metadata
license: apache-2.0
base_model: google/switch-base-8
tags:
- generated_from_trainer
datasets:
- samsum
metrics:
- rouge
model-index:
- name: switch-base-8-samsum-top-4-choose-1-deconly
results:
- task:
name: Sequence-to-sequence Language Modeling
type: text2text-generation
dataset:
name: samsum
type: samsum
config: samsum
split: validation
args: samsum
metrics:
- name: Rouge1
type: rouge
value: 47.2666
switch-base-8-samsum-top-4-choose-1-deconly
This model is a fine-tuned version of google/switch-base-8 on the samsum dataset. It achieves the following results on the evaluation set:
- Loss: 1.5869
- Rouge1: 47.2666
- Rouge2: 24.2196
- Rougel: 40.1766
- Rougelsum: 43.8418
- Gen Len: 16.9352
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 5.0
Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|---|---|---|---|---|---|---|---|---|
| 5.4611 | 0.2172 | 200 | 3.0917 | 23.5686 | 7.6846 | 20.6877 | 22.0746 | 14.7946 |
| 2.6551 | 0.4343 | 400 | 2.1027 | 39.7231 | 17.2476 | 33.3172 | 37.0509 | 17.1369 |
| 2.4452 | 0.6515 | 600 | 1.9255 | 42.9952 | 19.6478 | 35.8054 | 40.1569 | 17.3007 |
| 2.1259 | 0.8686 | 800 | 1.8270 | 43.9723 | 21.3238 | 37.0066 | 40.9323 | 16.1027 |
| 2.0957 | 1.0858 | 1000 | 1.7708 | 45.1103 | 21.769 | 37.9229 | 41.7446 | 17.2482 |
| 2.1168 | 1.3029 | 1200 | 1.7185 | 45.6806 | 22.0335 | 38.2398 | 42.4051 | 16.5941 |
| 2.1491 | 1.5201 | 1400 | 1.6982 | 46.0573 | 22.2803 | 38.33 | 42.531 | 16.9291 |
| 1.9829 | 1.7372 | 1600 | 1.6803 | 45.8845 | 22.4145 | 38.795 | 42.5814 | 16.4976 |
| 1.9741 | 1.9544 | 1800 | 1.6657 | 45.6645 | 22.0154 | 38.2445 | 42.2358 | 17.2689 |
| 1.8286 | 2.1716 | 2000 | 1.6462 | 46.7647 | 23.2912 | 39.4015 | 43.3207 | 16.8704 |
| 1.8177 | 2.3887 | 2200 | 1.6486 | 45.8872 | 22.8119 | 38.7398 | 42.3427 | 16.0403 |
| 1.8606 | 2.6059 | 2400 | 1.6270 | 45.9799 | 22.9475 | 38.9393 | 42.7565 | 16.6687 |
| 1.8327 | 2.8230 | 2600 | 1.6210 | 46.2715 | 23.4171 | 39.4324 | 43.0326 | 16.5452 |
| 1.6738 | 3.0402 | 2800 | 1.6242 | 46.1248 | 22.7245 | 38.8572 | 42.5884 | 16.8252 |
| 1.7515 | 3.2573 | 3000 | 1.6155 | 46.5372 | 23.4014 | 39.54 | 43.187 | 16.665 |
| 1.7728 | 3.4745 | 3200 | 1.6000 | 46.6652 | 23.4739 | 39.4761 | 43.2783 | 16.7873 |
| 1.7584 | 3.6916 | 3400 | 1.5922 | 47.2313 | 24.0035 | 39.9195 | 43.6996 | 16.7702 |
| 1.7082 | 3.9088 | 3600 | 1.5957 | 46.5132 | 23.4692 | 39.4884 | 43.2236 | 16.6553 |
| 1.5968 | 4.1260 | 3800 | 1.5916 | 47.2622 | 23.9444 | 40.1308 | 43.7971 | 16.9083 |
| 1.6439 | 4.3431 | 4000 | 1.5880 | 46.9607 | 23.7839 | 39.7431 | 43.5831 | 16.9621 |
| 1.6684 | 4.5603 | 4200 | 1.5930 | 47.2611 | 23.9828 | 40.0767 | 43.8297 | 16.8851 |
| 1.7749 | 4.7774 | 4400 | 1.5882 | 46.9562 | 23.874 | 39.8904 | 43.536 | 16.9377 |
| 1.6401 | 4.9946 | 4600 | 1.5869 | 47.2666 | 24.2196 | 40.1766 | 43.8418 | 16.9352 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.0.1+cu117
- Datasets 2.20.0
- Tokenizers 0.19.1