMoeHybridForCausalLM",
"mt": "granitemoehybrid",
"params": 6939037248,
"layers": 40,
"hidden": 1536,
"heads": 12,
"kv": 4,
"inter": 512,
"vocab": 100352,
"ctx": 131072,
"experts": 64,
"topk": 6,
"window": null,
"ltypes": [
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba"
],
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 423,
"mod": "2025-11-03",
"tags": [
"transformers",
"safetensors",
"granitemoehybrid",
"text-generation",
"language",
"granite-4.0",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-4.0-h-tiny-base",
"impl": {
"rope_theta": 10000,
"rms_norm_eps": 1e-05,
"position_embedding_type": "nope",
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": true,
"embedding_multiplier": 12,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 6,
"num_local_experts": 64,
"num_experts_per_tok": 6,
"shared_intermediate_size": 1024,
"output_router_logits": false,
"mamba_d_state": 128,
"mamba_d_conv": 4,
"mamba_expand": 2,
"mamba_n_heads": 48,
"mamba_n_groups": 1,
"mamba_chunk_size": 256,
"mamba_d_head": 64,
"mamba_conv_bias": true,
"mamba_proj_bias": false,
"pad_token_id": 100256,
"rope": {
"theta": 10000
},
"class": "GraniteMoeHybridForCausalLM",
"model_type": "granitemoehybrid"
}
},
{
"name": "granite-4.0-micro",
"arch": "GraniteMoeHybridForCausalLM",
"mt": "granitemoehybrid",
"params": 3402836480,
"layers": 40,
"hidden": 2560,
"heads": 40,
"kv": 8,
"inter": 8192,
"vocab": 100352,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": [
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention"
],
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 48852,
"mod": "2025-11-03",
"tags": [
"transformers",
"safetensors",
"granitemoehybrid",
"text-generation",
"language",
"granite-4.0",
"conversational",
"arxiv:0000.00000",
"license:apache-2.0",
"endpoints_compatible",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-4.0-micro",
"impl": {
"rope_theta": 10000000,
"rms_norm_eps": 1e-05,
"position_embedding_type": "rope",
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": true,
"embedding_multiplier": 12,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 10,
"num_local_experts": 0,
"num_experts_per_tok": 0,
"shared_intermediate_size": 8192,
"output_router_logits": false,
"mamba_d_state": 256,
"mamba_d_conv": 4,
"mamba_expand": 2,
"mamba_n_heads": 128,
"mamba_n_groups": 1,
"mamba_chunk_size": 256,
"mamba_d_head": 40,
"mamba_conv_bias": true,
"mamba_proj_bias": false,
"pad_token_id": 100256,
"rope": {
"theta": 10000000
},
"class": "GraniteMoeHybridForCausalLM",
"model_type": "granitemoehybrid"
}
},
{
"name": "granite-4.0-micro-base",
"arch": "GraniteMoeHybridForCausalLM",
"mt": "granitemoehybrid",
"params": 3402836480,
"layers": 40,
"hidden": 2560,
"heads": 40,
"kv": 8,
"inter": 8192,
"vocab": 100352,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": [
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention"
],
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 1925,
"mod": "2025-11-03",
"tags": [
"transformers",
"safetensors",
"granitemoehybrid",
"text-generation",
"language",
"granite-4.0",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-4.0-micro-base",
"impl": {
"rope_theta": 10000000,
"rms_norm_eps": 1e-05,
"position_embedding_type": "rope",
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": true,
"embedding_multiplier": 12,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 10,
"num_local_experts": 0,
"num_experts_per_tok": 0,
"shared_intermediate_size": 8192,
"output_router_logits": false,
"mamba_d_state": 256,
"mamba_d_conv": 4,
"mamba_expand": 2,
"mamba_n_heads": 128,
"mamba_n_groups": 1,
"mamba_chunk_size": 256,
"mamba_d_head": 40,
"mamba_conv_bias": true,
"mamba_proj_bias": false,
"pad_token_id": 100256,
"rope": {
"theta": 10000000
},
"class": "GraniteMoeHybridForCausalLM",
"model_type": "granitemoehybrid"
}
},
{
"name": "granite-4.0-tiny-base-preview",
"arch": "GraniteMoeHybridForCausalLM",
"mt": "granitemoehybrid",
"params": 6673296960,
"layers": 40,
"hidden": 1536,
"heads": 12,
"kv": 4,
"inter": 512,
"vocab": 50304,
"ctx": 131072,
"experts": 62,
"topk": 6,
"window": null,
"ltypes": [
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba"
],
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 228,
"mod": "2025-11-03",
"tags": [
"transformers",
"safetensors",
"granitemoehybrid",
"text-generation",
"language",
"granite-4.0",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-4.0-tiny-base-preview",
"impl": {
"rope_theta": 10000,
"rms_norm_eps": 1e-05,
"position_embedding_type": "nope",
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": true,
"embedding_multiplier": 12,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 6,
"num_local_experts": 62,
"num_experts_per_tok": 6,
"shared_intermediate_size": 1024,
"output_router_logits": false,
"mamba_d_state": 128,
"mamba_d_conv": 4,
"mamba_expand": 2,
"mamba_n_heads": 48,
"mamba_n_groups": 1,
"mamba_chunk_size": 256,
"mamba_d_head": 64,
"mamba_conv_bias": true,
"mamba_proj_bias": false,
"pad_token_id": 0,
"rope": {
"theta": 10000
},
"class": "GraniteMoeHybridForCausalLM",
"model_type": "granitemoehybrid"
}
},
{
"name": "granite-4.0-tiny-preview",
"arch": "GraniteMoeHybridForCausalLM",
"mt": "granitemoehybrid",
"params": 6671539776,
"layers": 40,
"hidden": 1536,
"heads": 12,
"kv": 4,
"inter": 512,
"vocab": 49160,
"ctx": 131072,
"experts": 62,
"topk": 6,
"window": null,
"ltypes": [
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"mamba",
"attention",
"mamba",
"mamba",
"mamba",
"mamba"
],
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 61713,
"mod": "2025-11-03",
"tags": [
"transformers",
"safetensors",
"granitemoehybrid",
"text-generation",
"language",
"granite-4.0",
"conversational",
"base_model:ibm-granite/granite-4.0-tiny-base-preview",
"base_model:finetune:ibm-granite/granite-4.0-tiny-base-preview",
"license:apache-2.0",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-4.0-tiny-preview",
"impl": {
"rope_theta": 10000,
"rms_norm_eps": 1e-05,
"position_embedding_type": "nope",
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": true,
"embedding_multiplier": 12,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 6,
"num_local_experts": 62,
"num_experts_per_tok": 6,
"shared_intermediate_size": 1024,
"output_router_logits": false,
"mamba_d_state": 128,
"mamba_d_conv": 4,
"mamba_expand": 2,
"mamba_n_heads": 48,
"mamba_n_groups": 1,
"mamba_chunk_size": 256,
"mamba_d_head": 64,
"mamba_conv_bias": true,
"mamba_proj_bias": false,
"pad_token_id": 0,
"rope": {
"theta": 10000
},
"class": "GraniteMoeHybridForCausalLM",
"model_type": "granitemoehybrid"
}
},
{
"name": "granite-4.1-30b",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 28865728512,
"layers": 64,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 32768,
"vocab": 100352,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 700161,
"mod": "2026-05-04",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"language",
"granite-4.1",
"conversational",
"arxiv:0000.00000",
"license:apache-2.0",
"eval-results",
"endpoints_compatible",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-4.1-30b",
"impl": {
"rope_theta": 50000000,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.175,
"logits_scaling": 16.0,
"pad_token_id": 100256,
"rope": {
"theta": 50000000
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-4.1-30b-base",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 28865728512,
"layers": 64,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 32768,
"vocab": 100352,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 2250,
"mod": "2026-04-28",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"language",
"granite-4.1",
"arxiv:2603.17074",
"license:apache-2.0",
"endpoints_compatible",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-4.1-30b-base",
"impl": {
"rope_theta": 50000000,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.175,
"logits_scaling": 16.0,
"pad_token_id": 100256,
"rope": {
"theta": 50000000
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-4.1-3b",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 3402836480,
"layers": 40,
"hidden": 2560,
"heads": 40,
"kv": 8,
"inter": 8192,
"vocab": 100352,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 392654,
"mod": "2026-05-04",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"language",
"granite-4.1",
"conversational",
"arxiv:0000.00000",
"license:apache-2.0",
"eval-results",
"endpoints_compatible",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-4.1-3b",
"impl": {
"rope_theta": 10000000,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 10.0,
"pad_token_id": 100256,
"rope": {
"theta": 10000000
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-4.1-3b-base",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 3402836480,
"layers": 40,
"hidden": 2560,
"heads": 40,
"kv": 8,
"inter": 8192,
"vocab": 100352,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 13964,
"mod": "2026-04-28",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"language",
"granite-4.1",
"arxiv:2603.17074",
"license:apache-2.0",
"endpoints_compatible",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-4.1-3b-base",
"impl": {
"rope_theta": 10000000,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 10.0,
"pad_token_id": 100256,
"rope": {
"theta": 10000000
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-4.1-8b",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 8791592960,
"layers": 40,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 12800,
"vocab": 100352,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 217938,
"mod": "2026-05-04",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"language",
"granite-4.1",
"conversational",
"arxiv:0000.00000",
"license:apache-2.0",
"eval-results",
"endpoints_compatible",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-4.1-8b",
"impl": {
"rope_theta": 10000000,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 16.0,
"pad_token_id": 100256,
"rope": {
"theta": 10000000
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-4.1-8b-base",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 8380551168,
"layers": 40,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 12800,
"vocab": 100352,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 3757,
"mod": "2026-04-28",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"language",
"granite-4.1",
"arxiv:2603.17074",
"license:apache-2.0",
"endpoints_compatible",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-4.1-8b-base",
"impl": {
"rope_theta": 10000000,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 16.0,
"pad_token_id": 100256,
"rope": {
"theta": 10000000
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-4.2-30b",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 29276770304,
"layers": 64,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 32768,
"vocab": 100352,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 23974,
"mod": "2026-09-04",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"granite-4.2",
"reasoning",
"thinking",
"tool-calling",
"ibm",
"conversational",
"en",
"de",
"es",
"fr",
