| { | |
| "architectures": [ | |
| "KimiVLForConditionalGeneration" | |
| ], | |
| "attn_mechanism": "ragged_page_attention_v3", | |
| "auto_map": { | |
| "AutoConfig": "configuration_kimi_vl.KimiVLConfig", | |
| "AutoModel": "modeling_kimi_vl.KimiVLForConditionalGeneration", | |
| "AutoModelForCausalLM": "modeling_kimi_vl.KimiVLForConditionalGeneration" | |
| }, | |
| "backend": null, | |
| "bits": null, | |
| "blocksize_b": 1, | |
| "blocksize_k": 128, | |
| "blocksize_q": 128, | |
| "decode_attn_mechanism": null, | |
| "dtype": "bfloat16", | |
| "easy_method": "train", | |
| "fcm_max_ratio": 0.0, | |
| "fcm_min_ratio": 0.0, | |
| "flash_attention_backward_pass_impl": "triton", | |
| "freq_max_position_embeddings": 40960, | |
| "fsdp_is_ep_bound": true, | |
| "gradient_checkpointing": "", | |
| "gradient_checkpointing_targets": null, | |
| "hardware_abstraction": true, | |
| "ignore_index": -100, | |
| "kv_cache_quantization_config": null, | |
| "kv_cache_sharding_sequence_axis_name": "sp", | |
| "mask_max_position_embeddings": 40960, | |
| "media_placeholder_token_id": 163605, | |
| "model_type": "kimi_vl", | |
| "moe_force_xla_gmm": false, | |
| "moe_method": "fused_moe", | |
| "moe_tiling_size_batch": 4, | |
| "moe_tiling_size_dim": 128, | |
| "moe_tiling_size_seqlen": 128, | |
| "operation_configs": null, | |
| "pad_token_id": 0, | |
| "pallas_k_block_size": 128, | |
| "pallas_m_block_size": 128, | |
| "pallas_n_block_size": 128, | |
| "partition_axis": { | |
| "attention_dim_axis": null, | |
| "attention_kv_dim_axis": null, | |
| "batch_axis": [ | |
| "fsdp", | |
| "dp" | |
| ], | |
| "bias_head_sequence_axis": null, | |
| "bias_key_sequence_axis": null, | |
| "data_parallel_axis": "dp", | |
| "decode_attention_dim_axis": null, | |
| "decode_attention_kv_dim_axis": null, | |
| "decode_batch_axis": [ | |
| "fsdp", | |
| "dp" | |
| ], | |
| "decode_head_axis": "tp", | |
| "decode_key_sequence_axis": "sp", | |
| "decode_kv_head_axis": "tp", | |
| "decode_query_sequence_axis": null, | |
| "expert_axis": "ep", | |
| "expert_gate_axis": null, | |
| "expert_parallel_axis": "ep", | |
| "fully_sharded_data_parallel_axis": "fsdp", | |
| "head_axis": "tp", | |
| "hidden_state_axis": "tp", | |
| "key_sequence_axis": "sp", | |
| "kv_head_axis": "tp", | |
| "mlp_intermediate_axis": "tp", | |
| "query_sequence_axis": "sp", | |
| "sequence_axis": "sp", | |
| "sequence_parallel_axis": "sp", | |
| "tensor_parallel_axis": "tp", | |
| "vocab_axis": "tp" | |
| }, | |
| "platform": null, | |
| "precompute_masks": true, | |
| "pretraining_tp": 1, | |
| "quantization_config": null, | |
| "scan_attention_layers": false, | |
| "scan_mlp_chunk_size": 1024, | |
| "scan_ring_attention": true, | |
| "sequence_axis_name": "sp", | |
| "sharding_axis_dims": [ | |
| 1, | |
| 1, | |
| 1, | |
| -1, | |
| 1 | |
| ], | |
| "sharding_axis_names": [ | |
| "dp", | |
| "fsdp", | |
| "ep", | |
| "tp", | |
| "sp" | |
| ], | |
| "sharding_dcn_axis_dims": null, | |
| "sp_is_ep_bound": true, | |
| "text_config": { | |
| "architectures": [ | |
| "KimiVLForConditionalGeneration" | |
| ], | |
| "attention_bias": false, | |
| "attention_dropout": 0.0, | |
| "attn_mechanism": "ragged_page_attention_v3", | |
| "auto_map": { | |
| "AutoConfig": "configuration_kimi_vl.KimiVLConfig", | |
| "AutoModel": "modeling_kimi_vl.KimiVLForConditionalGeneration", | |
| "AutoModelForCausalLM": "modeling_kimi_vl.KimiVLForConditionalGeneration" | |
| }, | |
| "aux_loss_alpha": 0.001, | |
| "backend": null, | |
| "bits": null, | |
| "blocksize_b": 1, | |
| "blocksize_k": 128, | |
| "blocksize_q": 128, | |
| "bos_token_id": 163584, | |
| "decode_attn_mechanism": null, | |
| "dtype": "bfloat16", | |
| "easy_method": "train", | |
