BART-large-commongen
This model is a fine-tuned version of facebook/bart-large on the gem dataset. It achieves the following results on the evaluation set:
- Loss: 1.1409
- Spice: 0.4009
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: 0.0001
- train_batch_size: 32
- eval_batch_size: 32
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 1000
- training_steps: 6317
Training results
Training Loss | Epoch | Step | Validation Loss | Spice |
---|---|---|---|---|
10.1086 | 0.05 | 100 | 4.9804 | 0.3736 |
4.4168 | 0.09 | 200 | 2.4402 | 0.4079 |
1.8158 | 0.14 | 300 | 1.1096 | 0.4258 |
1.1723 | 0.19 | 400 | 1.0845 | 0.4086 |
1.0894 | 0.24 | 500 | 1.0727 | 0.423 |
1.0949 | 0.28 | 600 | 1.0889 | 0.4224 |
1.0773 | 0.33 | 700 | 1.0977 | 0.4201 |
1.0708 | 0.38 | 800 | 1.1157 | 0.4213 |
1.0663 | 0.43 | 900 | 1.1798 | 0.421 |
1.0985 | 0.47 | 1000 | 1.1611 | 0.4025 |
1.0561 | 0.52 | 1100 | 1.1048 | 0.421 |
1.0594 | 0.57 | 1200 | 1.2044 | 0.3626 |
1.0689 | 0.62 | 1300 | 1.1409 | 0.4009 |
Framework versions
- Transformers 4.9.2
- Pytorch 1.9.0+cu102
- Datasets 1.11.1.dev0
- Tokenizers 0.10.3
- Downloads last month
- 9
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social
visibility and check back later, or deploy to Inference Endpoints (dedicated)
instead.