Hugging Face's logo

flytech
/
platistil

Safetensors
Generated from Trainer
Model card Files Community
1
sirr commited on
Commit
dc765e3
•
1 Parent(s): e6a8afb

End of training

Browse files
Files changed (4) hide show
  1. README.md +6 -4
  2. adapter_config.json +73 -1
  3. adapter_model.bin +3 -0
  4. training_args.bin +1 -1
README.md CHANGED
@@ -32,13 +32,15 @@ More information needed
32
  ### Training hyperparameters
33
 
34
  The following hyperparameters were used during training:
35
- - learning_rate: 0.01
36
- - train_batch_size: 32
37
- - eval_batch_size: 32
38
  - seed: 42
 
 
39
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
40
  - lr_scheduler_type: constant
41
- - lr_scheduler_warmup_ratio: 0.1
42
  - num_epochs: 3.0
43
 
44
  ### Training results
 
32
  ### Training hyperparameters
33
 
34
  The following hyperparameters were used during training:
35
+ - learning_rate: 0.001
36
+ - train_batch_size: 16
37
+ - eval_batch_size: 16
38
  - seed: 42
39
+ - gradient_accumulation_steps: 2
40
+ - total_train_batch_size: 32
41
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
42
  - lr_scheduler_type: constant
43
+ - lr_scheduler_warmup_ratio: 0.05
44
  - num_epochs: 3.0
45
 
46
  ### Training results
adapter_config.json CHANGED
@@ -15,29 +15,101 @@
15
  "revision": null,
16
  "target_modules": [
17
  "transformer.h.0.attn.c_attn",
 
18
  "transformer.h.1.attn.c_attn",
 
19
  "transformer.h.2.attn.c_attn",
 
20
  "transformer.h.3.attn.c_attn",
 
21
  "transformer.h.4.attn.c_attn",
 
22
  "transformer.h.5.attn.c_attn",
 
23
  "transformer.h.6.attn.c_attn",
 
24
  "transformer.h.7.attn.c_attn",
 
25
  "transformer.h.8.attn.c_attn",
 
26
  "transformer.h.9.attn.c_attn",
 
27
  "transformer.h.10.attn.c_attn",
 
28
  "transformer.h.11.attn.c_attn",
 
29
  "transformer.h.12.attn.c_attn",
 
30
  "transformer.h.13.attn.c_attn",
 
31
  "transformer.h.14.attn.c_attn",
 
32
  "transformer.h.15.attn.c_attn",
 
33
  "transformer.h.16.attn.c_attn",
 
34
  "transformer.h.17.attn.c_attn",
 
35
  "transformer.h.18.attn.c_attn",
 
36
  "transformer.h.19.attn.c_attn",
 
37
  "transformer.h.20.attn.c_attn",
 
38
  "transformer.h.21.attn.c_attn",
 
39
  "transformer.h.22.attn.c_attn",
40
- "transformer.h.23.attn.c_attn"
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
41
  ],
42
  "task_type": "CAUSAL_LM"
43
  }
 