"ja",
"pt",
"ar",
"cs",
"it",
"ko",
"nl",
"zh",
"base_model:ibm-granite/granite-4.1-30b-base",
"base_model:finetune:ibm-granite/granite-4.1-30b-base",
"license:apache-2.0",
"eval-results",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-4.2-30b",
"impl": {
"rope_theta": 50000000,
"rope_parameters": {
"rope_theta": 50000000,
"rope_type": "default"
},
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"mlp_bias": false,
"tie_word_embeddings": false,
"embedding_multiplier": 1.0,
"attention_multiplier": 0.0078125,
"residual_multiplier": 1.0,
"logits_scaling": 1.0,
"pad_token_id": 100257,
"rope": {
"from": "rope_parameters",
"theta": 50000000,
"rope_type": "default"
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-4.2-3b",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 3659737600,
"layers": 40,
"hidden": 2560,
"heads": 40,
"kv": 8,
"inter": 8192,
"vocab": 100352,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 45488,
"mod": "2026-09-02",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"granite-4.2",
"reasoning",
"thinking",
"tool-calling",
"ibm",
"conversational",
"en",
"de",
"es",
"fr",
"ja",
"pt",
"ar",
"cs",
"it",
"ko",
"nl",
"zh",
"base_model:ibm-granite/granite-4.1-3b-base",
"base_model:finetune:ibm-granite/granite-4.1-3b-base",
"license:apache-2.0",
"eval-results",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-4.2-3b",
"impl": {
"rope_theta": 10000000,
"rope_parameters": {
"rope_theta": 10000000,
"rope_type": "default"
},
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"mlp_bias": false,
"tie_word_embeddings": false,
"embedding_multiplier": 1.0,
"attention_multiplier": 0.015625,
"residual_multiplier": 1.0,
"logits_scaling": 1.0,
"pad_token_id": 100257,
"rope": {
"from": "rope_parameters",
"theta": 10000000,
"rope_type": "default"
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-4.2-8b",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 8791592960,
"layers": 40,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 12800,
"vocab": 100352,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 92032,
"mod": "2026-09-04",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"granite-4.2",
"reasoning",
"thinking",
"tool-calling",
"ibm",
"conversational",
"en",
"de",
"es",
"fr",
"ja",
"pt",
"ar",
"cs",
"it",
"ko",
"nl",
"zh",
"base_model:ibm-granite/granite-4.1-8b-base",
"base_model:finetune:ibm-granite/granite-4.1-8b-base",
"license:apache-2.0",
"eval-results",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-4.2-8b",
"impl": {
"rope_theta": 10000000,
"rope_parameters": {
"rope_theta": 10000000,
"rope_type": "default"
},
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"mlp_bias": false,
"tie_word_embeddings": false,
"embedding_multiplier": 1.0,
"attention_multiplier": 0.0078125,
"residual_multiplier": 1.0,
"logits_scaling": 1.0,
"pad_token_id": 100257,
"rope": {
"from": "rope_parameters",
"theta": 10000000,
"rope_type": "default"
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-7b-base",
"arch": "LlamaForCausalLM",
"mt": "llama",
"params": 6738415616,
"layers": 32,
"hidden": 4096,
"heads": 32,
"kv": 32,
"inter": 11008,
"vocab": 32000,
"ctx": 4096,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 596,
"mod": "2024-12-19",
"tags": [
"transformers",
"safetensors",
"llama",
"text-generation",
"license:apache-2.0",
"text-generation-inference",
"endpoints_compatible",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-7b-base",
"impl": {
"rope_theta": 10000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 4096,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": false,
"rope": {
"theta": 10000.0
},
"class": "LlamaForCausalLM",
"model_type": "llama"
}
},
{
"name": "granite-7b-instruct",
"arch": "LlamaForCausalLM",
"mt": "llama",
"params": 6738481152,
"layers": 32,
"hidden": 4096,
"heads": 32,
"kv": 32,
"inter": 11008,
"vocab": 32008,
"ctx": 4096,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 552,
"mod": "2024-12-19",
"tags": [
"transformers",
"safetensors",
"llama",
"text-generation",
"granite",
"ibm",
"lab",
"labrador",
"labradorite",
"conversational",
"en",
"arxiv:2403.01081",
"base_model:ibm-granite/granite-7b-base",
"base_model:finetune:ibm-granite/granite-7b-base",
"license:apache-2.0",
"text-generation-inference",
"endpoints_compatible",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-7b-instruct",
"impl": {
"rope_theta": 10000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 4096,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": false,
"rope": {
"theta": 10000.0
},
"class": "LlamaForCausalLM",
"model_type": "llama"
}
},
{
"name": "granite-7b-instruct-accelerator",
"arch": "MLPSpeculatorPreTrainedModel",
"mt": "mlp_speculator",
"params": 1394974720,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": 32008,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": null,
"dl": 113,
"mod": "2024-05-20",
"tags": [
"transformers",
"safetensors",
"mlp_speculator",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-7b-instruct-accelerator",
"impl": {
"emb_dim": 4096,
"inner_dim": 4096,
"n_predict": 5,
"n_candidates": 5,
"top_k_tokens_per_head": [
4,
3,
2,
2,
2
],
"class": "MLPSpeculatorPreTrainedModel",
"model_type": "mlp_speculator"
}
},
{
"name": "granite-8b-code-base-128k",
"arch": "LlamaForCausalLM",
"mt": "llama",
"params": 8054910976,
"layers": 36,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 14336,
"vocab": 49152,
"ctx": 128000,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 279,
"mod": "2026-04-25",
"tags": [
"transformers",
"safetensors",
"llama",
"text-generation",
"code",
"granite",
"dataset:codeparrot/github-code-clean",
"dataset:bigcode/starcoderdata",
"dataset:open-web-math/open-web-math",
"dataset:math-ai/StackMathQA",
"arxiv:2405.04324",
"license:apache-2.0",
"model-index",
"text-generation-inference",
"region:us"
],
"id": "ibm-granite/granite-8b-code-base-128k",
"impl": {
"rope_theta": 10000000,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 128000,
"hidden_act": "silu",
"attention_bias": true,
"attention_dropout": 0.1,
"mlp_bias": true,
"tie_word_embeddings": true,
"pad_token_id": 0,
"rope": {
"theta": 10000000
},
"class": "LlamaForCausalLM",
"model_type": "llama"
}
},
{
"name": "granite-8b-code-base-4k",
"arch": "LlamaForCausalLM",
"mt": "llama",
"params": 8054910976,
"layers": 36,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 14336,
"vocab": 49152,
"ctx": 4096,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 1812,
"mod": "2026-04-25",
"tags": [
"transformers",
"safetensors",
"llama",
"text-generation",
"code",
"granite",
"dataset:codeparrot/github-code-clean",
"dataset:bigcode/starcoderdata",
"dataset:open-web-math/open-web-math",
"dataset:math-ai/StackMathQA",
"arxiv:2405.04324",
"license:apache-2.0",
"model-index",
"text-generation-inference",
"region:us"
],
"id": "ibm-granite/granite-8b-code-base-4k",
"impl": {
"rope_theta": 10000,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 4096,
"hidden_act": "silu",
"attention_bias": true,
"attention_dropout": 0.1,
"mlp_bias": true,
"tie_word_embeddings": true,
"pad_token_id": 0,
"rope": {
"theta": 10000
},
"class": "LlamaForCausalLM",
"model_type": "llama"
}
},
{
"name": "granite-8b-code-instruct-128k",
"arch": "LlamaForCausalLM",
"mt": "llama",
"params": 8054910976,
"layers": 36,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 14336,
"vocab": 49152,
"ctx": 128000,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 1196,
"mod": "2026-04-25",
"tags": [
"transformers",
"safetensors",
"llama",
"text-generation",
"code",
"granite",
"conversational",
"dataset:bigcode/commitpackft",
"dataset:TIGER-Lab/MathInstruct",
"dataset:meta-math/MetaMathQA",
"dataset:glaiveai/glaive-code-assistant-v3",
"dataset:glaive-function-calling-v2",
"dataset:bugdaryan/sql-create-context-instruction",
"dataset:garage-bAInd/Open-Platypus",
"dataset:nvidia/HelpSteer",
"dataset:bigcode/self-oss-instruct-sc2-exec-filter-50k",
"arxiv:2407.13739",
"license:apache-2.0",
"model-index",
"text-generation-inference",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-8b-code-instruct-128k",
"impl": {
"rope_theta": 10000000,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 128000,
"hidden_act": "silu",
"attention_bias": true,
"attention_dropout": 0.1,
"mlp_bias": true,
"tie_word_embeddings": true,
"pad_token_id": 0,
"rope": {
"theta": 10000000
},
"class": "LlamaForCausalLM",
"model_type": "llama"
}
},
{
"name": "granite-8b-code-instruct-4k",
"arch": "LlamaForCausalLM",
"mt": "llama",
"params": 8054910976,
"layers": 36,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 14336,
"vocab": 49152,
"ctx": 4096,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 868,
"mod": "2026-04-25",
"tags": [
"transformers",
"safetensors",
"llama",
"text-generation",
"code",
"granite",
"conversational",
"dataset:bigcode/commitpackft",
"dataset:TIGER-Lab/MathInstruct",
"dataset:meta-math/MetaMathQA",
"dataset:glaiveai/glaive-code-assistant-v3",
"dataset:glaive-function-calling-v2",
"dataset:bugdaryan/sql-create-context-instruction",
"dataset:garage-bAInd/Open-Platypus",
"dataset:nvidia/HelpSteer",
"arxiv:2405.04324",
"base_model:ibm-granite/granite-8b-code-base-4k",
"base_model:finetune:ibm-granite/granite-8b-code-base-4k",
"license:apache-2.0",
"model-index",
"text-generation-inference",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-8b-code-instruct-4k",
"impl": {
"rope_theta": 10000,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 4096,
"hidden_act": "silu",
"attention_bias": true,
"attention_dropout": 0.1,
"mlp_bias": true,
"tie_word_embeddings": true,
"pad_token_id": 0,
"rope": {
"theta": 10000
},
"class": "LlamaForCausalLM",
"model_type": "llama"
}
},
{
"name": "granite-8b-code-instruct-accelerator",
"arch": "MLPSpeculatorPreTrainedModel",
"mt": "mlp_speculator",
"params": 2097192960,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": 49152,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": null,
"dl": 117,
"mod": "2024-05-29",
"tags": [
"transformers",
"safetensors",
"mlp_speculator",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-8b-code-instruct-accelerator",
"impl": {
"emb_dim": 4096,
"inner_dim": 4096,
"n_predict": 5,
"n_candidates": 5,
"top_k_tokens_per_head": [
4,
3,
2,
2,
2
],
"class": "MLPSpeculatorPreTrainedModel",
"model_type": "mlp_speculator"
}
},
{
"name": "granite-docling-258M",
"arch": "Idefics3ForConditionalGeneration",
"mt": "llama",
"params": 257517120,
"layers": 30,
"hidden": 576,
"heads": 9,
"kv": 3,
"inter": 1536,
"vocab": 100352,
"ctx": 8192,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "image-text-to-text",
"dl": 49707,
"mod": "2025-09-23",
"tags": [
"transformers",
"safetensors",
"idefics3",
"image-text-to-text",
"text-generation",
"documents",
"code",
"formula",
"chart",
"ocr",
"layout",
"table",
"document-parse",
"docling",
"granite",
"extraction",
"math",
"conversational",
"en",
"dataset:ds4sd/SynthCodeNet",
"dataset:ds4sd/SynthFormulaNet",
"dataset:ds4sd/SynthChartNet",
"dataset:HuggingFaceM4/DoclingMatix",
"arxiv:2501.17887",
"arxiv:2503.11576",
"arxiv:2305.03393",
"license:apache-2.0",
