metadata
license: apache-2.0
base_model: Salesforce/codet5-base
tags:
- code-generation
- generated_from_trainer
metrics:
- rouge
- bleu
model-index:
- name: codeT5-finetuned-cXg-nl-to-code
results: []
codeT5-finetuned-cXg-nl-to-code
This model is a fine-tuned version of Salesforce/codet5-base on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 1.5944
- Rouge1: 0.0735
- Rouge2: 0.0022
- Rougel: 0.0671
- Bleu: 0.7231
- Meteor: 0.0630
- Codebleu: 0.0382
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: 32
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 3
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Rouge1 | Rouge2 | Rougel | Bleu | Meteor | Codebleu |
|---|---|---|---|---|---|---|---|---|---|
| No log | 1.0 | 4 | 4.8941 | 0.0196 | 0.004 | 0.0201 | 0.0446 | 0.0142 | 0.0146 |
| No log | 2.0 | 8 | 2.1851 | 0.0519 | 0.0019 | 0.0519 | 0.1178 | 0.0281 | 0.0217 |
| No log | 3.0 | 12 | 1.5944 | 0.0731 | 0.0022 | 0.0670 | 0.1442 | 0.0596 | 0.0246 |
Framework versions
- Transformers 4.42.4
- Pytorch 2.3.1+cu121
- Datasets 2.20.0
- Tokenizers 0.19.1