Hugging Face's logo

theojolliffe
/
bart-cnn-pubmed-arxiv-pubmed-arxiv-arxiv-earlystopping

Text2Text Generation
Transformers
PyTorch
TensorBoard
bart
generated_from_trainer
Inference Endpoints
Model card Files Metrics Community
theojolliffe commited on
Commit
b56d23b
•
1 Parent(s): abcd713

update model card README.md

Browse files
Files changed (1) hide show
  1. README.md +19 -18
README.md CHANGED
@@ -16,11 +16,11 @@ should probably proofread and complete it, then remove this comment. -->
16
 
17
  This model is a fine-tuned version of [theojolliffe/bart-cnn-pubmed-arxiv-pubmed-arxiv-arxiv](https://huggingface.co/theojolliffe/bart-cnn-pubmed-arxiv-pubmed-arxiv-arxiv) on an unknown dataset.
18
  It achieves the following results on the evaluation set:
19
- - Loss: 0.7951
20
- - Rouge1: 53.9498
21
- - Rouge2: 35.4838
22
- - Rougel: 38.9904
23
- - Rougelsum: 51.3848
24
  - Gen Len: 142.0
25
 
26
  ## Model description
@@ -46,25 +46,26 @@ The following hyperparameters were used during training:
46
  - seed: 42
47
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
48
  - lr_scheduler_type: linear
49
- - num_epochs: 20
50
  - mixed_precision_training: Native AMP
51
 
52
  ### Training results
53
 
54
  | Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
55
  |:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|:-------:|:---------:|:--------:|
56
- | No log | 0.31 | 125 | 1.0221 | 52.9231 | 33.0585 | 34.8759 | 50.4025 | 142.0 |
57
- | No log | 0.63 | 250 | 1.0050 | 52.483 | 32.5851 | 34.1932 | 49.8346 | 141.537 |
58
- | No log | 0.94 | 375 | 0.8597 | 52.2419 | 32.1165 | 34.7481 | 49.8615 | 141.6481 |
59
- | 0.8252 | 1.26 | 500 | 0.9014 | 51.8094 | 30.8565 | 32.3167 | 48.9838 | 142.0 |
60
- | 0.8252 | 1.57 | 625 | 0.8751 | 53.5082 | 33.6193 | 35.9424 | 50.7697 | 141.9815 |
61
- | 0.8252 | 1.88 | 750 | 0.8224 | 53.0186 | 33.545 | 36.3322 | 50.6706 | 142.0 |
62
- | 0.8252 | 2.2 | 875 | 0.8004 | 53.0535 | 34.3281 | 37.516 | 50.6702 | 142.0 |
63
- | 0.5298 | 2.51 | 1000 | 0.7886 | 52.767 | 33.3381 | 35.6771 | 49.9858 | 142.0 |
64
- | 0.5298 | 2.83 | 1125 | 0.7875 | 53.7748 | 34.0148 | 35.4091 | 51.386 | 142.0 |
65
- | 0.5298 | 3.14 | 1250 | 0.8419 | 54.0386 | 35.3586 | 38.4748 | 51.5553 | 142.0 |
66
- | 0.5298 | 3.45 | 1375 | 0.8205 | 53.8319 | 34.6958 | 37.3817 | 51.3208 | 142.0 |
67
- | 0.3506 | 3.77 | 1500 | 0.7951 | 53.9498 | 35.4838 | 38.9904 | 51.3848 | 142.0 |
 
68
 
69
 
70
  ### Framework versions
 
16
 
17
  This model is a fine-tuned version of [theojolliffe/bart-cnn-pubmed-arxiv-pubmed-arxiv-arxiv](https://huggingface.co/theojolliffe/bart-cnn-pubmed-arxiv-pubmed-arxiv-arxiv) on an unknown dataset.
18
  It achieves the following results on the evaluation set:
19
+ - Loss: 0.8347
20
+ - Rouge1: 53.9049
21
+ - Rouge2: 35.5953
22
+ - Rougel: 39.788
23
+ - Rougelsum: 51.4101
24
  - Gen Len: 142.0
25
 
26
  ## Model description
 
46
  - seed: 42
47
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
48
  - lr_scheduler_type: linear
49
+ - num_epochs: 200
50
  - mixed_precision_training: Native AMP
51
 
52
  ### Training results
53
 
54
  | Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
55
  |:-------------:|:-----:|:----:|:---------------:|:-------:|:-------:|:-------:|:---------:|:--------:|
56
+ | No log | 0.31 | 125 | 1.0240 | 52.5632 | 32.977 | 34.672 | 49.9905 | 142.0 |
57
+ | No log | 0.63 | 250 | 1.0056 | 52.5508 | 32.4826 | 34.6851 | 49.835 | 141.6852 |
58
+ | No log | 0.94 | 375 | 0.8609 | 53.0475 | 32.9384 | 35.3322 | 50.272 | 141.6481 |
59
+ | 0.8255 | 1.26 | 500 | 0.9022 | 52.2493 | 31.5622 | 33.389 | 49.6612 | 142.0 |
60
+ | 0.8255 | 1.57 | 625 | 0.8706 | 53.3568 | 33.2533 | 35.7531 | 50.4568 | 141.8889 |
61
+ | 0.8255 | 1.88 | 750 | 0.8186 | 52.7375 | 33.4439 | 37.1094 | 50.5323 | 142.0 |
62
+ | 0.8255 | 2.2 | 875 | 0.8041 | 53.4992 | 34.6929 | 37.9614 | 51.091 | 142.0 |
63
+ | 0.5295 | 2.51 | 1000 | 0.7907 | 52.6185 | 33.8053 | 37.1725 | 50.4881 | 142.0 |
64
+ | 0.5295 | 2.83 | 1125 | 0.7740 | 52.7107 | 33.1023 | 36.0865 | 50.0365 | 142.0 |
65
+ | 0.5295 | 3.14 | 1250 | 0.8200 | 52.5607 | 33.7948 | 37.2312 | 50.3345 | 142.0 |
66
+ | 0.5295 | 3.45 | 1375 | 0.8188 | 53.9233 | 34.446 | 36.7566 | 51.3135 | 142.0 |
67
+ | 0.351 | 3.77 | 1500 | 0.8071 | 53.9096 | 35.5977 | 38.6832 | 51.4986 | 142.0 |
68
+ | 0.351 | 4.08 | 1625 | 0.8347 | 53.9049 | 35.5953 | 39.788 | 51.4101 | 142.0 |
69
 
70
 
71
  ### Framework versions