15
  "revision": null,
16
  "target_modules": [
17
  "transformer.h.0.attn.c_attn",
18
+ "transformer.h.0.attn.c_proj",
19
  "transformer.h.1.attn.c_attn",
20
+ "transformer.h.1.attn.c_proj",
21
  "transformer.h.2.attn.c_attn",
22
+ "transformer.h.2.attn.c_proj",
23
  "transformer.h.3.attn.c_attn",
24
+ "transformer.h.3.attn.c_proj",
25
  "transformer.h.4.attn.c_attn",
26
+ "transformer.h.4.attn.c_proj",
27
  "transformer.h.5.attn.c_attn",
28
+ "transformer.h.5.attn.c_proj",
29
  "transformer.h.6.attn.c_attn",
30
+ "transformer.h.6.attn.c_proj",
31
  "transformer.h.7.attn.c_attn",
32
+ "transformer.h.7.attn.c_proj",
33
  "transformer.h.8.attn.c_attn",
34
+ "transformer.h.8.attn.c_proj",
35
  "transformer.h.9.attn.c_attn",
36
+ "transformer.h.9.attn.c_proj",
37
  "transformer.h.10.attn.c_attn",
38
+ "transformer.h.10.attn.c_proj",
39
  "transformer.h.11.attn.c_attn",
40
+ "transformer.h.11.attn.c_proj",
41
  "transformer.h.12.attn.c_attn",
42
+ "transformer.h.12.attn.c_proj",
43
  "transformer.h.13.attn.c_attn",
44
+ "transformer.h.13.attn.c_proj",
45
  "transformer.h.14.attn.c_attn",
46
+ "transformer.h.14.attn.c_proj",
47
  "transformer.h.15.attn.c_attn",
48
+ "transformer.h.15.attn.c_proj",
49
  "transformer.h.16.attn.c_attn",
50
+ "transformer.h.16.attn.c_proj",
51
  "transformer.h.17.attn.c_attn",
52
+ "transformer.h.17.attn.c_proj",
53
  "transformer.h.18.attn.c_attn",
54
+ "transformer.h.18.attn.c_proj",
55
  "transformer.h.19.attn.c_attn",
56
+ "transformer.h.19.attn.c_proj",
57
  "transformer.h.20.attn.c_attn",
58
+ "transformer.h.20.attn.c_proj",
59
  "transformer.h.21.attn.c_attn",
60
+ "transformer.h.21.attn.c_proj",
61
  "transformer.h.22.attn.c_attn",
62
+ "transformer.h.22.attn.c_proj",
63
+ "transformer.h.23.attn.c_attn",
64
+ "transformer.h.23.attn.c_proj",
65
+ "transformer.h.0.mlp.c_fc",
66
+ "transformer.h.0.mlp.c_proj",
67
+ "transformer.h.1.mlp.c_fc",
68
+ "transformer.h.1.mlp.c_proj",
69
+ "transformer.h.2.mlp.c_fc",
70
+ "transformer.h.2.mlp.c_proj",
71
+ "transformer.h.3.mlp.c_fc",
72
+ "transformer.h.3.mlp.c_proj",
73
+ "transformer.h.4.mlp.c_fc",
74
+ "transformer.h.4.mlp.c_proj",
75
+ "transformer.h.5.mlp.c_fc",
76
+ "transformer.h.5.mlp.c_proj",
77
+ "transformer.h.6.mlp.c_fc",
78
+ "transformer.h.6.mlp.c_proj",
79
+ "transformer.h.7.mlp.c_fc",
80
+ "transformer.h.7.mlp.c_proj",
81
+ "transformer.h.8.mlp.c_fc",
82
+ "transformer.h.8.mlp.c_proj",
83
+ "transformer.h.9.mlp.c_fc",
84
+ "transformer.h.9.mlp.c_proj",
85
+ "transformer.h.10.mlp.c_fc",
86
+ "transformer.h.10.mlp.c_proj",
87
+ "transformer.h.11.mlp.c_fc",
88
+ "transformer.h.11.mlp.c_proj",
89
+ "transformer.h.12.mlp.c_fc",
90
+ "transformer.h.12.mlp.c_proj",
91
+ "transformer.h.13.mlp.c_fc",
92
+ "transformer.h.13.mlp.c_proj",
93
+ "transformer.h.14.mlp.c_fc",
94
+ "transformer.h.14.mlp.c_proj",
95
+ "transformer.h.15.mlp.c_fc",
96
+ "transformer.h.15.mlp.c_proj",
97
+ "transformer.h.16.mlp.c_fc",
98
+ "transformer.h.16.mlp.c_proj",
99
+ "transformer.h.17.mlp.c_fc",
100
+ "transformer.h.17.mlp.c_proj",
101
+ "transformer.h.18.mlp.c_fc",
102
+ "transformer.h.18.mlp.c_proj",
103
+ "transformer.h.19.mlp.c_fc",
104
+ "transformer.h.19.mlp.c_proj",
105
+ "transformer.h.20.mlp.c_fc",
106
+ "transformer.h.20.mlp.c_proj",
107
+ "transformer.h.21.mlp.c_fc",
108
+ "transformer.h.21.mlp.c_proj",
109
+ "transformer.h.22.mlp.c_fc",
110
+ "transformer.h.22.mlp.c_proj",
111
+ "transformer.h.23.mlp.c_fc",
112
+ "transformer.h.23.mlp.c_proj"
113
  ],
114
  "task_type": "CAUSAL_LM"
115
  }
adapter_model.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e349a5bc088361e0b11c73ae5d7f5a61fec1281b5a5694af06686fc88025110d
3
+ size 12650829
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:7aec30b908f8713803bd37cf76f06b0db58fd5199bc91d0e2a3973e0a745a8f0
3
  size 4091
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3881ce2a7bb2514addd63f19b658ec27b7bf1437c9eac577576e4535e953e79c
3
  size 4091