Hugging Face's logo

tencent
/
Tencent-Hunyuan-Large

Text Generation
Transformers
Safetensors
English
Inference Endpoints
Model card Files Community
8
Tencent-Hunyuan-Large
File size: 3,781 Bytes 
c9e801e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
 
{
  "_name_or_path": "/apdcephfs_gy2/share_302508627/woodchenwu/hunyuan-L-instruct-128K-1102online-HF/",
  "architectures": [
    "HunYuanForCausalLM"
  ],
  "attention_bias": false,
  "attention_dropout": 0.0,
  "auto_map": {
    "AutoConfig": "configuration_hunyuan.HunYuanConfig",
    "AutoModel": "modeling_hunyuan.HunyuanModel",
    "AutoModelForCausalLM": "modeling_hunyuan.HunYuanForCausalLM"
  },
  "bos_token_id": 1,
  "capacity_factor": 1.0,
  "cla_share_factor": 2,
  "eos_token_id": 2,
  "hidden_act": "silu",
  "hidden_size": 6400,
  "initializer_range": 0.02,
  "intermediate_size": 18304,
  "max_position_embeddings": 131072,
  "model_type": "hunyuan",
  "moe_drop_tokens": false,
  "moe_random_routing_dropped_token": false,
  "moe_topk": 1,
  "num_attention_heads": 80,
  "num_experts": 16,
  "num_hidden_layers": 64,
  "num_key_value_heads": 8,
  "num_shared_expert": 1,
  "pad_token_id": 0,
  "pretraining_tp": 1,
  "quantization_config": {
    "activation_scheme": "static",
    "ignored_layers": [
      "model.layers.38.mlp.gate.wg",
      "model.layers.60.mlp.gate.wg",
      "model.layers.20.mlp.gate.wg",
      "model.layers.11.mlp.gate.wg",
      "model.layers.37.mlp.gate.wg",
      "model.layers.51.mlp.gate.wg",
      "model.layers.54.mlp.gate.wg",
      "model.layers.49.mlp.gate.wg",
      "model.layers.59.mlp.gate.wg",
      "lm_head",
      "model.layers.61.mlp.gate.wg",
      "model.layers.50.mlp.gate.wg",
      "model.layers.26.mlp.gate.wg",
      "model.layers.34.mlp.gate.wg",
      "model.layers.29.mlp.gate.wg",
      "model.layers.2.mlp.gate.wg",
      "model.layers.58.mlp.gate.wg",
      "model.layers.23.mlp.gate.wg",
      "model.layers.16.mlp.gate.wg",
      "model.layers.4.mlp.gate.wg",
      "model.layers.31.mlp.gate.wg",
      "model.layers.45.mlp.gate.wg",
      "model.layers.56.mlp.gate.wg",
      "model.layers.3.mlp.gate.wg",
      "model.layers.19.mlp.gate.wg",
      "model.layers.14.mlp.gate.wg",
      "model.layers.8.mlp.gate.wg",
      "model.layers.41.mlp.gate.wg",
      "model.layers.53.mlp.gate.wg",
      "model.layers.46.mlp.gate.wg",
      "model.layers.7.mlp.gate.wg",
      "model.layers.25.mlp.gate.wg",
      "model.layers.6.mlp.gate.wg",
      "model.layers.1.mlp.gate.wg",
      "model.layers.10.mlp.gate.wg",
      "model.layers.12.mlp.gate.wg",
      "model.layers.15.mlp.gate.wg",
      "model.layers.18.mlp.gate.wg",
      "model.layers.57.mlp.gate.wg",
      "model.layers.63.mlp.gate.wg",
      "model.layers.55.mlp.gate.wg",
      "model.layers.40.mlp.gate.wg",
      "model.layers.32.mlp.gate.wg",
      "model.layers.39.mlp.gate.wg",
      "model.layers.62.mlp.gate.wg",
      "model.layers.28.mlp.gate.wg",
      "model.layers.30.mlp.gate.wg",
      "model.layers.27.mlp.gate.wg",
      "model.layers.22.mlp.gate.wg",
      "model.layers.5.mlp.gate.wg",
      "model.layers.21.mlp.gate.wg",
      "model.layers.36.mlp.gate.wg",
      "model.layers.48.mlp.gate.wg",
      "model.layers.42.mlp.gate.wg",
      "model.layers.44.mlp.gate.wg",
      "model.layers.52.mlp.gate.wg",
      "model.layers.0.mlp.gate.wg",
      "model.layers.13.mlp.gate.wg",
      "model.layers.17.mlp.gate.wg",
      "model.layers.24.mlp.gate.wg",
      "model.layers.33.mlp.gate.wg",
      "model.layers.35.mlp.gate.wg",
      "model.layers.43.mlp.gate.wg",
      "model.layers.47.mlp.gate.wg",
      "model.layers.9.mlp.gate.wg"
    ],
    "quant_method": "fp8"
  },
  "rms_norm_eps": 1e-05,
  "rope_scaling": {
    "alpha": 1000.0,
    "factor": 1.0,
    "type": "dynamic"
  },
  "rope_theta": 10000.0,
  "tie_word_embeddings": true,
  "torch_dtype": "bfloat16",
  "transformers_version": "4.44.2",
  "use_cache": true,
  "use_cla": true,
  "use_mixed_mlp_moe": true,
  "use_qk_norm": true,
  "vocab_size": 129024
}