"endpoints_compatible",
"deploy:azure",
"region:us"
],
"id": "ibm-granite/granite-docling-258M",
"impl": {
"tie_word_embeddings": true,
"pad_token_id": 100257,
"class": "Idefics3ForConditionalGeneration",
"model_type": "idefics3",
"text": {
"rope_theta": 100000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 8192,
"hidden_act": "silu",
"head_dim": 64,
"attention_bias": false,
"attention_dropout": 0.0,
"mlp_bias": false,
"tie_word_embeddings": true,
"pad_token_id": 100257,
"num_hidden_layers": 30,
"hidden_size": 576,
"num_attention_heads": 9,
"num_key_value_heads": 3,
"intermediate_size": 1536,
"vocab_size": 100352,
"model_type": "llama",
"rope": {
"theta": 100000.0
}
},
"vision": {
"layer_norm_eps": 1e-06,
"hidden_act": "gelu_pytorch_tanh",
"attention_dropout": 0.0,
"num_hidden_layers": 12,
"hidden_size": 768,
"num_attention_heads": 12,
"intermediate_size": 3072,
"model_type": "idefics3_vision"
}
}
},
{
"name": "granite-embedding-107m-multilingual",
"arch": "XLMRobertaModel",
"mt": "xlm-roberta",
"params": 106994304,
"layers": 6,
"hidden": 384,
"heads": 12,
"kv": null,
"inter": 1536,
"vocab": 250002,
"ctx": 514,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "sentence-similarity",
"dl": 64509,
"mod": "2025-08-19",
"tags": [
"transformers",
"pytorch",
"onnx",
"safetensors",
"xlm-roberta",
"feature-extraction",
"language",
"granite",
"embeddings",
"multilingual",
"mteb",
"sentence-similarity",
"en",
"ar",
"cs",
"de",
"es",
"fr",
"it",
"ja",
"ko",
"nl",
"pt",
"zh",
"arxiv:2502.20204",
"arxiv:0000.00000",
"license:apache-2.0",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-embedding-107m-multilingual",
"impl": {
"layer_norm_eps": 1e-12,
"position_embedding_type": "absolute",
"max_position_embeddings": 514,
"hidden_act": "gelu",
"type_vocab_size": 2,
"pad_token_id": 1,
"class": "XLMRobertaModel",
"model_type": "xlm-roberta"
}
},
{
"name": "granite-embedding-125m-english",
"arch": "RobertaModel",
"mt": "roberta",
"params": 124645632,
"layers": 12,
"hidden": 768,
"heads": 12,
"kv": null,
"inter": 3072,
"vocab": 50265,
"ctx": 514,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "sentence-similarity",
"dl": 108883,
"mod": "2025-08-19",
"tags": [
"sentence-transformers",
"pytorch",
"onnx",
"safetensors",
"roberta",
"feature-extraction",
"language",
"granite",
"embeddings",
"mteb",
"transformers",
"sentence-similarity",
"en",
"arxiv:2502.20204",
"arxiv:0000.00000",
"license:apache-2.0",
"model-index",
"eval-results",
"text-embeddings-inference",
"endpoints_compatible",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-embedding-125m-english",
"impl": {
"layer_norm_eps": 1e-05,
"position_embedding_type": "absolute",
"max_position_embeddings": 514,
"hidden_act": "gelu",
"type_vocab_size": 1,
"pad_token_id": 1,
"class": "RobertaModel",
"model_type": "roberta"
}
},
{
"name": "granite-embedding-278m-multilingual",
"arch": "XLMRobertaModel",
"mt": "xlm-roberta",
"params": 278043648,
"layers": 12,
"hidden": 768,
"heads": 12,
"kv": null,
"inter": 3072,
"vocab": 250002,
"ctx": 514,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "sentence-similarity",
"dl": 62095,
"mod": "2025-08-19",
"tags": [
"transformers",
"pytorch",
"onnx",
"safetensors",
"xlm-roberta",
"feature-extraction",
"language",
"granite",
"embeddings",
"multilingual",
"mteb",
"sentence-transformers",
"sentence-similarity",
"en",
"ar",
"cs",
"de",
"es",
"fr",
"it",
"ja",
"ko",
"nl",
"pt",
"zh",
"arxiv:2502.20204",
"arxiv:0000.00000",
"license:apache-2.0",
"model-index",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-embedding-278m-multilingual",
"impl": {
"layer_norm_eps": 1e-05,
"position_embedding_type": "absolute",
"max_position_embeddings": 514,
"hidden_act": "gelu",
"type_vocab_size": 1,
"pad_token_id": 1,
"class": "XLMRobertaModel",
"model_type": "xlm-roberta"
}
},
{
"name": "granite-embedding-30m-english",
"arch": "RobertaModel",
"mt": "roberta",
"params": 30295296,
"layers": 6,
"hidden": 384,
"heads": 12,
"kv": null,
"inter": 1536,
"vocab": 50265,
"ctx": 514,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "sentence-similarity",
"dl": 58252,
"mod": "2025-08-22",
"tags": [
"sentence-transformers",
"pytorch",
"onnx",
"safetensors",
"roberta",
"feature-extraction",
"language",
"granite",
"embeddings",
"mteb",
"transformers",
"sentence-similarity",
"en",
"arxiv:2502.20204",
"license:apache-2.0",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-embedding-30m-english",
"impl": {
"layer_norm_eps": 1e-12,
"position_embedding_type": "absolute",
"max_position_embeddings": 514,
"hidden_act": "gelu",
"type_vocab_size": 2,
"pad_token_id": 1,
"class": "RobertaModel",
"model_type": "roberta"
}
},
{
"name": "granite-embedding-30m-sparse",
"arch": "RobertaForMaskedLM",
"mt": "roberta",
"params": 30346329,
"layers": 6,
"hidden": 384,
"heads": 12,
"kv": null,
"inter": 1536,
"vocab": 50265,
"ctx": 514,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "feature-extraction",
"dl": 9229,
"mod": "2025-08-27",
"tags": [
"sentence-transformers",
"pytorch",
"safetensors",
"roberta",
"language",
"granite",
"embeddings",
"sparse-encoder",
"sparse",
"splade",
"feature-extraction",
"en",
"arxiv:2502.20204",
"license:apache-2.0",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-embedding-30m-sparse",
"impl": {
"layer_norm_eps": 1e-12,
"position_embedding_type": "absolute",
"max_position_embeddings": 514,
"hidden_act": "gelu",
"type_vocab_size": 2,
"pad_token_id": 1,
"class": "RobertaForMaskedLM",
"model_type": "roberta"
}
},
{
"name": "granite-embedding-311m-multilingual-r2",
"arch": "ModernBertModel",
"mt": "modernbert",
"params": 311664384,
"layers": 22,
"hidden": 768,
"heads": 12,
"kv": null,
"inter": 1152,
"vocab": 262152,
"ctx": 32768,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "feature-extraction",
"dl": 161110,
"mod": "2026-05-18",
"tags": [
"sentence-transformers",
"onnx",
"safetensors",
"openvino",
"modernbert",
"feature-extraction",
"granite",
"embeddings",
"transformers",
"multilingual",
"mteb",
"sentence-similarity",
"matryoshka",
"ar",
"az",
"bg",
"bn",
"ca",
"cs",
"da",
"de",
"el",
"en",
"es",
"et",
"fa",
"fi",
"fr",
"he",
"hi",
"hr",
"hu",
"id",
"is",
"it",
"ja",
"ka",
"kk",
"km",
"ko",
"lt",
"lv",
"mr",
"ms",
"nl",
"no",
"pl",
"pt",
"ro",
"ru",
"sk",
"sl",
"sq",
"sr",
"sv",
"sw",
"te",
"th",
"tl",
"tr",
"uk",
"ur",
"uz",
"vi",
"zh",
"arxiv:2605.13521",
"arxiv:2205.13147",
"license:apache-2.0",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-embedding-311m-multilingual-r2",
"impl": {
"global_rope_theta": 150000.0,
"local_rope_theta": 160000.0,
"layer_norm_eps": 1e-12,
"position_embedding_type": "absolute",
"max_position_embeddings": 32768,
"hidden_activation": "gelu",
"attention_bias": false,
"attention_dropout": 0.0,
"mlp_bias": false,
"local_attention": 128,
"global_attn_every_n_layers": 3,
"pad_token_id": 0,
"class": "ModernBertModel",
"model_type": "modernbert"
}
},
{
"name": "granite-embedding-97m-multilingual-r2",
"arch": "ModernBertModel",
"mt": "modernbert",
"params": 97441152,
"layers": 12,
"hidden": 384,
"heads": 12,
"kv": null,
"inter": 1536,
"vocab": 180000,
"ctx": 32768,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "feature-extraction",
"dl": 76559,
"mod": "2026-05-18",
"tags": [
"sentence-transformers",
"onnx",
"safetensors",
"openvino",
"modernbert",
"feature-extraction",
"granite",
"embeddings",
"transformers",
"multilingual",
"mteb",
"sentence-similarity",
"ar",
"az",
"bg",
"bn",
"ca",
"cs",
"da",
"de",
"el",
"en",
"es",
"et",
"fa",
"fi",
"fr",
"he",
"hi",
"hr",
"hu",
"id",
"is",
"it",
"ja",
"ka",
"kk",
"km",
"ko",
"lt",
"lv",
"mr",
"ms",
"nl",
"no",
"pl",
"pt",
"ro",
"ru",
"sk",
"sl",
"sq",
"sr",
"sv",
"sw",
"te",
"th",
"tl",
"tr",
"uk",
"ur",
"uz",
"vi",
"zh",
"arxiv:2605.13521",
"license:apache-2.0",
"text-embeddings-inference",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-embedding-97m-multilingual-r2",
"impl": {
"global_rope_theta": 150000.0,
"local_rope_theta": 160000.0,
"layer_norm_eps": 1e-05,
"position_embedding_type": "absolute",
"max_position_embeddings": 32768,
"hidden_activation": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"mlp_bias": false,
"local_attention": 128,
"global_attn_every_n_layers": 3,
"pad_token_id": 179935,
"class": "ModernBertModel",
"model_type": "modernbert"
}
},
{
"name": "granite-embedding-english-r2",
"arch": "ModernBertModel",
"mt": "modernbert",
"params": 149014272,
"layers": 22,
"hidden": 768,
"heads": 12,
"kv": null,
"inter": 1152,
"vocab": 50368,
"ctx": 8192,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "feature-extraction",
"dl": 59028,
"mod": "2026-01-06",
"tags": [
"sentence-transformers",
"pytorch",
"safetensors",
"modernbert",
"feature-extraction",
"granite",
"embeddings",
"transformers",
"mteb",
"en",
"arxiv:2508.21085",
"license:apache-2.0",
"text-embeddings-inference",
"endpoints_compatible",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-embedding-english-r2",
"impl": {
"global_rope_theta": 80000.0,
"local_rope_theta": 10000.0,
"layer_norm_eps": 1e-05,
"position_embedding_type": "absolute",
"max_position_embeddings": 8192,
"hidden_act": "gelu",
"hidden_activation": "gelu",
"attention_bias": false,
"attention_dropout": 0.0,
"mlp_bias": false,
"local_attention": 128,
"global_attn_every_n_layers": 3,
"pad_token_id": 50283,
"class": "ModernBertModel",
"model_type": "modernbert"
}
},
{
"name": "granite-embedding-reranker-english-r2",
"arch": "ModernBertForSequenceClassification",
"mt": "modernbert",
"params": 149605633,
"layers": 22,
"hidden": 768,
"heads": 12,
"kv": null,
"inter": 1152,
"vocab": 50368,
"ctx": 8192,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-ranking",
"dl": 23742,
"mod": "2025-11-06",
"tags": [
"sentence-transformers",
"safetensors",
"modernbert",
"text-classification",
"granite",
"transformers",
"embeddings",
"mteb",
"text-embeddings-inference",
"text-ranking",
"en",
"arxiv:2508.21085",
"base_model:ibm-granite/granite-embedding-english-r2",
"base_model:finetune:ibm-granite/granite-embedding-english-r2",
"license:apache-2.0",
"endpoints_compatible",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-embedding-reranker-english-r2",
"impl": {
"global_rope_theta": 160000,
"local_rope_theta": 40000,
"layer_norm_eps": 1e-05,
"position_embedding_type": "relative_key_query",
"max_position_embeddings": 8192,
"hidden_act": "gelu",
"hidden_activation": "gelu",
"attention_bias": false,
"attention_dropout": 0.0,
"mlp_bias": false,
"local_attention": 128,
"global_attn_every_n_layers": 3,
"pad_token_id": 50283,
"class": "ModernBertForSequenceClassification",
"model_type": "modernbert"
}
},
{
"name": "granite-embedding-small-english-r2",
"arch": "ModernBertModel",
"mt": "modernbert",
"params": 47662464,
"layers": 12,
"hidden": 384,
"heads": 12,
"kv": null,
"inter": 1536,
"vocab": 50368,
"ctx": 8192,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "feature-extraction",
"dl": 6277600,
"mod": "2026-01-21",
"tags": [
"sentence-transformers",
"pytorch",
"safetensors",
"modernbert",
"feature-extraction",
"granite",
"embeddings",
"transformers",
"mteb",
"en",
"arxiv:2508.21085",
"license:apache-2.0",
