| { |
| "architectures": [ |
| "KimiLinearForCausalLM" |
| ], |
| "auto_map": { |
| "AutoConfig": "configuration_kimi.KimiLinearConfig", |
| "AutoModel": "modeling_kimi.KimiLinearModel", |
| "AutoModelForCausalLM": "modeling_kimi.KimiLinearForCausalLM" |
| }, |
| "bos_token_id": 163584, |
| "dtype": "bfloat16", |
| "eos_token_id": 163586, |
| "first_k_dense_replace": 1, |
| "head_dim": 72, |
| "hidden_act": "silu", |
| "hidden_size": 2304, |
| "initializer_range": 0.02, |
| "intermediate_size": 9216, |
| "kv_lora_rank": 512, |
| "linear_attn_config": { |
| "full_attn_layers": [ |
| 4, |
| 8, |
| 12, |
| 16, |
| 20, |
| 24, |
| 27 |
| ], |
| "head_dim": 128, |
| "kda_layers": [ |
| 1, |
| 2, |
| 3, |
| 5, |
| 6, |
| 7, |
| 9, |
| 10, |
| 11, |
| 13, |
| 14, |
| 15, |
| 17, |
| 18, |
| 19, |
| 21, |
| 22, |
| 23, |
| 25, |
| 26 |
| ], |
| "num_heads": 32, |
| "short_conv_kernel_size": 4 |
| }, |
| "mla_use_nope": true, |
| "model_max_length": 1048576, |
| "model_type": "kimi_linear", |
| "moe_intermediate_size": 1024, |
| "moe_layer_freq": 1, |
| "moe_renormalize": true, |
| "moe_router_activation_func": "sigmoid", |
| "num_attention_heads": 32, |
| "num_expert_group": 1, |
| "num_experts": 256, |
| "num_experts_per_token": 8, |
| "num_hidden_layers": 27, |
| "num_key_value_heads": 32, |
| "num_nextn_predict_layers": 0, |
| "num_shared_experts": 1, |
| "pad_token_id": 163839, |
| "q_lora_rank": null, |
| "qk_nope_head_dim": 128, |
| "qk_rope_head_dim": 64, |
| "quantization_config": { |
| "config_groups": { |
| "group_0": { |
| "format": "float-quantized", |
| "input_activations": { |
| "actorder": null, |
| "block_structure": null, |
| "dynamic": true, |
| "group_size": 128, |
| "num_bits": 8, |
| "observer": null, |
| "observer_kwargs": {}, |
| "strategy": "group", |
| "symmetric": true, |
| "type": "float" |
| }, |
| "output_activations": null, |
| "targets": [ |
| "Linear" |
| ], |
| "weights": { |
| "actorder": null, |
| "block_structure": [ |
| 128, |
| 128 |
| ], |
| "dynamic": false, |
| "group_size": null, |
| "num_bits": 8, |
| "observer": "minmax", |
| "observer_kwargs": {}, |
| "strategy": "block", |
| "symmetric": true, |
| "type": "float" |
| } |
| } |
| }, |
| "format": "float-quantized", |
| "global_compression_ratio": null, |
| "ignore": [ |
| "model.layers.0.self_attn.q_proj", |
| "model.layers.0.self_attn.k_proj", |
| "model.layers.0.self_attn.v_proj", |
| "model.layers.0.self_attn.f_a_proj", |
| "model.layers.0.self_attn.f_b_proj", |
| "model.layers.0.self_attn.b_proj", |
| "model.layers.0.self_attn.g_a_proj", |
| "model.layers.0.self_attn.g_b_proj", |
| "model.layers.0.self_attn.o_proj", |
| "model.layers.0.mlp.gate_proj", |
| "model.layers.0.mlp.up_proj", |
| "model.layers.0.mlp.down_proj", |
| "model.layers.1.self_attn.q_proj", |
| "model.layers.1.self_attn.k_proj", |
| "model.layers.1.self_attn.v_proj", |
| "model.layers.1.self_attn.f_a_proj", |
| "model.layers.1.self_attn.f_b_proj", |
| "model.layers.1.self_attn.b_proj", |
| "model.layers.1.self_attn.g_a_proj", |
| "model.layers.1.self_attn.g_b_proj", |
| "model.layers.1.self_attn.o_proj", |
| "model.layers.1.block_sparse_moe.shared_experts.gate_proj", |
| "model.layers.1.block_sparse_moe.shared_experts.up_proj", |
| "model.layers.1.block_sparse_moe.shared_experts.down_proj", |
