| { |
| "architectures": [ |
| "KimiK3ForConditionalGeneration" |
| ], |
| "auto_map": { |
| "AutoConfig": "configuration_kimi_k3.KimiK3Config", |
| "AutoModel": "modeling_kimi_k3.KimiK3ForConditionalGeneration", |
| "AutoModelForCausalLM": "modeling_kimi_k3.KimiK3ForConditionalGeneration" |
| }, |
| "dtype": "bfloat16", |
| "ignore_index": -100, |
| "media_placeholder_token_id": 163605, |
| "model_type": "kimi_k3", |
| "pad_token_id": 0, |
| "quantization_config": { |
| "config_groups": { |
| "group_0": { |
| "format": "mxfp4-pack-quantized", |
| "input_activations": { |
| "actorder": null, |
| "block_structure": null, |
| "dynamic": true, |
| "group_size": 32, |
| "num_bits": 4, |
| "observer": null, |
| "observer_kwargs": {}, |
| "scale_dtype": "torch.uint8", |
| "strategy": "group", |
| "symmetric": true, |
| "type": "float", |
| "zp_dtype": null |
| }, |
| "output_activations": null, |
| "targets": [ |
| "re:.*block_sparse_moe.*" |
| ], |
| "weights": { |
| "actorder": null, |
| "block_structure": null, |
| "dynamic": false, |
| "group_size": 32, |
| "num_bits": 4, |
| "observer": "memoryless_minmax", |
| "observer_kwargs": {}, |
| "scale_dtype": "torch.uint8", |
| "strategy": "group", |
| "symmetric": true, |
| "type": "float", |
| "zp_dtype": null |
| } |
| } |
| }, |
| "format": "mxfp4-pack-quantized", |
| "global_compression_ratio": null, |
| "ignore": [ |
| "vision_tower.encoder.blocks.0.mlp.fc0", |
| "vision_tower.encoder.blocks.0.mlp.fc1", |
| "vision_tower.encoder.blocks.0.wqkv", |
| "vision_tower.encoder.blocks.0.wo", |
| "vision_tower.encoder.blocks.1.mlp.fc0", |
| "vision_tower.encoder.blocks.1.mlp.fc1", |
| "vision_tower.encoder.blocks.1.wqkv", |
| "vision_tower.encoder.blocks.1.wo", |
| "mm_projector.proj.0", |
| "mm_projector.proj.2", |
| "language_model.model.layers.0.self_attn.q_proj", |
| "language_model.model.layers.0.self_attn.k_proj", |
| "language_model.model.layers.0.self_attn.v_proj", |
| "language_model.model.layers.0.self_attn.f_a_proj", |
| "language_model.model.layers.0.self_attn.f_b_proj", |
| "language_model.model.layers.0.self_attn.b_proj", |
| "language_model.model.layers.0.self_attn.g_proj", |
| "language_model.model.layers.0.self_attn.o_norm", |
| "language_model.model.layers.0.self_attn.o_proj", |
| "language_model.model.layers.0.mlp.gate_proj", |
| "language_model.model.layers.0.mlp.up_proj", |
| "language_model.model.layers.0.mlp.down_proj", |
| "language_model.model.layers.0.self_attention_res_proj", |
| "language_model.model.layers.0.mlp_res_proj", |
| "language_model.model.layers.1.self_attn.q_proj", |
| "language_model.model.layers.1.self_attn.k_proj", |
| "language_model.model.layers.1.self_attn.v_proj", |
| "language_model.model.layers.1.self_attn.f_a_proj", |
| "language_model.model.layers.1.self_attn.f_b_proj", |
| "language_model.model.layers.1.self_attn.b_proj", |
| "language_model.model.layers.1.self_attn.g_proj", |
| "language_model.model.layers.1.self_attn.o_norm", |
| "language_model.model.layers.1.self_attn.o_proj", |
| "language_model.model.layers.1.block_sparse_moe.gate", |
| "language_model.model.layers.1.self_attention_res_proj", |
| "language_model.model.layers.1.mlp_res_proj", |
| "language_model.model.layers.2.self_attn.q_proj", |
| "language_model.model.layers.2.self_attn.k_proj", |
| "language_model.model.layers.2.self_attn.v_proj", |
| "language_model.model.layers.2.self_attn.f_a_proj", |
| "language_model.model.layers.2.self_attn.f_b_proj", |
| "language_model.model.layers.2.self_attn.b_proj", |
| "language_model.model.layers.2.self_attn.g_proj", |
