Atomic-Germ commited on
Commit
36c6840
·
verified ·
1 Parent(s): c99d2b4

Upload config.json with huggingface_hub

Browse files
Files changed (1) hide show
  1. config.json +39 -6
config.json CHANGED
@@ -1,7 +1,9 @@
1
  {
2
  "architectures": [
3
- "Qwen3_5MoeForCausalLM"
4
  ],
 
 
5
  "attention_bias": false,
6
  "attention_dropout": 0.0,
7
  "attn_output_gate": true,
@@ -62,8 +64,8 @@
62
  "linear_value_head_dim": 128,
63
  "mamba_ssm_dtype": "float32",
64
  "max_position_embeddings": 262144,
65
- "model_type": "qwen3_5_moe_text",
66
  "moe_intermediate_size": 512,
 
67
  "mtp_num_hidden_layers": 1,
68
  "mtp_use_dedicated_embeddings": false,
69
  "num_attention_heads": 16,
@@ -88,10 +90,41 @@
88
  },
89
  "router_aux_loss_coef": 0.001,
90
  "shared_expert_intermediate_size": 512,
91
- "tie_word_embeddings": false,
92
- "transformers_version": "5.5.4",
93
  "use_cache": true,
94
  "vocab_size": 248320,
95
- "intermediate_size": 512,
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
96
  "flm_version": "1.0.0"
97
- }
 
1
  {
2
  "architectures": [
3
+ "Qwen3_5MoeForConditionalGeneration"
4
  ],
5
+ "image_token_id": 248056,
6
+ "model_type": "qwen3_5_moe",
7
  "attention_bias": false,
8
  "attention_dropout": 0.0,
9
  "attn_output_gate": true,
 
64
  "linear_value_head_dim": 128,
65
  "mamba_ssm_dtype": "float32",
66
  "max_position_embeddings": 262144,
 
67
  "moe_intermediate_size": 512,
68
+ "intermediate_size": 512,
69
  "mtp_num_hidden_layers": 1,
70
  "mtp_use_dedicated_embeddings": false,
71
  "num_attention_heads": 16,
 
90
  },
91
  "router_aux_loss_coef": 0.001,
92
  "shared_expert_intermediate_size": 512,
 
 
93
  "use_cache": true,
94
  "vocab_size": 248320,
95
+ "transformers_version": "4.57.1",
96
+ "video_token_id": 248057,
97
+ "vision_config": {
98
+ "vision_mm_engine_xclbin_name": "vision_mm.xclbin",
99
+ "vision_mha_engine_xclbin_name": "vision_attn.xclbin",
100
+ "QWEN3_6_MOE_PATCH_SIZE": 16,
101
+ "QWEN3_6_MOE_IMAGE_MERGE_SIZE": 2,
102
+ "QWEN3_6_MOE_SPATIAL_MERGE_SIZE": 2,
103
+ "QWEN3_6_MOE_SHORTEST_EDGE": 65536,
104
+ "QWEN3_6_MOE_LONGEST_EDGE": 16777216,
105
+ "QWEN3_6_MOE_VISION_RESCALE_FACTOR": 0.00392156862745098,
106
+ "QWEN3_6_MOE_VISION_RESCALE_IMAGE_MEAN": 0.5,
107
+ "QWEN3_6_MOE_VISION_RESCALE_IMAGE_STD": 0.5,
108
+ "QWEN3_6_MOE_TEMPORAL_PATCH_SIZE": 2,
109
+ "QWEN3_6_MOE_VISION_EMBED_DIM": 1152,
110
+ "QWEN3_6_MOE_VISION_NUM_HEADS": 16,
111
+ "QWEN3_6_MOE_VISION_HEAD_DIM": 72,
112
+ "QWEN3_6_MOE_VISION_MLP_INTERMEDIATE_SIZE": 4304,
113
+ "QWEN3_6_MOE_VISION_NUM_POSITION_EMBEDDINGS": 2304,
114
+ "QWEN3_6_MOE_VISION_NUM_LAYERS": 27,
115
+ "QWEN3_6_MOE_VISION_LAYER_NORM_EPSILON": 1e-06,
116
+ "QWEN3_6_MOE_VISION_OUT_HIDDEN_SIZE": 2048,
117
+ "VISION_MM_TILE_M": 128,
118
+ "VISION_MM_TILE_K": 256,
119
+ "VISION_MM_TILE_N": 64
120
+ },
121
+ "vision_end_token_id": 248054,
122
+ "vision_start_token_id": 248053,
123
+ "vision_model_weight": "vision_weight.q4nx",
124
+ "addr_qk": 9216,
125
+ "addr_kv": 36864,
126
+ "addr_l_begin_mha": 53760,
127
+ "addr_l_end_mha": 25088,
128
+ "addr_kk": 45056,
129
  "flm_version": "1.0.0"
130
+ }