| "model.layers.2.self_attn.q_proj", |
| "model.layers.2.self_attn.k_proj", |
| "model.layers.2.self_attn.v_proj", |
| "model.layers.2.self_attn.f_a_proj", |
| "model.layers.2.self_attn.f_b_proj", |
| "model.layers.2.self_attn.b_proj", |
| "model.layers.2.self_attn.g_a_proj", |
| "model.layers.2.self_attn.g_b_proj", |
| "model.layers.2.self_attn.o_proj", |
| "model.layers.2.block_sparse_moe.shared_experts.gate_proj", |
| "model.layers.2.block_sparse_moe.shared_experts.up_proj", |
| "model.layers.2.block_sparse_moe.shared_experts.down_proj", |
| "model.layers.3.self_attn.q_proj", |
| "model.layers.3.self_attn.kv_a_proj_with_mqa", |
| "model.layers.3.self_attn.kv_b_proj", |
| "model.layers.3.self_attn.o_proj", |
| "model.layers.3.block_sparse_moe.shared_experts.gate_proj", |
| "model.layers.3.block_sparse_moe.shared_experts.up_proj", |
| "model.layers.3.block_sparse_moe.shared_experts.down_proj", |
| "model.layers.4.self_attn.q_proj", |
| "model.layers.4.self_attn.k_proj", |
| "model.layers.4.self_attn.v_proj", |
| "model.layers.4.self_attn.f_a_proj", |
| "model.layers.4.self_attn.f_b_proj", |
| "model.layers.4.self_attn.b_proj", |
| "model.layers.4.self_attn.g_a_proj", |
| "model.layers.4.self_attn.g_b_proj", |
| "model.layers.4.self_attn.o_proj", |
| "model.layers.4.block_sparse_moe.shared_experts.gate_proj", |
| "model.layers.4.block_sparse_moe.shared_experts.up_proj", |
| "model.layers.4.block_sparse_moe.shared_experts.down_proj", |
| "model.layers.5.self_attn.q_proj", |
| "model.layers.5.self_attn.k_proj", |
| "model.layers.5.self_attn.v_proj", |
| "model.layers.5.self_attn.f_a_proj", |
| "model.layers.5.self_attn.f_b_proj", |
| "model.layers.5.self_attn.b_proj", |
| "model.layers.5.self_attn.g_a_proj", |
| "model.layers.5.self_attn.g_b_proj", |
| "model.layers.5.self_attn.o_proj", |
| "model.layers.5.block_sparse_moe.shared_experts.gate_proj", |
| "model.layers.5.block_sparse_moe.shared_experts.up_proj", |
| "model.layers.5.block_sparse_moe.shared_experts.down_proj", |
| "model.layers.6.self_attn.q_proj", |
| "model.layers.6.self_attn.k_proj", |
| "model.layers.6.self_attn.v_proj", |
| "model.layers.6.self_attn.f_a_proj", |
| "model.layers.6.self_attn.f_b_proj", |
| "model.layers.6.self_attn.b_proj", |
| "model.layers.6.self_attn.g_a_proj", |
| "model.layers.6.self_attn.g_b_proj", |
| "model.layers.6.self_attn.o_proj", |
| "model.layers.6.block_sparse_moe.shared_experts.gate_proj", |
| "model.layers.6.block_sparse_moe.shared_experts.up_proj", |
| "model.layers.6.block_sparse_moe.shared_experts.down_proj", |
| "model.layers.7.self_attn.q_proj", |
| "model.layers.7.self_attn.kv_a_proj_with_mqa", |
| "model.layers.7.self_attn.kv_b_proj", |
| "model.layers.7.self_attn.o_proj", |
| "model.layers.7.block_sparse_moe.shared_experts.gate_proj", |
| "model.layers.7.block_sparse_moe.shared_experts.up_proj", |
| "model.layers.7.block_sparse_moe.shared_experts.down_proj", |
| "model.layers.8.self_attn.q_proj", |
| "model.layers.8.self_attn.k_proj", |
| "model.layers.8.self_attn.v_proj", |
| "model.layers.8.self_attn.f_a_proj", |
| "model.layers.8.self_attn.f_b_proj", |
| "model.layers.8.self_attn.b_proj", |
| "model.layers.8.self_attn.g_a_proj", |
| "model.layers.8.self_attn.g_b_proj", |
| "model.layers.8.self_attn.o_proj", |
