Hugging Face's logo

akashmaggon
/
bertnew-newscategoryclassification-fullmodel-3

Safetensors
Generated from Trainer
Model card Files
xet
Community
akashmaggon commited on
Commit
936102d
·
verified ·
1 Parent(s): 09f9612

Model save

Browse files
Files changed (1) hide show
  1. README.md +4 -51
README.md CHANGED
@@ -14,51 +14,6 @@ should probably proofread and complete it, then remove this comment. -->
14
  # bertnew-newscategoryclassification-fullmodel-3
15
 
16
  This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the None dataset.
17
- It achieves the following results on the evaluation set:
18
- - Loss: 1.0352
19
- - Class 0 Accuracy: 0.3385
20
- - Class 1 Accuracy: 0.4762
21
- - Class 2 Accuracy: 0.5138
22
- - Class 3 Accuracy: 0.5707
23
- - Class 4 Accuracy: 0.6970
24
- - Class 5 Accuracy: 0.5942
25
- - Class 6 Accuracy: 0.6654
26
- - Class 7 Accuracy: 0.6994
27
- - Class 8 Accuracy: 0.8560
28
- - Class 9 Accuracy: 0.6372
29
- - Class 10 Accuracy: 0.7861
30
- - Class 11 Accuracy: 0.5635
31
- - Class 12 Accuracy: 0.5449
32
- - Class 13 Accuracy: 0.7592
33
- - Class 14 Accuracy: 0.5952
34
- - Class 15 Accuracy: 0.5463
35
- - Class 16 Accuracy: 0.5105
36
- - Class 17 Accuracy: 0.8383
37
- - Class 18 Accuracy: 0.5116
38
- - Class 19 Accuracy: 0.6855
39
- - Class 20 Accuracy: 0.644
40
- - Class 21 Accuracy: 0.5333
41
- - Class 22 Accuracy: 0.7814
42
- - Class 23 Accuracy: 0.5637
43
- - Class 24 Accuracy: 0.8425
44
- - Class 25 Accuracy: 0.7691
45
- - Class 26 Accuracy: 0.6534
46
- - Class 27 Accuracy: 0.5217
47
- - Class 28 Accuracy: 0.8303
48
- - Class 29 Accuracy: 0.6194
49
- - Class 30 Accuracy: 0.8817
50
- - Class 31 Accuracy: 0.5521
51
- - Class 32 Accuracy: 0.5693
52
- - Class 33 Accuracy: 0.5931
53
- - Class 34 Accuracy: 0.8481
54
- - Class 35 Accuracy: 0.5706
55
- - Class 36 Accuracy: 0.8435
56
- - Class 37 Accuracy: 0.5423
57
- - Class 38 Accuracy: 0.8128
58
- - Class 39 Accuracy: 0.5385
59
- - Class 40 Accuracy: 0.4913
60
- - Class 41 Accuracy: 0.7120
61
- - Overall Accuracy: 0.7142
62
 
63
  ## Model description
64
 
@@ -84,16 +39,14 @@ The following hyperparameters were used during training:
84
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
85
  - lr_scheduler_type: cosine
86
  - lr_scheduler_warmup_steps: 600
87
- - num_epochs: 3
88
  - mixed_precision_training: Native AMP
89
 
