{ "architectures": [ "MistralLarge4ForConditionalGeneration" ], "dtype": "bfloat16", "image_token_index": 10, "model_type": "mistral_large4", "multimodal_projector_bias": false, "projector_hidden_act": "gelu", "quantization_config": { "activation_scheme": "dynamic", "dequantize": false, "expert_dtype": "fp8", "modules_to_convert": null, "modules_to_not_convert": [ "model.vision_tower", "model.multi_modal_projector", "lm_head", "kv_b_proj", "indexer.wk", "indexer.weights_proj", "kv_a_proj_with_mqa" ], "quant_method": "fp8", "scale_fmt": "float", "weight_block_size": [ 32, 32 ] }, "spatial_merge_size": 2, "text_config": { "attention_bias": false, "attention_dropout": 0.0, "bos_token_id": 1, "dtype": "bfloat16", "eos_token_id": 2, "final_logit_softcapping": 30.0, "first_k_dense_replace": 1, "head_dim": 8, "hidden_act": "silu", "hidden_size": 32, "index_head_dim": 16, "index_n_heads": 2, "index_topk": 4, "indexer_types": [ "full", "full", "full", "shared", "shared", "full", "full", "shared" ], "initializer_range": 0.02, "intermediate_size": 32, "kv_lora_rank": 16, "kv_source_layers": [ 0, 1, 2, 3, 4, 5, 6, 3 ], "layer_types": [ "indexed_attention", "indexed_attention", "indexed_attention", "indexed_attention", "indexed_attention", "indexed_attention", "indexed_attention", "indexed_attention" ], "max_position_embeddings": 1048576, "mlp_layer_types": [ "dense", "sparse", "sparse", "sparse", "sparse", "sparse", "sparse", "sparse" ], "model_type": "mistral_large4_text", "moe_intermediate_size": 64, "n_group": 1, "n_routed_experts": 8, "n_shared_experts": 1, "no_rope_layer_interval": 4, "no_rope_layers": [ 1, 1, 1, 0, 1, 1, 1, 0 ], "num_attention_heads": 4, "num_experts_per_tok": 2, "num_hidden_layers": 8, "num_key_value_heads": 4, "output_router_logits": false, "pad_token_id": null, "qk_head_dim": 128, "qk_nope_head_dim": 8, "qk_rope_head_dim": 8, "rms_norm_eps": 1e-05, "rope_parameters": { "attention_factor": 1.0, "beta_fast": 32, "beta_slow": 1, "factor": 2048, "original_max_position_embeddings": 512, "rope_theta": 10000.0, "rope_type": "yarn" }, "routed_scaling_factor": 1.0, "tie_word_embeddings": false, "topk_group": 1, "use_cache": true, "v_head_dim": 8, "vocab_size": 204800 }, "tie_word_embeddings": false, "transformers_version": "5.21.0.dev0", "vision_config": { "attention_dropout": 0.0, "dtype": "bfloat16", "head_dim": 4, "hidden_act": "silu", "hidden_size": 16, "image_size": 1540, "initializer_range": 0.02, "intermediate_size": 32, "model_type": "pixtral", "num_attention_heads": 4, "num_channels": 3, "num_hidden_layers": 2, "patch_size": 14, "rope_parameters": { "rope_theta": 10000.0, "rope_type": "axial" } }, "vision_feature_layer": -1 }