| "model.layers.8.block_sparse_moe.shared_experts.gate_proj", |
| "model.layers.8.block_sparse_moe.shared_experts.up_proj", |
| "model.layers.8.block_sparse_moe.shared_experts.down_proj", |
| "model.layers.9.self_attn.q_proj", |
| "model.layers.9.self_attn.k_proj", |
| "model.layers.9.self_attn.v_proj", |
| "model.layers.9.self_attn.f_a_proj", |
| "model.layers.9.self_attn.f_b_proj", |
| "model.layers.9.self_attn.b_proj", |
| "model.layers.9.self_attn.g_a_proj", |
| "model.layers.9.self_attn.g_b_proj", |
| "model.layers.9.self_attn.o_proj", |
| "model.layers.9.block_sparse_moe.shared_experts.gate_proj", |
| "model.layers.9.block_sparse_moe.shared_experts.up_proj", |
| "model.layers.9.block_sparse_moe.shared_experts.down_proj", |
| "model.layers.10.self_attn.q_proj", |
| "model.layers.10.self_attn.k_proj", |
| "model.layers.10.self_attn.v_proj", |
| "model.layers.10.self_attn.f_a_proj", |
| "model.layers.10.self_attn.f_b_proj", |
| "model.layers.10.self_attn.b_proj", |
| "model.layers.10.self_attn.g_a_proj", |
| "model.layers.10.self_attn.g_b_proj", |
| "model.layers.10.self_attn.o_proj", |
| "model.layers.10.block_sparse_moe.shared_experts.gate_proj", |
| "model.layers.10.block_sparse_moe.shared_experts.up_proj", |
| "model.layers.10.block_sparse_moe.shared_experts.down_proj", |
| "model.layers.11.self_attn.q_proj", |
| "model.layers.11.self_attn.kv_a_proj_with_mqa", |
| "model.layers.11.self_attn.kv_b_proj", |
| "model.layers.11.self_attn.o_proj", |
| "model.layers.11.block_sparse_moe.shared_experts.gate_proj", |
| "model.layers.11.block_sparse_moe.shared_experts.up_proj", |
| "model.layers.11.block_sparse_moe.shared_experts.down_proj", |
| "model.layers.12.self_attn.q_proj", |
| "model.layers.12.self_attn.k_proj", |
| "model.layers.12.self_attn.v_proj", |
| "model.layers.12.self_attn.f_a_proj", |
| "model.layers.12.self_attn.f_b_proj", |
| "model.layers.12.self_attn.b_proj", |
| "model.layers.12.self_attn.g_a_proj", |
| "model.layers.12.self_attn.g_b_proj", |
| "model.layers.12.self_attn.o_proj", |
| "model.layers.12.block_sparse_moe.shared_experts.gate_proj", |
| "model.layers.12.block_sparse_moe.shared_experts.up_proj", |
| "model.layers.12.block_sparse_moe.shared_experts.down_proj", |
| "model.layers.13.self_attn.q_proj", |
| "model.layers.13.self_attn.k_proj", |
| "model.layers.13.self_attn.v_proj", |
| "model.layers.13.self_attn.f_a_proj", |
| "model.layers.13.self_attn.f_b_proj", |
| "model.layers.13.self_attn.b_proj", |
| "model.layers.13.self_attn.g_a_proj", |
| "model.layers.13.self_attn.g_b_proj", |
| "model.layers.13.self_attn.o_proj", |
| "model.layers.13.block_sparse_moe.shared_experts.gate_proj", |
| "model.layers.13.block_sparse_moe.shared_experts.up_proj", |
| "model.layers.13.block_sparse_moe.shared_experts.down_proj", |
| "model.layers.14.self_attn.q_proj", |
| "model.layers.14.self_attn.k_proj", |
| "model.layers.14.self_attn.v_proj", |
| "model.layers.14.self_attn.f_a_proj", |
| "model.layers.14.self_attn.f_b_proj", |
| "model.layers.14.self_attn.b_proj", |
| "model.layers.14.self_attn.g_a_proj", |
| "model.layers.14.self_attn.g_b_proj", |
| "model.layers.14.self_attn.o_proj", |
| "model.layers.14.block_sparse_moe.shared_experts.gate_proj", |
| "model.layers.14.block_sparse_moe.shared_experts.up_proj", |