"text-embeddings-inference",
"endpoints_compatible",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-embedding-small-english-r2",
"impl": {
"global_rope_theta": 80000,
"local_rope_theta": 10000.0,
"layer_norm_eps": 1e-05,
"position_embedding_type": "relative_key_query",
"max_position_embeddings": 8192,
"hidden_act": "gelu",
"hidden_activation": "gelu",
"attention_bias": false,
"attention_dropout": 0.0,
"mlp_bias": false,
"local_attention": 128,
"global_attn_every_n_layers": 3,
"pad_token_id": 50283,
"class": "ModernBertModel",
"model_type": "modernbert"
}
},
{
"name": "granite-geospatial-biomass",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "image-feature-extraction",
"dl": 121,
"mod": "2025-09-08",
"tags": [
"terratorch",
"image-feature-extraction",
"arxiv:2503.20563",
"arxiv:2310.18660",
"arxiv:2406.19888",
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/granite-geospatial-biomass"
},
{
"name": "granite-geospatial-canopyheight",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "image-feature-extraction",
"dl": 14,
"mod": "2025-09-08",
"tags": [
"terratorch",
"pytorch",
"image-feature-extraction",
"arxiv:2310.18660",
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/granite-geospatial-canopyheight"
},
{
"name": "granite-geospatial-land-surface-temperature",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "image-feature-extraction",
"dl": 299,
"mod": "2025-09-08",
"tags": [
"terratorch",
"image-feature-extraction",
"arxiv:2310.18660",
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/granite-geospatial-land-surface-temperature"
},
{
"name": "granite-geospatial-ocean",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": null,
"dl": 80,
"mod": "2025-09-26",
"tags": [
"terratorch",
"Pytorch",
"Earth Observation",
"Foundation Model",
"IBM",
"arxiv:2509.21273",
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/granite-geospatial-ocean"
},
{
"name": "granite-geospatial-uki",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "image-feature-extraction",
"dl": 321,
"mod": "2025-12-03",
"tags": [
"terratorch",
"image-feature-extraction",
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/granite-geospatial-uki"
},
{
"name": "granite-geospatial-uki-flooddetection",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "image-segmentation",
"dl": 194,
"mod": "2025-12-02",
"tags": [
"terratorch",
"image-segmentation",
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/granite-geospatial-uki-flooddetection"
},
{
"name": "granite-geospatial-wxc-downscaling",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "cdla-permissive-2.0",
"pipe": "image-to-image",
"dl": 161,
"mod": "2025-09-08",
"tags": [
"terratorch",
"pytorch",
"image-to-image",
"license:cdla-permissive-2.0",
"region:us"
],
"id": "ibm-granite/granite-geospatial-wxc-downscaling"
},
{
"name": "granite-guardian-3.0-2b",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 2533531648,
"layers": 40,
"hidden": 2048,
"heads": 32,
"kv": 8,
"inter": 8192,
"vocab": 49155,
"ctx": 8192,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 4830,
"mod": "2025-01-03",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"conversational",
"en",
"arxiv:2412.07724",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-guardian-3.0-2b",
"impl": {
"rope_theta": 10000,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 8192,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.1,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 8.0,
"pad_token_id": 0,
"rope": {
"theta": 10000
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-guardian-3.0-8b",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 8170848256,
"layers": 40,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 12800,
"vocab": 49155,
"ctx": 8192,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 567,
"mod": "2025-01-03",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"conversational",
"en",
"arxiv:2412.07724",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-guardian-3.0-8b",
"impl": {
"rope_theta": 10000,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 8192,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.1,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 16.0,
"pad_token_id": 0,
"rope": {
"theta": 10000
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-guardian-3.1-2b",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 2533531648,
"layers": 40,
"hidden": 2048,
"heads": 32,
"kv": 8,
"inter": 8192,
"vocab": 49155,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 2682,
"mod": "2025-03-19",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"conversational",
"en",
"arxiv:2412.07724",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-guardian-3.1-2b",
"impl": {
"rope_theta": 5000000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.1,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 8.0,
"pad_token_id": 0,
"rope": {
"theta": 5000000.0
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-guardian-3.1-8b",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 8170848256,
"layers": 40,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 12800,
"vocab": 49155,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 1408,
"mod": "2025-09-02",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"conversational",
"en",
"arxiv:2412.07724",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-guardian-3.1-8b",
"impl": {
"rope_theta": 10000000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.1,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 16.0,
"pad_token_id": 0,
"rope": {
"theta": 10000000.0
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-guardian-3.2-3b-a800m",
"arch": "GraniteMoeForCausalLM",
"mt": "granitemoe",
"params": 3298793472,
"layers": 32,
"hidden": 1536,
"heads": 24,
"kv": 8,
"inter": 512,
"vocab": 49155,
"ctx": 131072,
"experts": 40,
"topk": 8,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 6439,
"mod": "2025-09-02",
"tags": [
"transformers",
"safetensors",
"granitemoe",
"text-generation",
"conversational",
"arxiv:2412.07724",
"arxiv:2306.11247",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-guardian-3.2-3b-a800m",
"impl": {
"rope_theta": 10000000.0,
"rms_norm_eps": 1e-06,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 6.0,
"num_local_experts": 40,
"num_experts_per_tok": 8,
"output_router_logits": false,
"pad_token_id": 0,
"rope": {
"theta": 10000000.0
},
"class": "GraniteMoeForCausalLM",
"model_type": "granitemoe"
}
},
{
"name": "granite-guardian-3.2-5b",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 5779996672,
"layers": 28,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 12800,
"vocab": 49155,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 2668,
"mod": "2025-09-02",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"conversational",
"arxiv:2412.07724",
"arxiv:2306.11247",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-guardian-3.2-5b",
"impl": {
"rope_theta": 10000000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.1,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 16.0,
"pad_token_id": 0,
"rope": {
"theta": 10000000.0
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-guardian-3.2-5b-lora-factuality-correction",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 0,
"mod": "2025-12-08",
"tags": [
"transformers",
"safetensors",
"text-generation",
"conversational",
"en",
"arxiv:2412.07724",
"arxiv:2502.18573",
"arxiv:2305.11738",
"base_model:ibm-granite/granite-guardian-3.2-5b",
"base_model:finetune:ibm-granite/granite-guardian-3.2-5b",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-guardian-3.2-5b-lora-factuality-correction"
},
{
"name": "granite-guardian-3.2-5b-lora-harm-categories",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 0,
"mod": "2025-09-02",
"tags": [
"transformers",
"safetensors",
"text-generation",
"en",
"arxiv:2412.07724",
"base_model:ibm-granite/granite-guardian-3.2-5b",
"base_model:finetune:ibm-granite/granite-guardian-3.2-5b",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-guardian-3.2-5b-lora-harm-categories"
},
{
"name": "granite-guardian-3.2-5b-lora-harm-correction",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": null,
"layers": 28,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 12800,
"vocab": 49155,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 139,
"mod": "2025-09-08",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"conversational",
"en",
"arxiv:2412.07724",
"arxiv:2508.09937",
"arxiv:2502.03397",
"base_model:ibm-granite/granite-guardian-3.2-5b",
"base_model:finetune:ibm-granite/granite-guardian-3.2-5b",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-guardian-3.2-5b-lora-harm-correction",
"impl": {
"rope_theta": 10000000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.1,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 16.0,
"pad_token_id": 0,
"rope": {
"theta": 10000000.0
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-guardian-3.2-8b-factuality-detection",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 8170848256,
"layers": 40,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 12800,
"vocab": 49155,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 474,
"mod": "2026-03-03",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"conversational",
"en",
"arxiv:2412.07724",
"arxiv:2502.18573",
"base_model:ibm-granite/granite-3.2-8b-instruct",
"base_model:finetune:ibm-granite/granite-3.2-8b-instruct",
"license:apache-2.0",
"endpoints_compatible",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-guardian-3.2-8b-factuality-detection",
"impl": {
"rope_theta": 10000000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.1,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 16.0,
"pad_token_id": 0,
"rope": {
"theta": 10000000.0
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-guardian-3.3-8b",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 8170864640,
"layers": 40,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 12800,
"vocab": 49159,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 204311,
"mod": "2025-09-09",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"conversational",
"en",
"arxiv:2412.07724",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-guardian-3.3-8b",
"impl": {
"rope_theta": 10000000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 16.0,
"pad_token_id": 0,
"rope": {
"theta": 10000000.0
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-guardian-4.0-3b-toxicity-ja",
"arch": "GraniteMoeHybridForCausalLM",
"mt": "granitemoehybrid",
"params": 3402836480,
"layers": 40,
"hidden": 2560,
"heads": 40,
"kv": 8,
"inter": 8192,
"vocab": 100352,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": [
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention"
],
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 437,
"mod": "2026-04-14",
"tags": [
"transformers",
"safetensors",
"granitemoehybrid",
"text-generation",
"granite",
"guardian",
"safety",
"conversational",
"ja",
"arxiv:2412.07724",
"base_model:ibm-granite/granite-4.0-micro",
"base_model:finetune:ibm-granite/granite-4.0-micro",
