Hugging Face's logo

NasimB
/
gpt2-dp-mod-datasets-txt-processing

Text Generation
Transformers
PyTorch
gpt2
Generated from Trainer
text-generation-inference
Model card Files
xet
 Community
1
NasimB commited on
Commit
011baa3
·
1 Parent(s): 3d88567

update model card README.md

Browse files
Files changed (1) hide show
  1. README.md +23 -27
README.md CHANGED
@@ -16,7 +16,7 @@ should probably proofread and complete it, then remove this comment. -->
16
 
17
  This model is a fine-tuned version of [gpt2](https://huggingface.co/gpt2) on the generator dataset.
18
  It achieves the following results on the evaluation set:
19
- - Loss: 4.3761
20
 
21
  ## Model description
22
 
@@ -42,38 +42,34 @@ The following hyperparameters were used during training:
42
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
43
  - lr_scheduler_type: cosine
44
  - lr_scheduler_warmup_steps: 1000
45
- - num_epochs: 7
46
  - mixed_precision_training: Native AMP
47
 
48
  ### Training results
49
 
50
  | Training Loss | Epoch | Step | Validation Loss |
51
  |:-------------:|:-----:|:-----:|:---------------:|
52
- | 6.6775 | 0.28 | 500 | 5.6723 |
53
- | 5.3325 | 0.55 | 1000 | 5.2436 |
54
- | 4.9992 | 0.83 | 1500 | 5.0034 |
55
- | 4.7425 | 1.1 | 2000 | 4.8449 |
56
- | 4.5796 | 1.38 | 2500 | 4.7369 |
57
- | 4.4803 | 1.66 | 3000 | 4.6477 |
58
- | 4.4001 | 1.93 | 3500 | 4.5614 |
59
- | 4.1974 | 2.21 | 4000 | 4.5262 |
60
- | 4.1371 | 2.48 | 4500 | 4.4672 |
61
- | 4.1121 | 2.76 | 5000 | 4.4112 |
62
- | 4.036 | 3.04 | 5500 | 4.3790 |
63
- | 3.8293 | 3.31 | 6000 | 4.3736 |
64
- | 3.8346 | 3.59 | 6500 | 4.3358 |
65
- | 3.8169 | 3.86 | 7000 | 4.3052 |
66
- | 3.6622 | 4.14 | 7500 | 4.3171 |
67
- | 3.546 | 4.42 | 8000 | 4.3083 |
68
- | 3.5542 | 4.69 | 8500 | 4.2896 |
69
- | 3.5409 | 4.97 | 9000 | 4.2722 |
70
- | 3.3117 | 5.24 | 9500 | 4.2998 |
71
- | 3.298 | 5.52 | 10000 | 4.2969 |
72
- | 3.2941 | 5.79 | 10500 | 4.2904 |
73
- | 3.24 | 6.07 | 11000 | 4.2975 |
74
- | 3.1365 | 6.35 | 11500 | 4.3027 |
75
- | 3.1357 | 6.62 | 12000 | 4.3033 |
76
- | 3.1363 | 6.9 | 12500 | 4.3034 |
77
 
78
 
79
  ### Framework versions
 
16
 
17
  This model is a fine-tuned version of [gpt2](https://huggingface.co/gpt2) on the generator dataset.
18
  It achieves the following results on the evaluation set:
19
+ - Loss: 4.3134
20
 
21
  ## Model description
22
 
 
42
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
43
  - lr_scheduler_type: cosine
44
  - lr_scheduler_warmup_steps: 1000
45
+ - num_epochs: 6
46
  - mixed_precision_training: Native AMP
47
 
48
  ### Training results
49
 
50
  | Training Loss | Epoch | Step | Validation Loss |
51
  |:-------------:|:-----:|:-----:|:---------------:|
52
+ | 6.694 | 0.28 | 500 | 5.6630 |
53
+ | 5.3512 | 0.55 | 1000 | 5.2372 |
54
+ | 5.0119 | 0.83 | 1500 | 4.9763 |
55
+ | 4.767 | 1.1 | 2000 | 4.8279 |
56
+ | 4.5688 | 1.38 | 2500 | 4.7089 |
57
+ | 4.4767 | 1.65 | 3000 | 4.6105 |
58
+ | 4.3893 | 1.93 | 3500 | 4.5220 |
59
+ | 4.1792 | 2.21 | 4000 | 4.4846 |
60
+ | 4.1211 | 2.48 | 4500 | 4.4302 |
61
+ | 4.08 | 2.76 | 5000 | 4.3699 |
62
+ | 4.0158 | 3.03 | 5500 | 4.3318 |
63
+ | 3.7873 | 3.31 | 6000 | 4.3214 |
64
+ | 3.7888 | 3.58 | 6500 | 4.2912 |
65
+ | 3.7709 | 3.86 | 7000 | 4.2590 |
66
+ | 3.6276 | 4.13 | 7500 | 4.2642 |
67
+ | 3.4947 | 4.41 | 8000 | 4.2579 |
68
+ | 3.4884 | 4.69 | 8500 | 4.2439 |
69
+ | 3.4836 | 4.96 | 9000 | 4.2315 |
70
+ | 3.3261 | 5.24 | 9500 | 4.2430 |
71
+ | 3.2961 | 5.51 | 10000 | 4.2427 |
72
+ | 3.2947 | 5.79 | 10500 | 4.2419 |
 
 
 
 
73
 
74
 
75
  ### Framework versions