| "eos_token_id": 163585, | |
| "ep_size": 1, | |
| "fcm_max_ratio": 0.0, | |
| "fcm_min_ratio": 0.0, | |
| "first_k_dense_replace": 1, | |
| "flash_attention_backward_pass_impl": "triton", | |
| "freq_max_position_embeddings": 40960, | |
| "fsdp_is_ep_bound": true, | |
| "gradient_checkpointing": "", | |
| "gradient_checkpointing_targets": null, | |
| "hardware_abstraction": true, | |
| "head_dim": 192, | |
| "hidden_act": "silu", | |
| "hidden_size": 2048, | |
| "initializer_range": 0.02, | |
| "intermediate_size": 11264, | |
| "kv_cache_quantization_config": null, | |
| "kv_cache_sharding_sequence_axis_name": "sp", | |
| "kv_lora_rank": 512, | |
| "layer_types": [ | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention", | |
| "full_attention" | |
| ], | |
| "mask_max_position_embeddings": 40960, | |
| "max_position_embeddings": 131072, | |
| "model_type": "deepseek_v3", | |
| "moe_force_xla_gmm": false, | |
| "moe_intermediate_size": 1408, | |
| "moe_layer_freq": 1, | |
| "moe_method": "fused_moe", | |
| "moe_tiling_size_batch": 4, | |
| "moe_tiling_size_dim": 128, | |
| "moe_tiling_size_seqlen": 128, | |
| "n_group": 1, | |
| "n_routed_experts": 64, | |
| "n_shared_experts": 2, | |
| "norm_topk_prob": true, | |
| "num_attention_heads": 16, | |
| "num_experts_per_tok": 6, | |
| "num_hidden_layers": 27, | |
| "num_key_value_heads": 16, | |
| "num_nextn_predict_layers": 1, | |
| "operation_configs": null, | |
| "pad_token_id": 163839, | |
| "pallas_k_block_size": 128, | |
| "pallas_m_block_size": 128, | |
| "pallas_n_block_size": 128, | |
| "partition_axis": { | |
| "attention_dim_axis": null, | |
| "attention_kv_dim_axis": null, | |
| "batch_axis": [ | |
| "fsdp", | |
| "dp" | |
| ], | |
| "bias_head_sequence_axis": null, | |
| "bias_key_sequence_axis": null, | |
| "data_parallel_axis": "dp", | |
| "decode_attention_dim_axis": null, | |
| "decode_attention_kv_dim_axis": null, | |
| "decode_batch_axis": [ | |
| "fsdp", | |
| "dp" | |
| ], | |
| "decode_head_axis": "tp", | |
| "decode_key_sequence_axis": "sp", | |
| "decode_kv_head_axis": "tp", | |
| "decode_query_sequence_axis": null, | |
| "expert_axis": "ep", | |
| "expert_gate_axis": null, | |
| "expert_parallel_axis": "ep", | |
| "fully_sharded_data_parallel_axis": "fsdp", | |
| "head_axis": "tp", | |
| "hidden_state_axis": "tp", | |
| "key_sequence_axis": "sp", | |
| "kv_head_axis": "tp", | |
| "mlp_intermediate_axis": "tp", | |
| "query_sequence_axis": "sp", | |
| "sequence_axis": "sp", | |
| "sequence_parallel_axis": "sp", | |
| "tensor_parallel_axis": "tp", | |
| "vocab_axis": "tp" | |
| }, | |
| "platform": null, | |
| "precompute_masks": true, | |
| "pretraining_tp": 1, | |
| "q_lora_rank": null, | |
| "qk_nope_head_dim": 128, | |
| "qk_rope_head_dim": 64, | |
| "quantization_config": null, | |
| "rms_norm_eps": 1e-05, | |
| "rope_scaling": null, | |
| "rope_theta": 800000.0, | |
| "routed_scaling_factor": 2.446, | |
| "scan_attention_layers": false, | |
| "scan_mlp_chunk_size": 1024, | |
| "scan_ring_attention": true, | |
| "scoring_func": "sigmoid", | |
| "seq_aux": true, | |
| "sequence_axis_name": "sp", | |
| "sharding_axis_dims": [ | |
| 1, | |
| 1, | |
| 1, | |
| -1, | |
| 1 | |
| ], | |
| "sharding_axis_names": [ | |
| "dp", | |
| "fsdp", | |
| "ep", | |
| "tp", | |
| "sp" | |
| ], | |
| "sharding_dcn_axis_dims": null, | |
| "sp_is_ep_bound": true, | |
| "topk_group": 1, | |
| "topk_method": "noaux_tc", | |
| "use_cache": true, | |
| "use_expert_tensor_mode": false, | |
| "use_ring_of_experts": false, | |
| "use_scan_mlp": false, | |
| "use_sharded_kv_caching": false, | |
| "use_sharding_constraint": false, | |
| "v_head_dim": 128, | |
| "vocab_size": 163840 | |