| "model.layers.14.block_sparse_moe.shared_experts.down_proj", |
| "model.layers.15.self_attn.q_proj", |
| "model.layers.15.self_attn.kv_a_proj_with_mqa", |
| "model.layers.15.self_attn.kv_b_proj", |
| "model.layers.15.self_attn.o_proj", |
| "model.layers.15.block_sparse_moe.shared_experts.gate_proj", |
| "model.layers.15.block_sparse_moe.shared_experts.up_proj", |
| "model.layers.15.block_sparse_moe.shared_experts.down_proj", |
| "model.layers.16.self_attn.q_proj", |
| "model.layers.16.self_attn.k_proj", |
| "model.layers.16.self_attn.v_proj", |
| "model.layers.16.self_attn.f_a_proj", |
| "model.layers.16.self_attn.f_b_proj", |
| "model.layers.16.self_attn.b_proj", |
| "model.layers.16.self_attn.g_a_proj", |
| "model.layers.16.self_attn.g_b_proj", |
| "model.layers.16.self_attn.o_proj", |
| "model.layers.16.block_sparse_moe.shared_experts.gate_proj", |
| "model.layers.16.block_sparse_moe.shared_experts.up_proj", |
| "model.layers.16.block_sparse_moe.shared_experts.down_proj", |
| "model.layers.17.self_attn.q_proj", |
| "model.layers.17.self_attn.k_proj", |
| "model.layers.17.self_attn.v_proj", |
| "model.layers.17.self_attn.f_a_proj", |
| "model.layers.17.self_attn.f_b_proj", |
| "model.layers.17.self_attn.b_proj", |
| "model.layers.17.self_attn.g_a_proj", |
| "model.layers.17.self_attn.g_b_proj", |
| "model.layers.17.self_attn.o_proj", |
| "model.layers.17.block_sparse_moe.shared_experts.gate_proj", |
| "model.layers.17.block_sparse_moe.shared_experts.up_proj", |
| "model.layers.17.block_sparse_moe.shared_experts.down_proj", |
| "model.layers.18.self_attn.q_proj", |
| "model.layers.18.self_attn.k_proj", |
| "model.layers.18.self_attn.v_proj", |
| "model.layers.18.self_attn.f_a_proj", |
| "model.layers.18.self_attn.f_b_proj", |
| "model.layers.18.self_attn.b_proj", |
| "model.layers.18.self_attn.g_a_proj", |
| "model.layers.18.self_attn.g_b_proj", |
| "model.layers.18.self_attn.o_proj", |
| "model.layers.18.block_sparse_moe.shared_experts.gate_proj", |
| "model.layers.18.block_sparse_moe.shared_experts.up_proj", |
| "model.layers.18.block_sparse_moe.shared_experts.down_proj", |
| "model.layers.19.self_attn.q_proj", |
| "model.layers.19.self_attn.kv_a_proj_with_mqa", |
| "model.layers.19.self_attn.kv_b_proj", |
| "model.layers.19.self_attn.o_proj", |
| "model.layers.19.block_sparse_moe.shared_experts.gate_proj", |
| "model.layers.19.block_sparse_moe.shared_experts.up_proj", |
| "model.layers.19.block_sparse_moe.shared_experts.down_proj", |
| "model.layers.20.self_attn.q_proj", |
| "model.layers.20.self_attn.k_proj", |
| "model.layers.20.self_attn.v_proj", |
| "model.layers.20.self_attn.f_a_proj", |
| "model.layers.20.self_attn.f_b_proj", |
| "model.layers.20.self_attn.b_proj", |
| "model.layers.20.self_attn.g_a_proj", |
| "model.layers.20.self_attn.g_b_proj", |
| "model.layers.20.self_attn.o_proj", |
| "model.layers.20.block_sparse_moe.shared_experts.gate_proj", |
| "model.layers.20.block_sparse_moe.shared_experts.up_proj", |
| "model.layers.20.block_sparse_moe.shared_experts.down_proj", |
| "model.layers.21.self_attn.q_proj", |
| "model.layers.21.self_attn.k_proj", |
| "model.layers.21.self_attn.v_proj", |
| "model.layers.21.self_attn.f_a_proj", |
| "model.layers.21.self_attn.f_b_proj", |
| "model.layers.21.self_attn.b_proj", |
