metadata
base_model: EleutherAI/pile-t5-base
tags:
- generated_from_trainer
model-index:
- name: pile-t5-base-inst
results: []
language:
- en
metrics:
- rouge
library_name: transformers
pile-t5-base-inst
This model is a fine-tuned version of EleutherAI/pile-t5-base on taskydata/Pile-T5-Instruction dataset. It achieves the following results on the evaluation set:
- Loss: 2.5082
- Rouge2 Precision: 0.2496
- Rouge2 Recall: 0.1633
- Rouge2 Fmeasure: 0.1786
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: 2e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 50
- num_epochs: 3
Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge2 Precision | Rouge2 Recall | Rouge2 Fmeasure |
|---|---|---|---|---|---|---|
| 5.8554 | 0.58 | 1500 | 3.2533 | 0.1 | 0.0744 | 0.0721 |
| 4.2403 | 1.16 | 3000 | 2.7020 | 0.1704 | 0.1174 | 0.1248 |
| 3.8091 | 1.74 | 4500 | 2.5844 | 0.2476 | 0.1617 | 0.1767 |
| 3.6589 | 2.32 | 6000 | 2.5289 | 0.2467 | 0.1621 | 0.1769 |
| 3.5802 | 2.9 | 7500 | 2.5082 | 0.2496 | 0.1633 | 0.1786 |
Framework versions
- Transformers 4.39.3
- Pytorch 2.1.2
- Datasets 2.18.0
- Tokenizers 0.15.2