"license:apache-2.0",
"endpoints_compatible",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-guardian-4.0-3b-toxicity-ja",
"impl": {
"rope_theta": 10000000,
"rms_norm_eps": 1e-05,
"position_embedding_type": "rope",
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": true,
"embedding_multiplier": 12,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 10,
"num_local_experts": 0,
"num_experts_per_tok": 0,
"shared_intermediate_size": 8192,
"output_router_logits": false,
"mamba_d_state": 256,
"mamba_d_conv": 4,
"mamba_expand": 2,
"mamba_n_heads": 128,
"mamba_n_groups": 1,
"mamba_chunk_size": 256,
"mamba_d_head": 40,
"mamba_conv_bias": true,
"mamba_proj_bias": false,
"pad_token_id": 100256,
"rope": {
"theta": 10000000
},
"class": "GraniteMoeHybridForCausalLM",
"model_type": "granitemoehybrid"
}
},
{
"name": "granite-guardian-4.1-8b",
"arch": "GraniteForCausalLM",
"mt": "granite",
"params": 8380551168,
"layers": 40,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 12800,
"vocab": 100352,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 11448,
"mod": "2026-08-27",
"tags": [
"transformers",
"safetensors",
"granite",
"text-generation",
"guardian",
"safety",
"hallucination",
"conversational",
"en",
"arxiv:2412.07724",
"base_model:ibm-granite/granite-4.1-8b",
"base_model:finetune:ibm-granite/granite-4.1-8b",
"license:apache-2.0",
"endpoints_compatible",
"deploy:azure",
"region:us"
],
"id": "ibm-granite/granite-guardian-4.1-8b",
"impl": {
"rope_theta": 10000000,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 16.0,
"pad_token_id": 100256,
"rope": {
"theta": 10000000
},
"class": "GraniteForCausalLM",
"model_type": "granite"
}
},
{
"name": "granite-guardian-hap-125m",
"arch": "RobertaForSequenceClassification",
"mt": "roberta",
"params": 124647170,
"layers": 12,
"hidden": 768,
"heads": 12,
"kv": null,
"inter": 3072,
"vocab": 50265,
"ctx": 514,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-classification",
"dl": 1956,
"mod": "2024-12-19",
"tags": [
"transformers",
"pytorch",
"safetensors",
"roberta",
"text-classification",
"en",
"license:apache-2.0",
"text-embeddings-inference",
"endpoints_compatible",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-guardian-hap-125m",
"impl": {
"layer_norm_eps": 1e-05,
"position_embedding_type": "absolute",
"max_position_embeddings": 514,
"hidden_act": "gelu",
"type_vocab_size": 1,
"pad_token_id": 1,
"class": "RobertaForSequenceClassification",
"model_type": "roberta"
}
},
{
"name": "granite-guardian-hap-38m",
"arch": "RobertaForSequenceClassification",
"mt": "roberta",
"params": 38455682,
"layers": 4,
"hidden": 576,
"heads": 12,
"kv": null,
"inter": 768,
"vocab": 50265,
"ctx": 514,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-classification",
"dl": 1113,
"mod": "2024-12-19",
"tags": [
"transformers",
"pytorch",
"safetensors",
"roberta",
"text-classification",
"en",
"license:apache-2.0",
"text-embeddings-inference",
"endpoints_compatible",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-guardian-hap-38m",
"impl": {
"layer_norm_eps": 1e-12,
"position_embedding_type": "absolute",
"max_position_embeddings": 514,
"hidden_act": "gelu",
"type_vocab_size": 2,
"pad_token_id": 1,
"class": "RobertaForSequenceClassification",
"model_type": "roberta"
}
},
{
"name": "granite-rag-3.0-8b-lora",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 0,
"mod": "2024-11-22",
"tags": [
"transformers",
"safetensors",
"text-generation",
"en",
"arxiv:2403.06009",
"arxiv:2407.11005",
"arxiv:2309.15217",
"arxiv:2204.04991",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-rag-3.0-8b-lora"
},
{
"name": "granite-speech-3.2-8b",
"arch": "GraniteSpeechForConditionalGeneration",
"mt": "granite",
"params": 8483498282,
"layers": 40,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 12800,
"vocab": 49156,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "automatic-speech-recognition",
"dl": 55988,
"mod": "2025-04-16",
"tags": [
"transformers",
"safetensors",
"granite_speech",
"automatic-speech-recognition",
"custom_code",
"en",
"base_model:ibm-granite/granite-3.2-8b-instruct",
"base_model:finetune:ibm-granite/granite-3.2-8b-instruct",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-speech-3.2-8b",
"impl": {
"class": "GraniteSpeechForConditionalGeneration",
"model_type": "granite_speech",
"text": {
"rope_theta": 10000000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 16.0,
"pad_token_id": 0,
"num_hidden_layers": 40,
"hidden_size": 4096,
"num_attention_heads": 32,
"num_key_value_heads": 8,
"intermediate_size": 12800,
"vocab_size": 49156,
"model_type": "granite",
"rope": {
"theta": 10000000.0
}
},
"encoder": {
"context_size": 200,
"num_layers": 10,
"hidden_dim": 1024,
"num_heads": 8,
"dim_head": 128,
"feedforward_mult": 4,
"conv_kernel_size": 15,
"output_dim": 42,
"model_type": "granite_speech_encoder"
},
"projector": {
"layer_norm_eps": 1e-12,
"position_embedding_type": "absolute",
"max_position_embeddings": 2048,
"hidden_act": "gelu",
"downsample_rate": 5,
"window_size": 15,
"num_hidden_layers": 2,
"hidden_size": 1024,
"num_attention_heads": 16,
"intermediate_size": 4096,
"model_type": "granite_speech_qformer"
}
}
},
{
"name": "granite-speech-3.3-2b",
"arch": "GraniteSpeechForConditionalGeneration",
"mt": "granite",
"params": 3009285376,
"layers": 40,
"hidden": 2048,
"heads": 32,
"kv": 8,
"inter": 8192,
"vocab": 49160,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "automatic-speech-recognition",
"dl": 70847,
"mod": "2026-04-07",
"tags": [
"transformers",
"safetensors",
"granite_speech",
"automatic-speech-recognition",
"multilingual",
"en",
"fr",
"de",
"es",
"pt",
"arxiv:2505.08699",
"base_model:ibm-granite/granite-3.3-2b-instruct",
"base_model:finetune:ibm-granite/granite-3.3-2b-instruct",
"license:apache-2.0",
"eval-results",
"endpoints_compatible",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-speech-3.3-2b",
"impl": {
"downsample_rate": 5,
"window_size": 15,
"class": "GraniteSpeechForConditionalGeneration",
"model_type": "granite_speech",
"text": {
"rope_theta": 10000000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 8.0,
"pad_token_id": 0,
"num_hidden_layers": 40,
"hidden_size": 2048,
"num_attention_heads": 32,
"num_key_value_heads": 8,
"intermediate_size": 8192,
"vocab_size": 49160,
"model_type": "granite",
"rope": {
"theta": 10000000.0
}
},
"encoder": {
"context_size": 200,
"num_layers": 16,
"hidden_dim": 1024,
"num_heads": 8,
"dim_head": 128,
"feedforward_mult": 4,
"conv_kernel_size": 15,
"output_dim": 256,
"model_type": "granite_speech_encoder"
},
"projector": {
"layer_norm_eps": 1e-12,
"position_embedding_type": "absolute",
"max_position_embeddings": 2048,
"hidden_act": "gelu",
"num_hidden_layers": 2,
"hidden_size": 1024,
"num_attention_heads": 16,
"intermediate_size": 4096,
"vocab_size": 30522,
"model_type": "blip_2_qformer"
}
}
},
{
"name": "granite-speech-3.3-8b",
"arch": "GraniteSpeechForConditionalGeneration",
"mt": "granite",
"params": 8648711424,
"layers": 40,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 12800,
"vocab": 49160,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "automatic-speech-recognition",
"dl": 42812,
"mod": "2026-04-02",
"tags": [
"transformers",
"safetensors",
"granite_speech",
"automatic-speech-recognition",
"multilingual",
"en",
"fr",
"de",
"es",
"pt",
"arxiv:2505.08699",
"base_model:ibm-granite/granite-3.3-8b-instruct",
"base_model:finetune:ibm-granite/granite-3.3-8b-instruct",
"license:apache-2.0",
"eval-results",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-speech-3.3-8b",
"impl": {
"downsample_rate": 5,
"window_size": 15,
"class": "GraniteSpeechForConditionalGeneration",
"model_type": "granite_speech",
"text": {
"rope_theta": 10000000.0,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 16.0,
"pad_token_id": 0,
"num_hidden_layers": 40,
"hidden_size": 4096,
"num_attention_heads": 32,
"num_key_value_heads": 8,
"intermediate_size": 12800,
"vocab_size": 49160,
"model_type": "granite",
"rope": {
"theta": 10000000.0
}
},
"encoder": {
"context_size": 200,
"num_layers": 16,
"hidden_dim": 1024,
"num_heads": 8,
"dim_head": 128,
"feedforward_mult": 4,
"conv_kernel_size": 15,
"output_dim": 256,
"model_type": "granite_speech_encoder"
},
"projector": {
"layer_norm_eps": 1e-12,
"position_embedding_type": "absolute",
"max_position_embeddings": 2048,
"hidden_act": "gelu",
"num_hidden_layers": 2,
"hidden_size": 1024,
"num_attention_heads": 16,
"intermediate_size": 4096,
"vocab_size": 30522,
"model_type": "blip_2_qformer"
}
}
},
{
"name": "granite-speech-4.1-2b",
"arch": "GraniteSpeechForConditionalGeneration",
"mt": "granite",
"params": 2313207132,
"layers": 40,
"hidden": 2048,
"heads": 16,
"kv": 4,
"inter": 4096,
"vocab": 100353,
"ctx": 4096,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "automatic-speech-recognition",
"dl": 235979,
"mod": "2026-06-12",
"tags": [
"transformers",
"safetensors",
"granite_speech",
"automatic-speech-recognition",
"multilingual",
"en",
"fr",
"de",
"es",
"pt",
"ja",
"arxiv:2310.02943",
"arxiv:2505.08699",
"arxiv:2603.11243",
"arxiv:2604.12398",
"arxiv:2604.22817",
"arxiv:2604.11269",
"arxiv:2603.08397",
"base_model:ibm-granite/granite-4.0-1b-base",
"base_model:finetune:ibm-granite/granite-4.0-1b-base",
"license:apache-2.0",
"eval-results",
"endpoints_compatible",
"region:us",
"deploy:azure",
"deploy:sagemaker"
],
"id": "ibm-granite/granite-speech-4.1-2b",
"impl": {
"tie_word_embeddings": false,
"downsample_rate": 5,
"window_size": 15,
"class": "GraniteSpeechForConditionalGeneration",
"model_type": "granite_speech",
"text": {
"rope_theta": 10000.0,
"rope_parameters": {
"rope_theta": 10000,
"rope_type": "default"
},
"rms_norm_eps": 1e-05,
"max_position_embeddings": 4096,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"mlp_bias": false,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 8.0,
"pad_token_id": 100256,
"num_hidden_layers": 40,
"hidden_size": 2048,
"num_attention_heads": 16,
"num_key_value_heads": 4,
"intermediate_size": 4096,
"vocab_size": 100353,
"model_type": "granite",
"rope": {
"from": "rope_parameters",
"theta": 10000.0,
"rope_type": "default"
}
},
"encoder": {
"context_size": 200,
"num_layers": 16,
"hidden_dim": 1024,
"num_heads": 8,
"dim_head": 128,
"feedforward_mult": 4,
"conv_kernel_size": 15,
"output_dim": 348,
"model_type": "granite_speech_encoder"
},
"projector": {
"layer_norm_eps": 1e-12,
"position_embedding_type": "absolute",
"max_position_embeddings": 2048,
"hidden_act": "gelu",
"num_hidden_layers": 2,
"hidden_size": 1024,
"num_attention_heads": 16,
"intermediate_size": 4096,
"vocab_size": 30522,
"model_type": "blip_2_qformer"
}
}
},
{
"name": "granite-speech-4.1-2b-nar",
"arch": "GraniteSpeechNarForASR",
"mt": "granite",
"params": 2254635356,
"layers": 40,
"hidden": 2048,
"heads": 16,
"kv": 4,
"inter": 4096,
"vocab": 100352,
"ctx": 4096,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "feature-extraction",
"dl": 83030,
"mod": "2026-06-18",
"tags": [
"transformers",
"safetensors",
"granite_speech_nar",
"feature-extraction",
"speech",
"asr",
"non-autoregressive",
"ctc",
"speech_recognition",
"automatic_speech_recognition",
