flan-t5-large
This model is a fine-tuned version of google/flan-t5-large on the cnn_dailymail dataset. It achieves the following results on the evaluation set:
- Loss: 1.3826
- Rouge1: 43.1919
- Rouge2: 20.634
- Rougel: 30.3172
- Rougelsum: 40.0097
- Gen Len: 94.6171
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: 192
- eval_batch_size: 192
- seed: 42
- distributed_type: multi-GPU
- num_devices: 2
- total_train_batch_size: 384
- total_eval_batch_size: 384
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Rougelsum | Gen Len |
|---|---|---|---|---|---|---|---|---|
| 1.653 | 1.0 | 748 | 1.4147 | 42.4138 | 19.9824 | 29.549 | 39.2091 | 100.5338 |
| 1.5924 | 2.0 | 1496 | 1.3920 | 42.8582 | 20.3835 | 30.1212 | 39.7493 | 94.9760 |
| 1.5231 | 3.0 | 2244 | 1.3826 | 43.1919 | 20.634 | 30.3172 | 40.0097 | 94.6171 |
Framework versions
- Transformers 4.26.0
- Pytorch 1.13.1+cu116
- Datasets 2.9.0
- Tokenizers 0.13.2
- Downloads last month
- 4
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support
Dataset used to train vsevolodl/flan-t5-large-sum
Evaluation results
- Rouge1 on cnn_dailymailtest set self-reported43.192