| "model.layers.21.self_attn.g_a_proj", |
| "model.layers.21.self_attn.g_b_proj", |
| "model.layers.21.self_attn.o_proj", |
| "model.layers.21.block_sparse_moe.shared_experts.gate_proj", |
| "model.layers.21.block_sparse_moe.shared_experts.up_proj", |
| "model.layers.21.block_sparse_moe.shared_experts.down_proj", |
| "model.layers.22.self_attn.q_proj", |
| "model.layers.22.self_attn.k_proj", |
| "model.layers.22.self_attn.v_proj", |
| "model.layers.22.self_attn.f_a_proj", |
| "model.layers.22.self_attn.f_b_proj", |
| "model.layers.22.self_attn.b_proj", |
| "model.layers.22.self_attn.g_a_proj", |
| "model.layers.22.self_attn.g_b_proj", |
| "model.layers.22.self_attn.o_proj", |
| "model.layers.22.block_sparse_moe.shared_experts.gate_proj", |
| "model.layers.22.block_sparse_moe.shared_experts.up_proj", |
| "model.layers.22.block_sparse_moe.shared_experts.down_proj", |
| "model.layers.23.self_attn.q_proj", |
| "model.layers.23.self_attn.kv_a_proj_with_mqa", |
| "model.layers.23.self_attn.kv_b_proj", |
| "model.layers.23.self_attn.o_proj", |
| "model.layers.23.block_sparse_moe.shared_experts.gate_proj", |
| "model.layers.23.block_sparse_moe.shared_experts.up_proj", |
| "model.layers.23.block_sparse_moe.shared_experts.down_proj", |
| "model.layers.24.self_attn.q_proj", |
| "model.layers.24.self_attn.k_proj", |
| "model.layers.24.self_attn.v_proj", |
| "model.layers.24.self_attn.f_a_proj", |
| "model.layers.24.self_attn.f_b_proj", |
| "model.layers.24.self_attn.b_proj", |
| "model.layers.24.self_attn.g_a_proj", |
| "model.layers.24.self_attn.g_b_proj", |
| "model.layers.24.self_attn.o_proj", |
| "model.layers.24.block_sparse_moe.shared_experts.gate_proj", |
| "model.layers.24.block_sparse_moe.shared_experts.up_proj", |
| "model.layers.24.block_sparse_moe.shared_experts.down_proj", |
| "model.layers.25.self_attn.q_proj", |
| "model.layers.25.self_attn.k_proj", |
| "model.layers.25.self_attn.v_proj", |
| "model.layers.25.self_attn.f_a_proj", |
| "model.layers.25.self_attn.f_b_proj", |
| "model.layers.25.self_attn.b_proj", |
| "model.layers.25.self_attn.g_a_proj", |
| "model.layers.25.self_attn.g_b_proj", |
| "model.layers.25.self_attn.o_proj", |
| "model.layers.25.block_sparse_moe.shared_experts.gate_proj", |
| "model.layers.25.block_sparse_moe.shared_experts.up_proj", |
| "model.layers.25.block_sparse_moe.shared_experts.down_proj", |
| "model.layers.26.self_attn.q_proj", |
| "model.layers.26.self_attn.kv_a_proj_with_mqa", |
| "model.layers.26.self_attn.kv_b_proj", |
| "model.layers.26.self_attn.o_proj", |
| "model.layers.26.block_sparse_moe.shared_experts.gate_proj", |
| "model.layers.26.block_sparse_moe.shared_experts.up_proj", |
| "model.layers.26.block_sparse_moe.shared_experts.down_proj", |
| "lm_head" |
| ], |
| "kv_cache_scheme": null, |
| "quant_method": "compressed-tensors", |
| "quantization_status": "compressed", |
| "sparsity_config": {}, |
| "transform_config": {}, |
| "version": "0.12.2" |
| }, |
| "rms_norm_eps": 1e-05, |
| "rope_scaling": null, |
| "rope_theta": 10000.0, |
| "routed_scaling_factor": 2.446, |
| "tie_word_embeddings": false, |
| "topk_group": 1, |
| "transformers_version": "4.56.2", |
| "use_cache": true, |
| "use_grouped_topk": true, |
| "v_head_dim": 128, |
| "vocab_size": 163840 |
| } |