Download dspark/config.json from tiny-random/minimax-m3.1-flash: direct link, hf CLI and curl.
- Browser
- Download file 9.96 kB
-
https://huggingface.co/tiny-random/minimax-m3.1-flash/resolve/main/dspark/config.json
- Command line
-
hf download hf://tiny-random/minimax-m3.1-flash/dspark/config.json
-
curl -L -o config.json https://huggingface.co/tiny-random/minimax-m3.1-flash/resolve/main/dspark/config.json
9.96 kB
| { | |
| "transformers_version": "5.12.1", | |
| "architectures": [ | |
| "DSparkMiniMaxDraftModel" | |
| ], | |
| "output_hidden_states": false, | |
| "return_dict": true, | |
| "dtype": "bfloat16", | |
| "chunk_size_feed_forward": 0, | |
| "is_encoder_decoder": false, | |
| "id2label": { | |
| "0": "LABEL_0", | |
| "1": "LABEL_1" | |
| }, | |
| "label2id": { | |
| "LABEL_0": 0, | |
| "LABEL_1": 1 | |
| }, | |
| "problem_type": null, | |
| "model_type": "minimax_m3_vl", | |
| "text_config": { | |
| "architectures": [ | |
| "DSparkMiniMaxDraftModel" | |
| ], | |
| "output_hidden_states": false, | |
| "return_dict": true, | |
| "dtype": "bfloat16", | |
| "chunk_size_feed_forward": 0, | |
| "is_encoder_decoder": false, | |
| "id2label": { | |
| "0": "LABEL_0", | |
| "1": "LABEL_1" | |
| }, | |
| "label2id": { | |
| "LABEL_0": 0, | |
| "LABEL_1": 1 | |
| }, | |
| "problem_type": null, | |
| "_name_or_path": "", | |
| "hidden_size": 32, | |
| "intermediate_size": 32, | |
| "num_hidden_layers": 5, | |
| "num_attention_heads": 1, | |
| "num_key_value_heads": 1, | |
| "head_dim": 32, | |
| "vocab_size": 200064, | |
| "max_position_embeddings": 1048576, | |
| "rms_norm_eps": 1e-06, | |
| "use_gemma_norm": true, | |
| "attention_output_gate": false, | |
| "rope_theta": 10000000, | |
| "rotary_dim": 32, | |
| "partial_rotary_factor": 0.5, | |
| "hidden_act": "swigluoai", | |
| "use_qk_norm": true, | |
| "tie_word_embeddings": false, | |
| "dense_intermediate_size": 64, | |
| "shared_intermediate_size": 32, | |
| "num_local_experts": 128, | |
| "num_experts_per_tok": 4, | |
| "n_shared_experts": 1, | |
| "scoring_func": "sigmoid", | |
| "use_routing_bias": true, | |
| "moe_layer_freq": [ | |
| 0, | |
| 0, | |
| 0, | |
| 0, | |
| 0 | |
| ], | |
| "qk_norm_type": "per_head", | |
| "swiglu_alpha": 1.702, | |
| "swiglu_limit": 7.0, | |
| "routed_scaling_factor": 2.0, | |
| "sparse_attention_config": null, | |
| "dspark_block_size": 7, | |
| "dspark_markov_rank": 16, | |
| "dspark_markov_head_type": "vanilla", | |
| "dspark_aux_gemm_precision": "bf16", | |
| "dspark_noise_token_id": 200058, | |
| "dspark_use_dense_ffn": true, | |
| "dspark_ffn_hidden_size": 64, | |
| "dspark_target_layer_ids": [ | |
| 3, | |
| 17, | |
| 31, | |
| 45, | |
| 59 | |
| ], | |
| "dspark_config": { | |
| "markov_rank": 256, | |
| "markov_head_type": "vanilla", | |
| "mask_token_id": 200058, | |
| "hybrid_context_fusion": false, | |
| "aux_gemm_precision": "bf16", | |
| "use_dense_ffn": true, | |
| "ffn_hidden_size": 12288 | |
| }, | |
| "dspark_hybrid_context_fusion": false, | |
| "dflash_config": { | |
| "num_hidden_layers": 5, | |
| "num_target_layers": 60, | |
| "context_source": "target_layers", | |
| "block_size": 7, | |
| "mask_token_id": 200058, | |
| "target_layer_ids": [ | |
| 3, | |
| 17, | |
| 31, | |
| 45, | |
| 59 | |
| ], | |
| "layer_types": [ | |
| "sliding_attention", | |
| "sliding_attention", | |
| "sliding_attention", | |
| "sliding_attention", | |
| "sliding_attention" | |
| ], | |
| "sliding_window": 4096 | |
| }, | |
| "enable_confidence_head": true, | |
| "confidence_head_with_markov": true, | |
| "num_nextn_predict_layers": 0, | |
| "use_mtp": false, | |
