Hugging Face's logo

ninagroot
/
Baby-Llama-58M-RUN1

Text Generation
Transformers
Safetensors
llama
Generated from Trainer
text-generation-inference
Model card Files
xet
Community
ninagroot commited on
Commit
a43e481
·
verified ·
1 Parent(s): 5e652fe

ninagroot/babyllamatest

Browse files
Files changed (4) hide show
  1. README.md +44 -24
  2. config.json +1 -1
  3. model.safetensors +2 -2
  4. training_args.bin +1 -1
README.md CHANGED
@@ -13,7 +13,7 @@ should probably proofread and complete it, then remove this comment. -->
13
 
14
  This model is a fine-tuned version of [](https://huggingface.co/) on an unknown dataset.
15
  It achieves the following results on the evaluation set:
16
- - Loss: 3.4868
17
 
18
  ## Model description
19
 
@@ -33,39 +33,59 @@ More information needed
33
 
34
  The following hyperparameters were used during training:
35
  - learning_rate: 0.00025
36
- - train_batch_size: 4
37
  - eval_batch_size: 8
38
  - seed: 42
39
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
40
  - lr_scheduler_type: cosine
41
- - lr_scheduler_warmup_steps: 200
42
- - num_epochs: 20
43
  - mixed_precision_training: Native AMP
44
 
45
  ### Training results
46
 
47
  | Training Loss | Epoch | Step | Validation Loss |
48
  |:-------------:|:-----:|:----:|:---------------:|
49
- | 16.2182 | 1.0 | 393 | 13.7445 |
50
- | 6.1598 | 2.0 | 786 | 6.7784 |
51
- | 5.094 | 3.0 | 1179 | 5.4728 |
52
- | 4.2823 | 4.0 | 1572 | 4.9842 |
53
- | 3.7105 | 5.0 | 1965 | 4.6118 |
54
- | 3.3325 | 6.0 | 2358 | 4.4379 |
55
- | 3.1282 | 7.0 | 2751 | 4.2705 |
56
- | 2.9706 | 8.0 | 3144 | 4.0921 |
57
- | 2.8795 | 9.0 | 3537 | 3.9575 |
58
- | 2.5869 | 10.0 | 3930 | 3.8738 |
59
- | 2.6449 | 11.0 | 4323 | 3.8033 |
60
- | 2.4537 | 12.0 | 4716 | 3.7222 |
61
- | 2.4489 | 13.0 | 5109 | 3.6770 |
62
- | 2.237 | 14.0 | 5502 | 3.6201 |
63
- | 2.2934 | 15.0 | 5895 | 3.5597 |
64
- | 2.2597 | 16.0 | 6288 | 3.5336 |
65
- | 2.2667 | 17.0 | 6681 | 3.5108 |
66
- | 2.2947 | 18.0 | 7074 | 3.4935 |
67
- | 2.1618 | 19.0 | 7467 | 3.4894 |
68
- | 2.2033 | 20.0 | 7860 | 3.4868 |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
69
 
70
 
71
  ### Framework versions
 
13
 
14
  This model is a fine-tuned version of [](https://huggingface.co/) on an unknown dataset.
15
  It achieves the following results on the evaluation set:
16
+ - Loss: 5.5855
17
 
18
  ## Model description
19
 
 
33
 
34
  The following hyperparameters were used during training:
35
  - learning_rate: 0.00025
36
+ - train_batch_size: 32
37
  - eval_batch_size: 8
38
  - seed: 42
39
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
40
  - lr_scheduler_type: cosine
41
+ - lr_scheduler_warmup_steps: 50
42
+ - num_epochs: 40
43
  - mixed_precision_training: Native AMP
44
 