| "language_model.model.layers.2.self_attn.o_norm", |
| "language_model.model.layers.2.self_attn.o_proj", |
| "language_model.model.layers.2.block_sparse_moe.gate", |
| "language_model.model.layers.2.self_attention_res_proj", |
| "language_model.model.layers.2.mlp_res_proj", |
| "language_model.model.layers.3.self_attn.q_a_proj", |
| "language_model.model.layers.3.self_attn.q_b_proj", |
| "language_model.model.layers.3.self_attn.kv_a_proj_with_mqa", |
| "language_model.model.layers.3.self_attn.kv_b_proj", |
| "language_model.model.layers.3.self_attn.o_proj", |
| "language_model.model.layers.3.self_attn.g_proj", |
| "language_model.model.layers.3.block_sparse_moe.gate", |
| "language_model.model.layers.3.self_attention_res_proj", |
| "language_model.model.layers.3.mlp_res_proj", |
| "language_model.model.layers.4.self_attn.q_proj", |
| "language_model.model.layers.4.self_attn.k_proj", |
| "language_model.model.layers.4.self_attn.v_proj", |
| "language_model.model.layers.4.self_attn.f_a_proj", |
| "language_model.model.layers.4.self_attn.f_b_proj", |
| "language_model.model.layers.4.self_attn.b_proj", |
| "language_model.model.layers.4.self_attn.g_proj", |
| "language_model.model.layers.4.self_attn.o_norm", |
| "language_model.model.layers.4.self_attn.o_proj", |
| "language_model.model.layers.4.block_sparse_moe.gate", |
| "language_model.model.layers.4.self_attention_res_proj", |
| "language_model.model.layers.4.mlp_res_proj", |
| "language_model.model.layers.5.self_attn.q_proj", |
| "language_model.model.layers.5.self_attn.k_proj", |
| "language_model.model.layers.5.self_attn.v_proj", |
| "language_model.model.layers.5.self_attn.f_a_proj", |
| "language_model.model.layers.5.self_attn.f_b_proj", |
| "language_model.model.layers.5.self_attn.b_proj", |
| "language_model.model.layers.5.self_attn.g_proj", |
| "language_model.model.layers.5.self_attn.o_norm", |
| "language_model.model.layers.5.self_attn.o_proj", |
| "language_model.model.layers.5.block_sparse_moe.gate", |
| "language_model.model.layers.5.self_attention_res_proj", |
| "language_model.model.layers.5.mlp_res_proj", |
| "language_model.model.layers.6.self_attn.q_proj", |
| "language_model.model.layers.6.self_attn.k_proj", |
| "language_model.model.layers.6.self_attn.v_proj", |
| "language_model.model.layers.6.self_attn.f_a_proj", |
| "language_model.model.layers.6.self_attn.f_b_proj", |
| "language_model.model.layers.6.self_attn.b_proj", |
| "language_model.model.layers.6.self_attn.g_proj", |
| "language_model.model.layers.6.self_attn.o_norm", |
| "language_model.model.layers.6.self_attn.o_proj", |
| "language_model.model.layers.6.block_sparse_moe.gate", |
| "language_model.model.layers.6.self_attention_res_proj", |
| "language_model.model.layers.6.mlp_res_proj", |
| "language_model.model.layers.7.self_attn.q_a_proj", |
| "language_model.model.layers.7.self_attn.q_b_proj", |
| "language_model.model.layers.7.self_attn.kv_a_proj_with_mqa", |
| "language_model.model.layers.7.self_attn.kv_b_proj", |
| "language_model.model.layers.7.self_attn.o_proj", |
| "language_model.model.layers.7.self_attn.g_proj", |
| "language_model.model.layers.7.block_sparse_moe.gate", |
| "language_model.model.layers.7.self_attention_res_proj", |
| "language_model.model.layers.7.mlp_res_proj", |
| "language_model.model.output_attn_res_proj", |
| "language_model.lm_head", |
| "re:.*routed_expert.*" |
| ], |
| "kv_cache_scheme": null, |
| "quant_method": "compressed-tensors", |
| "quantization_status": "compressed", |
| "sparsity_config": {}, |
| "transform_config": {}, |