| "num_mtp_modules": 0, | |
| "layer_types": [ | |
| "sliding_attention", | |
| "sliding_attention", | |
| "sliding_attention", | |
| "sliding_attention", | |
| "sliding_attention" | |
| ], | |
| "sliding_window": 4096, | |
| "v_head_dim": 128, | |
| "swa_head_dim": 128, | |
| "swa_v_head_dim": 128, | |
| "model_type": "", | |
| "output_attentions": false | |
| }, | |
| "img_token_compression_config": { | |
| "image_token_compression_method": "patch_merge", | |
| "spatial_merge_size": 2, | |
| "temporal_patch_size": 2 | |
| }, | |
| "image_grid_pinpoints": "[(336, 336), (336, 672), (336, 1008), (336, 1344), (336, 1680), (336, 2016), (672, 336), (672, 672), (672, 1008), (672, 1344), (672, 1680), (672, 2016), (1008, 336), (1008, 672), (1008, 1008), (1008, 1344), (1008, 1680), (1008, 2016), (1344, 336), (1344, 672), (1344, 1008), (1344, 1344), (1344, 1680), (1344, 2016), (1680, 336), (1680, 672), (1680, 1008), (1680, 1344), (1680, 1680), (1680, 2016), (2016, 336), (2016, 672), (2016, 1008), (2016, 1344), (2016, 1680), (2016, 2016)]", | |
| "image_seq_length": 576, | |
| "image_token_index": 200025, | |
| "video_token_index": 200026, | |
| "multimodal_projector_bias": true, | |
| "num_reward_heads": 0, | |
| "process_image_mode": "dynamic_res", | |
| "projector_hidden_act": "gelu", | |
| "vision_feature_layer": -1, | |
| "vision_feature_select_strategy": "full", | |
| "projector_hidden_size": 32, | |
| "quantization_config": { | |
| "quant_method": "mxfp8", | |
| "activation_scheme": "dynamic", | |
| "weight_block_size": [ | |
| 1, | |
| 32 | |
| ], | |
| "ignored_layers": [ | |
| "language_model.model.dspark.confidence_head.proj", | |
| "language_model.model.dspark.fc", | |
| "language_model.model.dspark.markov_head.markov_w1", | |
| "language_model.model.dspark.markov_head.markov_w2", | |
| "language_model.model.layers.10.block_sparse_moe.gate", | |
| "language_model.model.layers.11.block_sparse_moe.gate", | |
| "language_model.model.layers.12.block_sparse_moe.gate", | |
| "language_model.model.layers.13.block_sparse_moe.gate", | |
| "language_model.model.layers.14.block_sparse_moe.gate", | |
| "language_model.model.layers.15.block_sparse_moe.gate", | |
| "language_model.model.layers.16.block_sparse_moe.gate", | |
| "language_model.model.layers.17.block_sparse_moe.gate", | |
| "language_model.model.layers.18.block_sparse_moe.gate", | |
| "language_model.model.layers.19.block_sparse_moe.gate", | |
| "language_model.model.layers.20.block_sparse_moe.gate", | |
| "language_model.model.layers.21.block_sparse_moe.gate", | |
| "language_model.model.layers.22.block_sparse_moe.gate", | |
| "language_model.model.layers.23.block_sparse_moe.gate", | |
| "language_model.model.layers.24.block_sparse_moe.gate", | |
| "language_model.model.layers.25.block_sparse_moe.gate", | |
| "language_model.model.layers.26.block_sparse_moe.gate", | |
| "language_model.model.layers.27.block_sparse_moe.gate", | |
| "language_model.model.layers.28.block_sparse_moe.gate", | |
| "language_model.model.layers.29.block_sparse_moe.gate", | |
| "language_model.model.layers.3.block_sparse_moe.gate", | |
| "language_model.model.layers.30.block_sparse_moe.gate", | |
| "language_model.model.layers.31.block_sparse_moe.gate", | |
| "language_model.model.layers.32.block_sparse_moe.gate", | |
| "language_model.model.layers.33.block_sparse_moe.gate", | |
| "language_model.model.layers.34.block_sparse_moe.gate", | |
| "language_model.model.layers.35.block_sparse_moe.gate", | |
| "language_model.model.layers.36.block_sparse_moe.gate", | |
| "language_model.model.layers.37.block_sparse_moe.gate", | |