45
  ### Training results
46
 
47
  | Training Loss | Epoch | Step | Validation Loss |
48
  |:-------------:|:-----:|:----:|:---------------:|
49
+ | No log | 1.0 | 7 | 71.5542 |
50
+ | No log | 2.0 | 14 | 62.5591 |
51
+ | 75.1015 | 3.0 | 21 | 52.1433 |
52
+ | 75.1015 | 4.0 | 28 | 41.8504 |
53
+ | 75.1015 | 5.0 | 35 | 35.2569 |
54
+ | 47.6078 | 6.0 | 42 | 29.3735 |
55
+ | 47.6078 | 7.0 | 49 | 25.0324 |
56
+ | 47.6078 | 8.0 | 56 | 21.8563 |
57
+ | 29.8451 | 9.0 | 63 | 18.5148 |
58
+ | 29.8451 | 10.0 | 70 | 15.5228 |
59
+ | 29.8451 | 11.0 | 77 | 13.0948 |
60
+ | 17.3358 | 12.0 | 84 | 11.0128 |
61
+ | 17.3358 | 13.0 | 91 | 9.2921 |
62
+ | 17.3358 | 14.0 | 98 | 8.5328 |
63
+ | 9.565 | 15.0 | 105 | 7.7662 |
64
+ | 9.565 | 16.0 | 112 | 7.5869 |
65
+ | 9.565 | 17.0 | 119 | 7.0970 |
66
+ | 6.5384 | 18.0 | 126 | 6.8532 |
67
+ | 6.5384 | 19.0 | 133 | 6.5811 |
68
+ | 5.5048 | 20.0 | 140 | 6.4781 |
69
+ | 5.5048 | 21.0 | 147 | 6.3671 |
70
+ | 5.5048 | 22.0 | 154 | 6.3537 |
71
+ | 5.0318 | 23.0 | 161 | 6.1682 |
72
+ | 5.0318 | 24.0 | 168 | 5.9982 |
73
+ | 5.0318 | 25.0 | 175 | 5.9669 |
74
+ | 4.71 | 26.0 | 182 | 5.8959 |
75
+ | 4.71 | 27.0 | 189 | 5.8910 |
76
+ | 4.71 | 28.0 | 196 | 5.8447 |
77
+ | 4.5081 | 29.0 | 203 | 5.7656 |
78
+ | 4.5081 | 30.0 | 210 | 5.7277 |
79
+ | 4.5081 | 31.0 | 217 | 5.7041 |
80
+ | 4.3384 | 32.0 | 224 | 5.6611 |
81
+ | 4.3384 | 33.0 | 231 | 5.6682 |
82
+ | 4.3384 | 34.0 | 238 | 5.6317 |
83
+ | 4.2903 | 35.0 | 245 | 5.6118 |
84
+ | 4.2903 | 36.0 | 252 | 5.5995 |
85
+ | 4.2903 | 37.0 | 259 | 5.6007 |
86
+ | 4.2404 | 38.0 | 266 | 5.5819 |
87
+ | 4.2404 | 39.0 | 273 | 5.5834 |
88
+ | 4.1963 | 40.0 | 280 | 5.5855 |
89
 
90
 
91
  ### Framework versions
config.json CHANGED
@@ -24,5 +24,5 @@
24
  "torch_dtype": "float32",
25
  "transformers_version": "4.39.1",
26
  "use_cache": true,
27
- "vocab_size": 32000
28
  }
 
24
  "torch_dtype": "float32",
25
  "transformers_version": "4.39.1",
26
  "use_cache": true,
27
+ "vocab_size": 12198
28
  }
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:18587d669e316d918d29fef1f1c52079e33a2b167f33c3ccd433d7cd84931186
3
- size 298928096
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:03acd465aafbe8e6f46318a9592cd0ae304d19d427c38a460754b2dc18cd506c
3
+ size 217819016
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:42dba10b5115a01d6be8e65830dfd27ac82bb1251d44e3f697c99e87edbab09d
3
  size 4984
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0a404120ec632e6c40885939d6a4d0f4d7e819ebd221abbc8530bab3798a602c
3
  size 4984