antphb commited on
Commit
277b905
1 Parent(s): 8757f58

update model card README.md

Browse files
Files changed (1) hide show
  1. README.md +20 -7
README.md CHANGED
@@ -13,12 +13,7 @@ should probably proofread and complete it, then remove this comment. -->
13
 
14
  This model is a fine-tuned version of [NlpHUST/gpt2-vietnamese](https://huggingface.co/NlpHUST/gpt2-vietnamese) on the None dataset.
15
  It achieves the following results on the evaluation set:
16
- - eval_loss: 2.8435
17
- - eval_runtime: 40.4638
18
- - eval_samples_per_second: 180.68
19
- - eval_steps_per_second: 22.588
20
- - epoch: 2.34
21
- - step: 40000
22
 
23
  ## Model description
24
 
@@ -41,11 +36,29 @@ The following hyperparameters were used during training:
41
  - train_batch_size: 8
42
  - eval_batch_size: 8
43
  - seed: 42
 
 
44
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
45
  - lr_scheduler_type: cosine
46
- - lr_scheduler_warmup_steps: 1000
47
  - num_epochs: 7
48
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
49
  ### Framework versions
50
 
51
  - Transformers 4.30.2
 
13
 
14
  This model is a fine-tuned version of [NlpHUST/gpt2-vietnamese](https://huggingface.co/NlpHUST/gpt2-vietnamese) on the None dataset.
15
  It achieves the following results on the evaluation set:
16
+ - Loss: 2.7322
 
 
 
 
 
17
 
18
  ## Model description
19
 
 
36
  - train_batch_size: 8
37
  - eval_batch_size: 8
38
  - seed: 42
39
+ - gradient_accumulation_steps: 8
40
+ - total_train_batch_size: 64
41
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
42
  - lr_scheduler_type: cosine
43
+ - lr_scheduler_warmup_ratio: 0.2
44
  - num_epochs: 7
45
 
46
+ ### Training results
47
+
48
+ | Training Loss | Epoch | Step | Validation Loss |
49
+ |:-------------:|:-----:|:-----:|:---------------:|
50
+ | 2.9759 | 0.66 | 1400 | 2.6781 |
51
+ | 2.5019 | 1.31 | 2800 | 2.4921 |
52
+ | 2.3352 | 1.97 | 4200 | 2.3726 |
53
+ | 2.0759 | 2.62 | 5600 | 2.3240 |
54
+ | 1.9303 | 3.28 | 7000 | 2.3279 |
55
+ | 1.7867 | 3.93 | 8400 | 2.2556 |
56
+ | 1.5133 | 4.59 | 9800 | 2.3424 |
57
+ | 1.3726 | 5.25 | 11200 | 2.5290 |
58
+ | 1.1925 | 5.9 | 12600 | 2.5132 |
59
+ | 1.0211 | 6.56 | 14000 | 2.7322 |
60
+
61
+
62
  ### Framework versions
63
 
64
  - Transformers 4.30.2