"custom_code",
"en",
"fr",
"de",
"es",
"pt",
"arxiv:2603.08397",
"arxiv:2505.08699",
"arxiv:2603.11243",
"arxiv:2604.12398",
"arxiv:2604.22817",
"arxiv:2604.11269",
"base_model:ibm-granite/granite-4.0-1b-base",
"base_model:finetune:ibm-granite/granite-4.0-1b-base",
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/granite-speech-4.1-2b-nar",
"impl": {
"tie_word_embeddings": true,
"encoder_layer_indices": [
4,
8,
12,
-1
],
"blank_token_id": 100257,
"class": "GraniteSpeechNarForASR",
"model_type": "granite_speech_nar",
"text": {
"rope_parameters": {
"rope_theta": 10000,
"rope_type": "default"
},
"rms_norm_eps": 1e-05,
"max_position_embeddings": 4096,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 8,
"pad_token_id": 100256,
"num_hidden_layers": 40,
"hidden_size": 2048,
"num_attention_heads": 16,
"num_key_value_heads": 4,
"intermediate_size": 4096,
"vocab_size": 100352,
"model_type": "granite",
"rope": {
"from": "rope_parameters",
"theta": 10000,
"rope_type": "default"
}
},
"encoder": {
"bpe_pooling_window": 4,
"bpe_output_dim": 100352,
"blank_token_id": 100257,
"context_size": 200,
"num_layers": 16,
"hidden_dim": 1024,
"num_heads": 8,
"dim_head": 128,
"feedforward_mult": 4,
"conv_kernel_size": 15,
"output_dim": 348,
"model_type": "granite_speech_nar_encoder"
},
"projector": {
"mlp_bias": true,
"downsample_rate": 5,
"hidden_size": 2048,
"num_layers": 2,
"num_heads": 32,
"model_type": "granite_speech_nar_projector"
}
}
},
{
"name": "granite-speech-4.1-2b-plus",
"arch": "GraniteSpeechPlusForConditionalGeneration",
"mt": "granite",
"params": 2111812956,
"layers": 40,
"hidden": 2048,
"heads": 16,
"kv": 4,
"inter": 4096,
"vocab": 100353,
"ctx": 4096,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "automatic-speech-recognition",
"dl": 131500,
"mod": "2026-06-16",
"tags": [
"transformers",
"safetensors",
"granite_speech_plus",
"automatic-speech-recognition",
"multilingual",
"en",
"fr",
"de",
"es",
"pt",
"arxiv:2604.11269",
"arxiv:2604.22817",
"arxiv:2604.12398",
"arxiv:2505.08699",
"arxiv:2603.11243",
"arxiv:2603.08397",
"base_model:ibm-granite/granite-4.0-1b-base",
"base_model:finetune:ibm-granite/granite-4.0-1b-base",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-speech-4.1-2b-plus",
"impl": {
"downsample_rate": 5,
"window_size": 15,
"class": "GraniteSpeechPlusForConditionalGeneration",
"model_type": "granite_speech_plus",
"text": {
"rope_theta": 10000,
"rope_parameters": {
"rope_theta": 10000,
"rope_type": "default"
},
"rms_norm_eps": 1e-05,
"max_position_embeddings": 4096,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 8,
"pad_token_id": 100256,
"num_hidden_layers": 40,
"hidden_size": 2048,
"num_attention_heads": 16,
"num_key_value_heads": 4,
"intermediate_size": 4096,
"vocab_size": 100353,
"model_type": "granite",
"rope": {
"from": "rope_parameters",
"theta": 10000,
"rope_type": "default"
}
},
"encoder": {
"cat_hidden_layers": [
3
],
"context_size": 200,
"num_layers": 16,
"hidden_dim": 1024,
"num_heads": 8,
"dim_head": 128,
"feedforward_mult": 4,
"conv_kernel_size": 15,
"output_dim": 348,
"model_type": "granite_speech_plus_encoder"
},
"projector": {
"layer_norm_eps": 1e-12,
"position_embedding_type": "absolute",
"max_position_embeddings": 2048,
"hidden_act": "gelu",
"pad_token_id": 0,
"num_hidden_layers": 2,
"hidden_size": 1024,
"num_attention_heads": 16,
"intermediate_size": 4096,
"vocab_size": 30522,
"model_type": "blip_2_qformer"
}
}
},
{
"name": "granite-speech-5.0-470m-turboctc",
"arch": "GraniteSpeech5ForCTC",
"mt": "granite_speech5_ctc",
"params": 472993792,
"layers": 16,
"hidden": 1024,
"heads": 8,
"kv": 8,
"inter": 4096,
"vocab": 16384,
"ctx": 512,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "automatic-speech-recognition",
"dl": 43587,
"mod": "2026-09-18",
"tags": [
"transformers",
"safetensors",
"granite_speech5_ctc",
"automatic-speech-recognition",
"en",
"arxiv:2609.20104",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-speech-5.0-470m-turboctc",
"from": "encoder_config",
"impl": {
"tie_word_embeddings": true,
"pad_token_id": 0,
"class": "GraniteSpeech5ForCTC",
"model_type": "granite_speech5_ctc",
"encoder": {
"max_position_embeddings": 512,
"hidden_act": "silu",
"head_dim": 128,
"attention_bias": true,
"attention_dropout": 0.0,
"context_size": 128,
"num_mel_bins": 80,
"subsample_layers": [
0,
1
],
"num_hidden_layers": 16,
"hidden_size": 1024,
"num_attention_heads": 8,
"num_key_value_heads": 8,
"intermediate_size": 4096,
"vocab_size": 16384,
"conv_kernel_size": 7,
"model_type": "granite_speech5_encoder"
}
}
},
{
"name": "granite-speech-5.0-470m-turboctc-nc",
"arch": "GraniteSpeech5ForCTC",
"mt": "granite_speech5_ctc",
"params": 472993792,
"layers": 16,
"hidden": 1024,
"heads": 8,
"kv": 8,
"inter": 4096,
"vocab": 16384,
"ctx": 512,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "cc-by-nc-sa-4.0",
"pipe": "automatic-speech-recognition",
"dl": 2388,
"mod": "2026-08-26",
"tags": [
"transformers",
"safetensors",
"granite_speech5_ctc",
"automatic-speech-recognition",
"en",
"license:cc-by-nc-sa-4.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-speech-5.0-470m-turboctc-nc",
"from": "encoder_config",
"impl": {
"tie_word_embeddings": true,
"pad_token_id": 0,
"class": "GraniteSpeech5ForCTC",
"model_type": "granite_speech5_ctc",
"encoder": {
"max_position_embeddings": 512,
"hidden_act": "silu",
"head_dim": 128,
"attention_bias": true,
"attention_dropout": 0.0,
"context_size": 128,
"num_mel_bins": 80,
"subsample_layers": [
0,
1
],
"num_hidden_layers": 16,
"hidden_size": 1024,
"num_attention_heads": 8,
"num_key_value_heads": 8,
"intermediate_size": 4096,
"vocab_size": 16384,
"conv_kernel_size": 7,
"model_type": "granite_speech5_encoder"
}
}
},
{
"name": "granite-swash-2b",
"arch": "GraniteSWAForCausalLM",
"mt": "granite_swa",
"params": 2144463840,
"layers": 24,
"hidden": 2560,
"heads": 20,
"kv": 4,
"inter": 8192,
"vocab": 100352,
"ctx": 8192,
"experts": null,
"topk": null,
"window": 128,
"ltypes": [
"full_attention",
"sliding_attention",
"sliding_attention",
"full_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"full_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"full_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"full_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"full_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"full_attention"
],
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 20236,
"mod": "2026-07-07",
"tags": [
"transformers",
"safetensors",
"granite_swa",
"text-generation",
"language",
"granite-next",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-swash-2b",
"impl": {
"rope_theta": 10000,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 8192,
"hidden_act": "silu",
"attention_bias": false,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.28,
"logits_scaling": 10,
"sliding_window": 128,
"pad_token_id": 100256,
"rope": {
"theta": 10000
},
"class": "GraniteSWAForCausalLM",
"model_type": "granite_swa"
}
},
{
"name": "granite-swash-3b-a600m",
"arch": "GraniteMoeSWAForCausalLM",
"mt": "granitemoe_swa",
"params": 3020382000,
"layers": 28,
"hidden": 1280,
"heads": 20,
"kv": 4,
"inter": 512,
"vocab": 100352,
"ctx": 8192,
"experts": 48,
"topk": 4,
"window": 128,
"ltypes": [
"full_attention",
"sliding_attention",
"sliding_attention",
"full_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"full_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"full_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"full_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"full_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"full_attention",
"sliding_attention",
"sliding_attention",
"sliding_attention",
"full_attention"
],
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 18507,
"mod": "2026-07-09",
"tags": [
"transformers",
"safetensors",
"granitemoe_swa",
"text-generation",
"language",
"granite-next",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-swash-3b-a600m",
"impl": {
"rope_parameters": {
"rope_theta": 10000,
"rope_type": "default"
},
"rms_norm_eps": 1e-05,
"max_position_embeddings": 8192,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.26,
"logits_scaling": 5,
"num_local_experts": 48,
"num_experts_per_tok": 4,
"shared_intermediate_size": 1280,
"output_router_logits": false,
"sliding_window": 128,
"pad_token_id": 100256,
"rope": {
"from": "rope_parameters",
"theta": 10000,
"rope_type": "default"
},
"class": "GraniteMoeSWAForCausalLM",
"model_type": "granitemoe_swa"
}
},
{
"name": "granite-switch-4.1-30b-preview",
"arch": "GraniteSwitchForCausalLM",
"mt": "granite_switch",
"params": 32237998080,
"layers": 65,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 32768,
"vocab": 100364,
"ctx": 131072,
"experts": null,
"topk": 2,
"window": null,
"ltypes": [
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention"
],
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 450,
"mod": "2026-05-24",
"tags": [
"transformers",
"safetensors",
"granite_switch",
"text-generation",
"language",
"granite-switch",
"granite-4.1",
"conversational",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-switch-4.1-30b-preview",
"impl": {
"rope_parameters": {
"rope_theta": 50000000,
"rope_type": "default"
},
"rms_norm_eps": 1e-05,
"position_embedding_type": "rope",
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.175,
"logits_scaling": 16.0,
"num_local_experts": 0,
"num_experts_per_tok": 2,
"shared_intermediate_size": 32768,
"output_router_logits": false,
"mamba_d_state": 256,
"mamba_d_conv": 4,
"mamba_expand": 2,
"mamba_n_heads": 128,
"mamba_n_groups": 1,
"mamba_chunk_size": 256,
"mamba_d_head": 64,
"mamba_conv_bias": true,
"mamba_proj_bias": false,
"num_adapters": 12,
"max_lora_rank": 32,
"adapter_ranks": [
16,
32,
32,
16,
16,
32,
16,
32,
16,
32,
16,
32
],
"lora_target_modules": [
"o_proj",
"qkv_proj",
"shared_input_linear",
"shared_output_linear"
],
"control_token_gain": 15.0,
"switch_head_dim": 32,
"projection_head_dim": 128,
"pad_token_id": 100256,
"rope": {
"from": "rope_parameters",
"theta": 50000000,
"rope_type": "default"
},
"class": "GraniteSwitchForCausalLM",
"model_type": "granite_switch"
}
},
{
"name": "granite-switch-4.1-3b-preview",
"arch": "GraniteSwitchForCausalLM",
"mt": "granite_switch",
"params": 4149977600,
"layers": 41,
"hidden": 2560,
"heads": 40,
"kv": 8,
"inter": 8192,
"vocab": 100364,
"ctx": 131072,
"experts": null,
"topk": 2,
"window": null,
"ltypes": [
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention"
],
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 1357,
"mod": "2026-05-24",
"tags": [
"transformers",
"safetensors",
"granite_switch",
"text-generation",
"language",
"granite-switch",
"granite-4.1",
"conversational",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-switch-4.1-3b-preview",
"impl": {
"rope_parameters": {
"rope_theta": 10000000,
"rope_type": "default"
},
"rms_norm_eps": 1e-05,
"position_embedding_type": "rope",
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 10.0,
"num_local_experts": 0,
"num_experts_per_tok": 2,
"shared_intermediate_size": 8192,
"output_router_logits": false,
"mamba_d_state": 256,
"mamba_d_conv": 4,
"mamba_expand": 2,