| "language_model.model.layers.38.block_sparse_moe.gate", | |
| "language_model.model.layers.39.block_sparse_moe.gate", | |
| "language_model.model.layers.4.block_sparse_moe.gate", | |
| "language_model.model.layers.40.block_sparse_moe.gate", | |
| "language_model.model.layers.41.block_sparse_moe.gate", | |
| "language_model.model.layers.42.block_sparse_moe.gate", | |
| "language_model.model.layers.43.block_sparse_moe.gate", | |
| "language_model.model.layers.44.block_sparse_moe.gate", | |
| "language_model.model.layers.45.block_sparse_moe.gate", | |
| "language_model.model.layers.46.block_sparse_moe.gate", | |
| "language_model.model.layers.47.block_sparse_moe.gate", | |
| "language_model.model.layers.48.block_sparse_moe.gate", | |
| "language_model.model.layers.49.block_sparse_moe.gate", | |
| "language_model.model.layers.5.block_sparse_moe.gate", | |
| "language_model.model.layers.50.block_sparse_moe.gate", | |
| "language_model.model.layers.51.block_sparse_moe.gate", | |
| "language_model.model.layers.52.block_sparse_moe.gate", | |
| "language_model.model.layers.53.block_sparse_moe.gate", | |
| "language_model.model.layers.54.block_sparse_moe.gate", | |
| "language_model.model.layers.55.block_sparse_moe.gate", | |
| "language_model.model.layers.56.block_sparse_moe.gate", | |
| "language_model.model.layers.57.block_sparse_moe.gate", | |
| "language_model.model.layers.58.block_sparse_moe.gate", | |
| "language_model.model.layers.59.block_sparse_moe.gate", | |
| "language_model.model.layers.6.block_sparse_moe.gate", | |
| "language_model.model.layers.7.block_sparse_moe.gate", | |
| "language_model.model.layers.8.block_sparse_moe.gate", | |
| "language_model.model.layers.9.block_sparse_moe.gate", | |
| "lm_head", | |
| "model.embed_tokens", | |
| "multi_modal_projector", | |
| "patch_merge_mlp", | |
| "vision_tower" | |
| ], | |
| "group_size": 16, | |
| "moe_quant_algo": "NVFP4", | |
| "moe_quant_format": "nvfp4-pack-quantized" | |
| }, | |
| "tie_word_embeddings": false, | |
| "dspark_block_size": 7, | |
| "dspark_markov_rank": 16, | |
| "dspark_markov_head_type": "vanilla", | |
| "dspark_aux_gemm_precision": "bf16", | |
| "dspark_noise_token_id": 200058, | |
| "dspark_use_dense_ffn": true, | |
| "dspark_ffn_hidden_size": 12288, | |
| "dspark_target_layer_ids": [ | |
| 3, | |
| 17, | |
| 31, | |
| 45, | |
| 59 | |
| ], | |
| "dspark_config": { | |
| "markov_rank": 256, | |
| "markov_head_type": "vanilla", | |
| "mask_token_id": 200058, | |
| "hybrid_context_fusion": false, | |
| "aux_gemm_precision": "bf16", | |
| "use_dense_ffn": true, | |
| "ffn_hidden_size": 12288 | |
| }, | |
| "dspark_hybrid_context_fusion": false, | |
| "enable_confidence_head": true, | |
| "confidence_head_with_markov": true, | |
| "dflash_config": { | |
| "num_hidden_layers": 5, | |
| "num_target_layers": 60, | |
| "context_source": "target_layers", | |
| "block_size": 7, | |
| "mask_token_id": 200058, | |
| "target_layer_ids": [ | |
| 3, | |
| 17, | |
| 31, | |
| 45, | |
| 59 | |
| ], | |
| "layer_types": [ | |
| "sliding_attention", | |
| "sliding_attention", | |
| "sliding_attention", | |
| "sliding_attention", | |
| "sliding_attention" | |
| ], | |
| "sliding_window": 4096 | |
| }, | |
| "markov_rank": 256, | |
| "markov_head_type": "vanilla", | |
| "hidden_size": 6144, | |
| "vocab_size": 200064, | |
| "num_hidden_layers": 5, | |
| "num_nextn_predict_layers": 0, | |
| "use_mtp": false, | |
| "num_mtp_modules": 0, | |
| "context_len": 1048576, | |
| "encoder_only": false, | |
| "language_only": false, | |
| "output_attentions": false | |
| } | |