| }, | |
| "tie_word_embeddings": false, | |
| "transformers_version": "4.57.3", | |
| "use_expert_tensor_mode": false, | |
| "use_ring_of_experts": false, | |
| "use_scan_mlp": false, | |
| "use_sharded_kv_caching": false, | |
| "use_sharding_constraint": false, | |
| "vision_config": { | |
| "architectures": [ | |
| "KimiVLForConditionalGeneration" | |
| ], | |
| "attn_mechanism": "ragged_page_attention_v3", | |
| "auto_map": { | |
| "AutoConfig": "configuration_kimi_vl.KimiVLConfig", | |
| "AutoModel": "modeling_kimi_vl.KimiVLForConditionalGeneration", | |
| "AutoModelForCausalLM": "modeling_kimi_vl.KimiVLForConditionalGeneration" | |
| }, | |
| "backend": null, | |
| "bits": null, | |
| "blocksize_b": 1, | |
| "blocksize_k": 128, | |
| "blocksize_q": 128, | |
| "decode_attn_mechanism": null, | |
| "dtype": "bfloat16", | |
| "easy_method": "train", | |
| "fcm_max_ratio": 0.0, | |
| "fcm_min_ratio": 0.0, | |
| "flash_attention_backward_pass_impl": "triton", | |
| "freq_max_position_embeddings": 40960, | |
| "fsdp_is_ep_bound": true, | |
| "gradient_checkpointing": "", | |
| "gradient_checkpointing_targets": null, | |
| "hardware_abstraction": true, | |
| "hidden_size": 1152, | |
| "init_pos_emb_height": 64, | |
| "init_pos_emb_width": 64, | |
| "intermediate_size": 4304, | |
| "kv_cache_quantization_config": null, | |
| "kv_cache_sharding_sequence_axis_name": "sp", | |
| "mask_max_position_embeddings": 40960, | |
| "merge_kernel_size": [ | |
| 2, | |
| 2 | |
| ], | |
| "model_type": "moonvit", | |
| "moe_force_xla_gmm": false, | |
| "moe_method": "fused_moe", | |
| "moe_tiling_size_batch": 4, | |
| "moe_tiling_size_dim": 128, | |
| "moe_tiling_size_seqlen": 128, | |
| "num_attention_heads": 16, | |
| "num_hidden_layers": 27, | |
| "operation_configs": null, | |
| "pallas_k_block_size": 128, | |
| "pallas_m_block_size": 128, | |
| "pallas_n_block_size": 128, | |
| "partition_axis": { | |
| "attention_dim_axis": null, | |
| "attention_kv_dim_axis": null, | |
| "batch_axis": [ | |
| "fsdp", | |
| "dp" | |
| ], | |
| "bias_head_sequence_axis": null, | |
| "bias_key_sequence_axis": null, | |
| "data_parallel_axis": "dp", | |
| "decode_attention_dim_axis": null, | |
| "decode_attention_kv_dim_axis": null, | |
| "decode_batch_axis": [ | |
| "fsdp", | |
| "dp" | |
| ], | |
| "decode_head_axis": "tp", | |
| "decode_key_sequence_axis": "sp", | |
| "decode_kv_head_axis": "tp", | |
| "decode_query_sequence_axis": null, | |
| "expert_axis": "ep", | |
| "expert_gate_axis": null, | |
| "expert_parallel_axis": "ep", | |
| "fully_sharded_data_parallel_axis": "fsdp", | |
| "head_axis": "tp", | |
| "hidden_state_axis": "tp", | |
| "key_sequence_axis": "sp", | |
| "kv_head_axis": "tp", | |
| "mlp_intermediate_axis": "tp", | |
| "query_sequence_axis": "sp", | |
| "sequence_axis": "sp", | |
| "sequence_parallel_axis": "sp", | |
| "tensor_parallel_axis": "tp", | |
| "vocab_axis": "tp" | |
| }, | |
| "patch_size": 14, | |
| "platform": null, | |
| "precompute_masks": true, | |
| "pretraining_tp": 1, | |
| "quantization_config": null, | |
| "scan_attention_layers": false, | |
| "scan_mlp_chunk_size": 1024, | |
| "scan_ring_attention": true, | |
| "sequence_axis_name": "sp", | |
| "sharding_axis_dims": [ | |
| 1, | |
| 1, | |
| 1, | |
| -1, | |
| 1 | |
| ], | |
| "sharding_axis_names": [ | |
| "dp", | |
| "fsdp", | |
| "ep", | |
| "tp", | |
| "sp" | |
| ], | |
| "sharding_dcn_axis_dims": null, | |
| "sp_is_ep_bound": true, | |
| "use_expert_tensor_mode": false, | |
| "use_ring_of_experts": false, | |
| "use_scan_mlp": false, | |
| "use_sharded_kv_caching": false, | |
| "use_sharding_constraint": false, | |
| "vocab_size": 163840 | |
| }, | |
| "vocab_size": 163840 | |
| } | |