"mamba_n_heads": 128,
"mamba_n_groups": 1,
"mamba_chunk_size": 256,
"mamba_d_head": 40,
"mamba_conv_bias": true,
"mamba_proj_bias": false,
"num_adapters": 12,
"max_lora_rank": 32,
"adapter_ranks": [
16,
32,
32,
16,
16,
32,
16,
32,
16,
32,
16,
16
],
"lora_target_modules": [
"o_proj",
"qkv_proj",
"shared_input_linear",
"shared_output_linear"
],
"control_token_gain": 15.0,
"switch_head_dim": 32,
"projection_head_dim": 64,
"pad_token_id": 100256,
"rope": {
"from": "rope_parameters",
"theta": 10000000,
"rope_type": "default"
},
"class": "GraniteSwitchForCausalLM",
"model_type": "granite_switch"
}
},
{
"name": "granite-switch-4.1-8b-preview",
"arch": "GraniteSwitchForCausalLM",
"mt": "granite_switch",
"params": 9568112640,
"layers": 41,
"hidden": 4096,
"heads": 32,
"kv": 8,
"inter": 12800,
"vocab": 100364,
"ctx": 131072,
"experts": null,
"topk": 2,
"window": null,
"ltypes": [
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention",
"attention"
],
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 495,
"mod": "2026-05-24",
"tags": [
"transformers",
"safetensors",
"granite_switch",
"text-generation",
"language",
"granite-switch",
"granite-4.1",
"conversational",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-switch-4.1-8b-preview",
"impl": {
"rope_parameters": {
"rope_theta": 10000000,
"rope_type": "default"
},
"rms_norm_eps": 1e-05,
"position_embedding_type": "rope",
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.0078125,
"residual_multiplier": 0.22,
"logits_scaling": 16.0,
"num_local_experts": 0,
"num_experts_per_tok": 2,
"shared_intermediate_size": 12800,
"output_router_logits": false,
"mamba_d_state": 256,
"mamba_d_conv": 4,
"mamba_expand": 2,
"mamba_n_heads": 128,
"mamba_n_groups": 1,
"mamba_chunk_size": 256,
"mamba_d_head": 64,
"mamba_conv_bias": true,
"mamba_proj_bias": false,
"num_adapters": 12,
"max_lora_rank": 32,
"adapter_ranks": [
16,
32,
32,
16,
16,
32,
16,
32,
16,
32,
16,
16
],
"lora_target_modules": [
"o_proj",
"qkv_proj",
"shared_input_linear",
"shared_output_linear"
],
"control_token_gain": 15.0,
"switch_head_dim": 32,
"projection_head_dim": 128,
"pad_token_id": 100256,
"rope": {
"from": "rope_parameters",
"theta": 10000000,
"rope_type": "default"
},
"class": "GraniteSwitchForCausalLM",
"model_type": "granite_switch"
}
},
{
"name": "granite-timeseries-flowstate-r1",
"arch": "FlowStateModel",
"mt": "flowstate",
"params": 9069312,
"layers": 6,
"hidden": 512,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "time-series-forecasting",
"dl": 27202,
"mod": "2026-04-29",
"tags": [
"safetensors",
"flowstate",
"time series",
"forecasting",
"pretrained models",
"foundation models",
"time series foundation models",
"time-series",
"state space model",
"ssm",
"time-series-forecasting",
"arxiv:2508.05287",
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/granite-timeseries-flowstate-r1",
"ts": {
"context window": 2048,
"forecast horizon": 24,
"state width": 512,
"decoder width": 256,
"quantile levels": 9
},
"impl": {
"context_length": 2048,
"class": "FlowStateModel",
"model_type": "flowstate"
}
},
{
"name": "granite-timeseries-patchtsmixer",
"arch": "PatchTSMixerForPrediction",
"mt": "patchtsmixer",
"params": 196144,
"layers": 2,
"hidden": 48,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "time-series-forecasting",
"dl": 5518,
"mod": "2024-08-01",
"tags": [
"transformers",
"pytorch",
"safetensors",
"patchtsmixer",
"time series",
"forecasting",
"pretrained models",
"foundation models",
"time series foundation models",
"time-series",
"time-series-forecasting",
"arxiv:2306.09364",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-timeseries-patchtsmixer",
"ts": {
"context window": 512,
"forecast horizon": 96,
"patch length": 16,
"patch stride": 16,
"patches": 32,
"feed-forward expansion": 3,
"input channels": 7
},
"impl": {
"context_length": 512,
"prediction_length": 96,
"patch_length": 16,
"patch_stride": 16,
"num_patches": 32,
"d_model": 48,
"expansion_factor": 3,
"num_input_channels": 7,
"loss": "mse",
"scaling": true,
"head_dropout": 0.5,
"class": "PatchTSMixerForPrediction",
"model_type": "patchtsmixer"
}
},
{
"name": "granite-timeseries-patchtst",
"arch": "PatchTSTForPrediction",
"mt": "patchtst",
"params": 616032,
"layers": 3,
"hidden": 128,
"heads": 16,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "time-series-forecasting",
"dl": 18071,
"mod": "2024-08-01",
"tags": [
"transformers",
"safetensors",
"patchtst",
"generated_from_trainer",
"time series",
"forecasting",
"pretrained models",
"foundation models",
"time series foundation models",
"time-series",
"time-series-forecasting",
"arxiv:2211.14730",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-timeseries-patchtst",
"ts": {
"context window": 512,
"forecast horizon": 96,
"patch length": 12,
"patch stride": 12,
"input channels": 7
},
"impl": {
"activation_function": "gelu",
"attention_dropout": 0.0,
"context_length": 512,
"prediction_length": 96,
"patch_length": 12,
"patch_stride": 12,
"d_model": 128,
"num_input_channels": 7,
"loss": "mse",
"scaling": "std",
"head_dropout": 0.2,
"class": "PatchTSTForPrediction",
"model_type": "patchtst"
}
},
{
"name": "granite-timeseries-patchtst-fm-r1",
"arch": "PatchTSTFMForPrediction",
"mt": "patchtst_fm",
"params": 257895552,
"layers": 20,
"hidden": 1024,
"heads": 16,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "time-series-forecasting",
"dl": 40173,
"mod": "2026-03-25",
"tags": [
"safetensors",
"patchtst_fm",
"time series",
"forecasting",
"pretrained models",
"foundation models",
"time series foundation models",
"time-series",
"time-series-forecasting",
"arxiv:2211.14730",
"arxiv:2602.06909",
"arxiv:2403.07815",
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/granite-timeseries-patchtst-fm-r1",
"ts": {
"context window": 8192,
"forecast horizon": 64,
"patch length": 16,
"patches": 512,
"quantile levels": 99
},
"impl": {
"context_length": 8192,
"prediction_length": 64,
"d_model": 1024,
"n_patch": 512,
"d_patch": 16,
"norm_first": true,
"class": "PatchTSTFMForPrediction",
"model_type": "patchtst_fm"
}
},
{
"name": "granite-timeseries-patchtst-fm-r2",
"arch": "PatchTSTFMForPrediction",
"mt": "patchtst_fm",
"params": 384600208,
"layers": 30,
"hidden": 1024,
"heads": 16,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "openmdw-1.0",
"pipe": "time-series-forecasting",
"dl": 74821,
"mod": "2026-09-09",
"tags": [
"safetensors",
"patchtst_fm",
"time series",
"forecasting",
"pretrained models",
"foundation models",
"time series foundation models",
"time-series",
"time-series-forecasting",
"arxiv:2211.14730",
"arxiv:2005.08100",
"arxiv:2508.02879",
"arxiv:2403.07815",
"arxiv:2602.06909",
"license:openmdw-1.0",
"region:us"
],
"id": "ibm-granite/granite-timeseries-patchtst-fm-r2",
"ts": {
"context window": 8192,
"forecast horizon": 64,
"patch length": 16,
"patch stride": 8,
"patches": 1023,
"quantile levels": 99
},
"impl": {
"context_length": 8192,
"prediction_length": 64,
"patch_stride": 8,
"d_model": 1024,
"block_type": "conformer",
"n_patch": 1023,
"d_patch": 16,
"norm_first": true,
"prehead_norm": true,
"patch_loss_windowing": "hamming",
"class": "PatchTSTFMForPrediction",
"model_type": "patchtst_fm"
}
},
{
"name": "granite-timeseries-tspulse-r1",
"arch": "TSPulseForReconstruction",
"mt": "tspulse",
"params": 1084330,
"layers": 8,
"hidden": 24,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": null,
"dl": 275479,
"mod": "2026-03-07",
"tags": [
"granite-tsfm",
"safetensors",
"tspulse",
"time series",
"pretrained models",
"foundation models",
"time series foundation models",
"time-series",
"arxiv:2505.13033",
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/granite-timeseries-tspulse-r1",
"ts": {
"context window": 512,
"forecast horizon": 16,
"patch length": 8,
"patch stride": 8,
"patches": 128,
"decoder width": 24,
"feed-forward expansion": 2,
"input channels": 1
},
"impl": {
"context_length": 512,
"prediction_length": 16,
"patch_length": 8,
"patch_stride": 8,
"num_patches": 128,
"d_model": 24,
"decoder_d_model": 24,
"expansion_factor": 2,
"num_input_channels": 1,
"loss": "mse",
"scaling": "revin",
"head_dropout": 0.2,
"revin_affine": true,
"fft_remove_component": "last",
"mask_ratio": 0.7,
"class": "TSPulseForReconstruction",
"model_type": "tspulse"
}
},
{
"name": "granite-timeseries-ttm-r1",
"arch": "TinyTimeMixerForPrediction",
"mt": "tinytimemixer",
"params": 805280,
"layers": 2,
"hidden": 192,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "time-series-forecasting",
"dl": 24839,
"mod": "2025-01-30",
"tags": [
"granite-tsfm",
"safetensors",
"tinytimemixer",
"time series",
"forecasting",
"pretrained models",
"foundation models",
"time series foundation models",
"time-series",
"time-series-forecasting",
"arxiv:2401.03955",
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/granite-timeseries-ttm-r1",
"ts": {
"context window": 512,
"forecast horizon": 96,
"patch length": 64,
"patch stride": 64,
"patches": 8,
"decoder width": 128,
"feed-forward expansion": 2,
"input channels": 1
},
"impl": {
"context_length": 512,
"prediction_length": 96,
"patch_length": 64,
"patch_stride": 64,
"num_patches": 8,
"d_model": 192,
"decoder_d_model": 128,
"expansion_factor": 2,
"num_input_channels": 1,
"loss": "mse",
"scaling": "std",
"head_dropout": 0.2,
"class": "TinyTimeMixerForPrediction",
"model_type": "tinytimemixer"
}
},
{
"name": "granite-timeseries-ttm-r2",
"arch": "TinyTimeMixerForPrediction",
"mt": "tinytimemixer",
"params": 805280,
"layers": 2,
"hidden": 192,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "time-series-forecasting",
"dl": 305110,
"mod": "2025-02-26",
"tags": [
"granite-tsfm",
"safetensors",
"tinytimemixer",
"time series",
"forecasting",
"pretrained models",
"foundation models",
"time series foundation models",
"time-series",
"time-series-forecasting",
"arxiv:2401.03955",
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/granite-timeseries-ttm-r2",
"ts": {
"context window": 512,
"forecast horizon": 96,
"patch length": 64,
"patch stride": 64,
"patches": 8,
"decoder width": 128,
"feed-forward expansion": 2,
"input channels": 1
},
"impl": {
"context_length": 512,
"prediction_length": 96,
"patch_length": 64,
"patch_stride": 64,
"num_patches": 8,
"d_model": 192,
"decoder_d_model": 128,
"expansion_factor": 2,
"num_input_channels": 1,
"loss": "mse",
"scaling": "std",
"head_dropout": 0.4,
"class": "TinyTimeMixerForPrediction",
"model_type": "tinytimemixer"
}
},
{
"name": "granite-timeseries-ttm-r3",
"arch": "tinytimemixer",
"mt": "tinytimemixer",
"params": 1414514,
"layers": 10,
"hidden": 128,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "time-series-forecasting",
"dl": 38246,
"mod": "2026-07-02",
"tags": [
"safetensors",
"tinytimemixer",
"time series",
"forecasting",
"pretrained models",
"foundation models",
"time series foundation models",
"time-series",
"granite-ttm-r3",
"time-series-forecasting",
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/granite-timeseries-ttm-r3",
"ts": {
"context window": 512,
"forecast horizon": 30,
"patch length": 9,
"patch stride": 9,
"decoder width": 64,
"feed-forward expansion": 3,
