Hugging Face's logo

BytedTsinghua-SIA
/
Enigmata-Qwen2.5-32B

Safetensors
English
Chinese
qwen2
Model card Files Community
2
jiangjiechen commited on
Commit
abe8c97
·
1 Parent(s): 8c93358

Upload folder using huggingface_hub

Browse files
Files changed (21) hide show
  1. config.json +9 -18
  2. generation_config.json +11 -4
  3. model-00001-of-00018.safetensors +3 -0
  4. model-00002-of-00018.safetensors +3 -0
  5. model-00003-of-00018.safetensors +3 -0
  6. model-00004-of-00018.safetensors +3 -0
  7. model-00005-of-00018.safetensors +3 -0
  8. model-00006-of-00018.safetensors +3 -0
  9. model-00007-of-00018.safetensors +3 -0
  10. model-00008-of-00018.safetensors +3 -0
  11. model-00009-of-00018.safetensors +3 -0
  12. model-00010-of-00018.safetensors +3 -0
  13. model-00011-of-00018.safetensors +3 -0
  14. model-00012-of-00018.safetensors +3 -0
  15. model-00013-of-00018.safetensors +3 -0
  16. model-00014-of-00018.safetensors +3 -0
  17. model-00015-of-00018.safetensors +3 -0
  18. model-00016-of-00018.safetensors +3 -0
  19. model-00017-of-00018.safetensors +3 -0
  20. model-00018-of-00018.safetensors +3 -0
  21. model.safetensors.index.json +0 -0
config.json CHANGED
@@ -1,37 +1,28 @@
1
  {
2
- "_moe_implementation": "fused",
3
- "_name_or_path": "/tmp/08513abd184b0668398ede78b54d60db/huggingface",
4
  "architectures": [
5
- "P6DenseForCausalLM"
6
  ],
7
- "attention_bias": true,
8
  "attention_dropout": 0.0,
9
- "attention_out_bias": true,
10
- "bos_token_id": 151644,
11
- "embd_pdrop": 0.0,
12
  "eos_token_id": 151645,
13
  "hidden_act": "silu",
14
  "hidden_size": 5120,
15
- "initializer_range": 0.01613743060919757,
16
  "intermediate_size": 27648,
17
- "layer_norm_eps": null,
18
  "max_position_embeddings": 32768,
19
- "mlp_bias": false,
20
- "model_type": "seed_p6dense",
21
  "num_attention_heads": 40,
22
  "num_hidden_layers": 64,
23
  "num_key_value_heads": 8,
24
- "pad_token_id": 151643,
25
- "resid_pdrop": 0.0,
26
  "rms_norm_eps": 1e-06,
27
- "rope_scaling": {
28
- "factor": 1,
29
- "rope_type": "default"
30
- },
31
  "rope_theta": 1000000.0,
 
32
  "tie_word_embeddings": false,
33
  "torch_dtype": "bfloat16",
34
- "transformers_version": "4.47.1",
35
  "use_cache": true,
 
36
  "vocab_size": 152064
37
  }
 
1
  {
 
 
2
  "architectures": [
3
+ "Qwen2ForCausalLM"
4
  ],
 
5
  "attention_dropout": 0.0,
6
+ "bos_token_id": 151643,
 
 
7
  "eos_token_id": 151645,
8
  "hidden_act": "silu",
9
  "hidden_size": 5120,
10
+ "initializer_range": 0.02,
11
  "intermediate_size": 27648,
 
12
  "max_position_embeddings": 32768,
13
+ "max_window_layers": 70,
14
+ "model_type": "qwen2",
15
  "num_attention_heads": 40,
16
  "num_hidden_layers": 64,
17
  "num_key_value_heads": 8,
 
 
18
  "rms_norm_eps": 1e-06,
19
+ "rope_scaling": null,
 
 
 