| "version": "0.17.2.dev32+g1f3b632.d20260723" |
| }, |
| "text_config": { |
| "_name_or_path": "", |
| "activation_situ_beta": 4.0, |
| "activation_situ_linear_beta": 25.0, |
| "architectures": null, |
| "attn_res_block_size": 4, |
| "auto_map": { |
| "AutoConfig": "configuration_kimi_k3.KimiLinearConfig", |
| "AutoModel": "modeling_kimi_linear.KimiLinearModel", |
| "AutoModelForCausalLM": "modeling_kimi_linear.KimiLinearForCausalLM" |
| }, |
| "bos_token_id": 1, |
| "chunk_size_feed_forward": 0, |
| "dtype": null, |
| "eos_token_id": 2, |
| "first_k_dense_replace": 1, |
| "head_dim": 74, |
| "hidden_act": "situ", |
| "hidden_size": 1024, |
| "id2label": { |
| "0": "LABEL_0", |
| "1": "LABEL_1" |
| }, |
| "initializer_range": 0.02, |
| "intermediate_size": 2048, |
| "is_encoder_decoder": false, |
| "kv_lora_rank": 128, |
| "label2id": { |
| "LABEL_0": 0, |
| "LABEL_1": 1 |
| }, |
| "latent_moe_use_norm": true, |
| "linear_attn_config": { |
| "full_attn_layers": [ |
| 4, |
| 8 |
| ], |
| "head_dim": 32, |
| "kda_layers": [ |
| 1, |
| 2, |
| 3, |
| 5, |
| 6, |
| 7 |
| ], |
| "num_heads": 8, |
| "short_conv_kernel_size": 4, |
| "use_full_rank_gate": true |
| }, |
| "max_position_embeddings": 4096, |
| "mla_use_nope": true, |
| "mla_use_output_gate": true, |
| "model_type": "kimi_linear", |
| "moe_intermediate_size": 256, |
| "moe_layer_freq": 1, |
| "moe_renormalize": true, |
| "moe_router_activation_func": "sigmoid", |
| "num_attention_heads": 8, |
| "num_expert_group": 1, |
| "num_experts": 8, |
| "num_experts_per_token": 2, |
| "num_hidden_layers": 8, |
| "num_key_value_heads": 8, |
| "num_nextn_predict_layers": 0, |
| "num_shared_experts": 1, |
| "output_attentions": false, |
| "output_hidden_states": false, |
| "pad_token_id": 0, |
| "problem_type": null, |
| "q_lora_rank": 256, |
| "qk_nope_head_dim": 64, |
| "qk_rope_head_dim": 32, |
| "return_dict": true, |
| "rms_norm_eps": 1e-05, |
| "rope_parameters": { |
| "rope_theta": 10000.0, |
| "rope_type": "default" |
| }, |
| "rope_theta": 10000.0, |
| "routed_expert_hidden_size": 512, |
| "routed_scaling_factor": 1.0, |
| "tie_word_embeddings": false, |
| "topk_group": 1, |
| "topk_method": "noaux_tc", |
| "use_cache": true, |
| "use_grouped_topk": true, |
| "v_head_dim": 64, |
| "vocab_size": 163840 |
| }, |
| "transformers_version": "5.15.0.dev0", |
| "vision_config": { |
| "_name_or_path": "", |
| "activation_func": "gelu_pytorch_tanh", |
| "architectures": null, |
| "attn_bias": false, |
| "chunk_size_feed_forward": 0, |
| "dtype": null, |
| "id2label": { |
| "0": "LABEL_0", |
| "1": "LABEL_1" |
| }, |
| "init_pos_emb_height": 64, |
| "init_pos_emb_time": 4, |
| "init_pos_emb_width": 64, |
| "is_encoder_decoder": false, |
| "label2id": { |
| "LABEL_0": 0, |
| "LABEL_1": 1 |
| }, |
| "linear_bias": false, |
| "merge_kernel_size": [ |
| 2, |
| 2 |
| ], |
| "merge_type": "sd2_tpool", |
| "mlp_type": "mlp2", |
| "mm_hidden_size": 256, |
| "mm_projector_type": "patchmergerv2", |
| "model_type": "", |
| "norm_type": "rmsnorm", |
| "output_attentions": false, |
| "output_hidden_states": false, |
| "patch_embed_proj_bias": false, |
| "patch_size": 14, |
| "pos_emb_interpolation_mode": "bilinear", |
| "pos_emb_type": "divided_fixed", |
| "problem_type": null, |
| "projector_hidden_act": "gelu", |
| "projector_ln_eps": 1e-05, |
| "qkv_hidden_size": 1536, |
| "return_dict": true, |
| "text_hidden_size": 1024, |
| "vt_hidden_size": 256, |
| "vt_intermediate_size": 512, |
| "vt_num_attention_heads": 4, |
| "vt_num_hidden_layers": 2 |
| } |
| } |