"input channels": 1,
"quantile levels": 9
},
"impl": {
"context_length": 512,
"prediction_length": 30,
"patch_length": 9,
"patch_stride": 9,
"d_model": 128,
"decoder_d_model": 64,
"expansion_factor": 3,
"num_input_channels": 1,
"quantile_levels": [
0.1,
0.2,
0.3,
0.4,
0.5,
0.6,
0.7,
0.8,
0.9
],
"loss": "mae",
"scaling": "std",
"head_dropout": 0.2,
"model_type": "tinytimemixer"
}
},
{
"name": "granite-uncertainty-3.0-8b-lora",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 0,
"mod": "2024-12-07",
"tags": [
"transformers",
"safetensors",
"text-generation",
"en",
"arxiv:2403.08819",
"license:apache-2.0",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-uncertainty-3.0-8b-lora"
},
{
"name": "granite-vision-3.1-2b-preview",
"arch": "LlavaNextForConditionalGeneration",
"mt": "granite",
"params": 2975396928,
"layers": 40,
"hidden": 2048,
"heads": 32,
"kv": 8,
"inter": 8192,
"vocab": 49156,
"ctx": 16384,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "image-text-to-text",
"dl": 1069,
"mod": "2025-06-12",
"tags": [
"transformers",
"safetensors",
"llava_next",
"image-text-to-text",
"conversational",
"en",
"arxiv:2502.09927",
"base_model:ibm-granite/granite-3.1-2b-instruct",
"base_model:finetune:ibm-granite/granite-3.1-2b-instruct",
"license:apache-2.0",
"text-generation-inference",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-vision-3.1-2b-preview",
"impl": {
"tie_word_embeddings": true,
"class": "LlavaNextForConditionalGeneration",
"model_type": "llava_next",
"text": {
"rope_theta": 300000,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 16384,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.1,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 8.0,
"pad_token_id": 0,
"num_hidden_layers": 40,
"hidden_size": 2048,
"num_attention_heads": 32,
"num_key_value_heads": 8,
"intermediate_size": 8192,
"vocab_size": 49156,
"model_type": "granite",
"rope": {
"theta": 300000
}
},
"vision": {
"num_hidden_layers": 27,
"hidden_size": 1152,
"num_attention_heads": 16,
"intermediate_size": 4304,
"model_type": "siglip_vision_model"
}
}
},
{
"name": "granite-vision-3.2-2b",
"arch": "LlavaNextForConditionalGeneration",
"mt": "granite",
"params": 2975396928,
"layers": 40,
"hidden": 2048,
"heads": 32,
"kv": 8,
"inter": 8192,
"vocab": 49156,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "image-text-to-text",
"dl": 1836,
"mod": "2026-04-02",
"tags": [
"transformers",
"safetensors",
"llava_next",
"image-text-to-text",
"conversational",
"en",
"arxiv:2502.09927",
"base_model:ibm-granite/granite-3.1-2b-instruct",
"base_model:finetune:ibm-granite/granite-3.1-2b-instruct",
"license:apache-2.0",
"text-generation-inference",
"endpoints_compatible",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-vision-3.2-2b",
"impl": {
"tie_word_embeddings": true,
"class": "LlavaNextForConditionalGeneration",
"model_type": "llava_next",
"text": {
"rope_theta": 300000,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.1,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 8.0,
"pad_token_id": 0,
"num_hidden_layers": 40,
"hidden_size": 2048,
"num_attention_heads": 32,
"num_key_value_heads": 8,
"intermediate_size": 8192,
"vocab_size": 49156,
"model_type": "granite",
"rope": {
"theta": 300000
}
},
"vision": {
"num_hidden_layers": 27,
"hidden_size": 1152,
"num_attention_heads": 16,
"intermediate_size": 4304,
"model_type": "siglip_vision_model"
}
}
},
{
"name": "granite-vision-3.3-2b",
"arch": "LlavaNextForConditionalGeneration",
"mt": "granite",
"params": 2975396928,
"layers": 40,
"hidden": 2048,
"heads": 32,
"kv": 8,
"inter": 8192,
"vocab": 49156,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "image-to-text",
"dl": 183991,
"mod": "2026-04-02",
"tags": [
"safetensors",
"llava_next",
"image-to-text",
"arxiv:2502.09927",
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/granite-vision-3.3-2b",
"impl": {
"tie_word_embeddings": true,
"class": "LlavaNextForConditionalGeneration",
"model_type": "llava_next",
"text": {
"rope_theta": 300000,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.1,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 8.0,
"pad_token_id": 0,
"num_hidden_layers": 40,
"hidden_size": 2048,
"num_attention_heads": 32,
"num_key_value_heads": 8,
"intermediate_size": 8192,
"vocab_size": 49156,
"model_type": "granite",
"rope": {
"theta": 300000
}
},
"vision": {
"layer_norm_eps": 1e-06,
"hidden_act": "gelu_pytorch_tanh",
"attention_dropout": 0.0,
"num_hidden_layers": 27,
"hidden_size": 1152,
"num_attention_heads": 16,
"intermediate_size": 4304,
"model_type": "siglip_vision_model"
}
}
},
{
"name": "granite-vision-3.3-2b-chart2csv-preview",
"arch": "LlavaNextForConditionalGeneration",
"mt": "granite",
"params": 2975396928,
"layers": 40,
"hidden": 2048,
"heads": 32,
"kv": 8,
"inter": 8192,
"vocab": 49156,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "image-text-to-text",
"dl": 917,
"mod": "2026-02-09",
"tags": [
"transformers",
"safetensors",
"llava_next",
"image-text-to-text",
"generated_from_trainer",
"sft",
"trl",
"conversational",
"arxiv:2507.19492",
"arxiv:2502.09927",
"base_model:ibm-granite/granite-vision-3.3-2b",
"base_model:finetune:ibm-granite/granite-vision-3.3-2b",
"license:apache-2.0",
"text-generation-inference",
"endpoints_compatible",
"region:us"
],
"id": "ibm-granite/granite-vision-3.3-2b-chart2csv-preview",
"impl": {
"tie_word_embeddings": true,
"pad_token_id": 0,
"class": "LlavaNextForConditionalGeneration",
"model_type": "llava_next",
"text": {
"rope_theta": 300000,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.1,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 8.0,
"pad_token_id": 0,
"num_hidden_layers": 40,
"hidden_size": 2048,
"num_attention_heads": 32,
"num_key_value_heads": 8,
"intermediate_size": 8192,
"vocab_size": 49156,
"model_type": "granite",
"rope": {
"theta": 300000
}
},
"vision": {
"layer_norm_eps": 1e-06,
"hidden_act": "gelu_pytorch_tanh",
"attention_dropout": 0.0,
"num_hidden_layers": 27,
"hidden_size": 1152,
"num_attention_heads": 16,
"intermediate_size": 4304,
"model_type": "siglip_vision_model"
}
}
},
{
"name": "granite-vision-3.3-2b-embedding",
"arch": "GraniteVisionEmb",
"mt": "granite",
"params": 2975659200,
"layers": 40,
"hidden": 2048,
"heads": null,
"kv": 8,
"inter": 8192,
"vocab": 49156,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "feature-extraction",
"dl": 633,
"mod": "2025-08-16",
"tags": [
"transformers",
"safetensors",
"granitevisionemb",
"feature-extraction",
"custom_code",
"en",
"arxiv:2407.01449",
"arxiv:2502.09927",
"arxiv:2502.12342",
"arxiv:2505.17166",
"base_model:ibm-granite/granite-vision-3.3-2b",
"base_model:finetune:ibm-granite/granite-vision-3.3-2b",
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/granite-vision-3.3-2b-embedding",
"impl": {
"tie_word_embeddings": true,
"class": "GraniteVisionEmb",
"model_type": "granitevisionemb",
"text": {
"rope_theta": 300000,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"attention_dropout": 0.1,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 8.0,
"pad_token_id": 0,
"num_hidden_layers": 40,
"hidden_size": 2048,
"num_key_value_heads": 8,
"intermediate_size": 8192,
"vocab_size": 49156,
"model_type": "granite",
"rope": {
"theta": 300000
}
},
"vision": {
"layer_norm_eps": 1e-06,
"hidden_act": "gelu_pytorch_tanh",
"num_hidden_layers": 27,
"hidden_size": 1152,
"num_attention_heads": 16,
"intermediate_size": 4304,
"model_type": "siglip_vision_model"
}
}
},
{
"name": "granite-vision-4.1-4b",
"arch": "Granite4VisionForConditionalGeneration",
"mt": "granite",
"params": 3997203904,
"layers": 40,
"hidden": 2560,
"heads": 40,
"kv": 8,
"inter": 8192,
"vocab": 100352,
"ctx": 131072,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "image-text-to-text",
"dl": 151219,
"mod": "2026-07-13",
"tags": [
"transformers",
"safetensors",
"granite4_vision",
"image-text-to-text",
"conversational",
"custom_code",
"en",
"arxiv:2603.27064",
"arxiv:2208.00385",
"arxiv:2502.09927",
"arxiv:2406.04334",
"license:apache-2.0",
"eval-results",
"endpoints_compatible",
"region:us",
"deploy:azure"
],
"id": "ibm-granite/granite-vision-4.1-4b",
"impl": {
"tie_word_embeddings": true,
"downsample_rate": "4/8",
"class": "Granite4VisionForConditionalGeneration",
"model_type": "granite4_vision",
"text": {
"rope_theta": 10000000,
"rms_norm_eps": 1e-05,
"max_position_embeddings": 131072,
"hidden_act": "silu",
"attention_bias": false,
"attention_dropout": 0.0,
"mlp_bias": false,
"tie_word_embeddings": true,
"embedding_multiplier": 12.0,
"attention_multiplier": 0.015625,
"residual_multiplier": 0.22,
"logits_scaling": 10.0,
"pad_token_id": 100256,
"num_hidden_layers": 40,
"hidden_size": 2560,
"num_attention_heads": 40,
"num_key_value_heads": 8,
"intermediate_size": 8192,
"vocab_size": 100352,
"model_type": "granite",
"rope": {
"theta": 10000000
}
},
"vision": {
"layer_norm_eps": 1e-06,
"hidden_act": "gelu_pytorch_tanh",
"attention_dropout": 0.0,
"num_hidden_layers": 27,
"hidden_size": 1152,
"num_attention_heads": 16,
"intermediate_size": 4304,
"model_type": "siglip_vision_model"
}
}
},
{
"name": "granitelib-core-r1.0",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 4649,
"mod": "2026-05-20",
"tags": [
"granite-library",
"safetensors",
"granite",
"uncertainty",
"context-attribution",
"requirement-checking",
"lora",
"peft",
"text-generation",
"en",
"base_model:ibm-granite/granite-4.0-micro",
"base_model:adapter:ibm-granite/granite-4.0-micro",
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/granitelib-core-r1.0"
},
{
"name": "granitelib-guardian-r1.0",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": "text-generation",
"dl": 591,
"mod": "2026-06-11",
"tags": [
"granite-library",
"safetensors",
"granite",
"guardian",
"safety",
"hallucination",
"factuality",
"lora",
"peft",
"text-generation",
"en",
"base_model:ibm-granite/granite-4.0-micro",
"base_model:adapter:ibm-granite/granite-4.0-micro",
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/granitelib-guardian-r1.0"
},
{
"name": "granitelib-rag-gpt-oss-r1.0",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": null,
"dl": 34,
"mod": "2026-05-01",
"tags": [
"granite-library",
"safetensors",
"en",
"base_model:openai/gpt-oss-20b",
"base_model:finetune:openai/gpt-oss-20b",
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/granitelib-rag-gpt-oss-r1.0"
},
{
"name": "granitelib-rag-r1.0",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": null,
"dl": 7806,
"mod": "2026-05-20",
"tags": [
"granite-library",
"safetensors",
"gguf",
"en",
"arxiv:2504.12397",
"base_model:ibm-granite/granite-4.0-micro",
"base_model:quantized:ibm-granite/granite-4.0-micro",
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/granitelib-rag-r1.0"
},
{
"name": "rag-intrinsics-lib",
"arch": null,
"mt": null,
"params": null,
"layers": null,
"hidden": null,
"heads": null,
"kv": null,
"inter": null,
"vocab": null,
"ctx": null,
"experts": null,
"topk": null,
"window": null,
"ltypes": null,
"lic": "apache-2.0",
"pipe": null,
"dl": 0,
"mod": "2026-01-24",
"tags": [
"safetensors",
"license:apache-2.0",
"region:us"
],
"id": "ibm-granite/rag-intrinsics-lib"
}
]