20
  "rope_theta": 1000000.0,
21
+ "sliding_window": 131072,
22
  "tie_word_embeddings": false,
23
  "torch_dtype": "bfloat16",
24
+ "transformers_version": "4.51.3",
25
  "use_cache": true,
26
+ "use_sliding_window": false,
27
  "vocab_size": 152064
28
  }
generation_config.json CHANGED
@@ -1,7 +1,14 @@
1
  {
2
- "_from_model_config": true,
3
- "bos_token_id": 151644,
4
- "eos_token_id": 151645,
 
 
 
5
  "pad_token_id": 151643,
6
- "transformers_version": "4.47.1"
 
 
 
 
7
  }
 
1
  {
2
+ "bos_token_id": 151643,
3
+ "do_sample": true,
4
+ "eos_token_id": [
5
+ 151645,
6
+ 151643
7
+ ],
8
  "pad_token_id": 151643,
9
+ "repetition_penalty": 1.05,
10
+ "temperature": 0.7,
11
+ "top_k": 20,
12
+ "top_p": 0.8,
13
+ "transformers_version": "4.51.3"
14
  }
model-00001-of-00018.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5edb2392f48755ecd48adfa08fc70709c7ac8f357aa97f0aadb3d8f96f495b6b
3
+ size 3633403488
model-00002-of-00018.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ca441d6c3cd7b1dbb179bbd7f8dd85ba3da59d8e85be3f8b29c206c7b51f7fec
3
+ size 3848418568
model-00003-of-00018.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d64cc4165777c2f5c24037d0a5b320853d1fb49050f4d77df6e93a35d1264213
3
+ size 3890359432
model-00004-of-00018.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8eca2c4f578940de0c7ed6513669cbb8aff1303e02094cd7aa3116638ef5d7ff
3
+ size 3890359488
model-00005-of-00018.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f513361f1e6307402821df44661f787d15061e46040f2739d39eca5791dbd2a8
3
+ size 3848408256
model-00006-of-00018.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:bd04e73731a50988b7cdfbe6e1c7bb986db411910f64d8f8a4c4793093d601db
3
+ size 3617711168
model-00007-of-00018.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:1cf3edd704e5dba77bfa79ca4bc7fccb9853f213aadfff5dd057c4c59f1822bb
3
+ size 3617731904
model-00008-of-00018.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c04bf406a5648460d6ac7a39859c6120fbf38797fb6606fdc62a622f1722f37b
3
+ size 3617731896
model-00009-of-00018.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2b9b9db48cd34d92a8b31733faed8ed7ea9dfd47b199da303589862eb1a8a206
3
+ size 3848418616
model-00010-of-00018.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0ce20cfb9134489dc1ca99486407399a4774cd36e07d11bec7dcc38c5901f38b
3
+ size 3890359496
model-00011-of-00018.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3b6355fef7bcce07f6c34fa2a2c91edb2b83225e92a8a9488044691fd67c984f
3
+ size 3890359488
model-00012-of-00018.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ef8027c404de6d881b371dc1a2098677a49c9d6d3ce924f4c85722c1e195c925
3
+ size 3848408256
model-00013-of-00018.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:276f4716d448d01b871d2b16a19baf991cca68357dbf5f93da17e0f6fd5e6952
3
+ size 3617711168
model-00014-of-00018.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:515549e49dc5df9b94dc7c1de7c47d59085edec64dadaf0109ff5cd7b2666ac6
3
+ size 3617731904
model-00015-of-00018.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:87a2a91fba6146433c5f6abad94fe18361b467a7af0553d9bcf65f295a726cfc
3
+ size 3617731896
model-00016-of-00018.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ebed8d541e9de007b28891547cf0fe2dbb6c1fdef374b268d4cd049d64a1c702
3
+ size 3848418616
model-00017-of-00018.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:94b75fc24621c193e0ea98458e3150e7e79e08bdff20e1395931593b0096b68f
3
+ size 3827442520
model-00018-of-00018.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c2dab1e5cd953caf2c45870ae7c12c2d08f83845a2be901fd6c344c2849d269c
3
+ size 1557135488
model.safetensors.index.json CHANGED
The diff for this file is too large to render. See raw diff