90
  ### Training results
91
 
92
- | Training Loss | Epoch | Step | Validation Loss | Class 0 Accuracy | Class 1 Accuracy | Class 2 Accuracy | Class 3 Accuracy | Class 4 Accuracy | Class 5 Accuracy | Class 6 Accuracy | Class 7 Accuracy | Class 8 Accuracy | Class 9 Accuracy | Class 10 Accuracy | Class 11 Accuracy | Class 12 Accuracy | Class 13 Accuracy | Class 14 Accuracy | Class 15 Accuracy | Class 16 Accuracy | Class 17 Accuracy | Class 18 Accuracy | Class 19 Accuracy | Class 20 Accuracy | Class 21 Accuracy | Class 22 Accuracy | Class 23 Accuracy | Class 24 Accuracy | Class 25 Accuracy | Class 26 Accuracy | Class 27 Accuracy | Class 28 Accuracy | Class 29 Accuracy | Class 30 Accuracy | Class 31 Accuracy | Class 32 Accuracy | Class 33 Accuracy | Class 34 Accuracy | Class 35 Accuracy | Class 36 Accuracy | Class 37 Accuracy | Class 38 Accuracy | Class 39 Accuracy | Class 40 Accuracy | Class 41 Accuracy | Overall Accuracy |
93
- |:-------------:|:-----:|:----:|:---------------:|:----------------:|:----------------:|:----------------:|:----------------:|:----------------:|:----------------:|:----------------:|:----------------:|:----------------:|:----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:----------------:|
94
- | 1.2637 | 1.0 | 2213 | 1.1140 | 0.2077 | 0.2993 | 0.4523 | 0.6152 | 0.6742 | 0.5471 | 0.6693 | 0.7362 | 0.8148 | 0.6637 | 0.7754 | 0.6667 | 0.5 | 0.7740 | 0.5429 | 0.2996 | 0.4494 | 0.8416 | 0.4942 | 0.6210 | 0.636 | 0.4593 | 0.6305 | 0.5908 | 0.8393 | 0.6928 | 0.6932 | 0.5963 | 0.8277 | 0.6418 | 0.8846 | 0.4323 | 0.4380 | 0.7379 | 0.8223 | 0.4529 | 0.7519 | 0.5192 | 0.82 | 0.5105 | 0.1107 | 0.5380 | 0.6827 |
95
- | 0.9389 | 2.0 | 4426 | 1.0106 | 0.3154 | 0.4218 | 0.44 | 0.5471 | 0.6894 | 0.5785 | 0.7087 | 0.7117 | 0.8601 | 0.6814 | 0.7663 | 0.5556 | 0.5256 | 0.7641 | 0.5762 | 0.5066 | 0.4979 | 0.8482 | 0.4767 | 0.6935 | 0.608 | 0.5926 | 0.7610 | 0.5366 | 0.8556 | 0.7843 | 0.6023 | 0.5590 | 0.8538 | 0.6269 | 0.8773 | 0.5208 | 0.5401 | 0.4724 | 0.8567 | 0.5647 | 0.8282 | 0.5385 | 0.832 | 0.5280 | 0.5398 | 0.6848 | 0.7092 |
96
- | 0.6831 | 3.0 | 6639 | 1.0352 | 0.3385 | 0.4762 | 0.5138 | 0.5707 | 0.6970 | 0.5942 | 0.6654 | 0.6994 | 0.8560 | 0.6372 | 0.7861 | 0.5635 | 0.5449 | 0.7592 | 0.5952 | 0.5463 | 0.5105 | 0.8383 | 0.5116 | 0.6855 | 0.644 | 0.5333 | 0.7814 | 0.5637 | 0.8425 | 0.7691 | 0.6534 | 0.5217 | 0.8303 | 0.6194 | 0.8817 | 0.5521 | 0.5693 | 0.5931 | 0.8481 | 0.5706 | 0.8435 | 0.5423 | 0.8128 | 0.5385 | 0.4913 | 0.7120 | 0.7142 |
97
 
98
 
99
  ### Framework versions
 
14
  # bertnew-newscategoryclassification-fullmodel-3
15
 
16
  This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the None dataset.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
17
 
18
  ## Model description
19
 
 
39
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
40
  - lr_scheduler_type: cosine
41
  - lr_scheduler_warmup_steps: 600
42
+ - training_steps: 10
43
  - mixed_precision_training: Native AMP
44
 
45
  ### Training results
46
 
47
+ | Training Loss | Epoch | Step | Validation Loss | Class 0 Accuracy | Class 1 Accuracy | Class 2 Accuracy | Class 3 Accuracy | Class 4 Accuracy | Class 5 Accuracy | Class 6 Accuracy | Class 7 Accuracy | Class 8 Accuracy | Class 9 Accuracy | Class 10 Accuracy | Class 11 Accuracy | Class 12 Accuracy | Class 13 Accuracy | Class 14 Accuracy | Class 15 Accuracy | Class 16 Accuracy | Class 17 Accuracy | Class 18 Accuracy | Class 19 Accuracy | Class 20 Accuracy | Class 21 Accuracy | Class 22 Accuracy | Class 23 Accuracy | Class 24 Accuracy | Class 25 Accuracy | Class 26 Accuracy | Class 27 Accuracy | Class 28 Accuracy | Class 29 Accuracy | Class 30 Accuracy | Class 31 Accuracy | Class 32 Accuracy | Class 33 Accuracy | Class 34 Accuracy | Class 35 Accuracy | Class 36 Accuracy | Class 37 Accuracy | Class 38 Accuracy | Class 39 Accuracy | Class 40 Accuracy | Class 41 Accuracy | Overall Accuracy |
48
+ |:-------------:|:------:|:----:|:---------------:|:----------------:|:----------------:|:----------------:|:----------------:|:----------------:|:----------------:|:----------------:|:----------------:|:----------------:|:----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:-----------------:|:----------------:|
49
+ | No log | 0.0045 | 10 | 3.7525 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0190 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.9864 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0 | 0.0237 |
 
 
50
 
51
 
52
  ### Framework versions