README.md

metadata

license: bsd-3-clause
tags:
  - generated_from_trainer
metrics:
  - rouge
  - bleu
model-index:
  - name: CommitPredictorT5
    results: []

CommitPredictorT5

This model is a fine-tuned version of Salesforce/codet5-base-multi-sum on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.4517
• Rouge1: 0.0001
• Rouge2: 0.0
• Rougel: 0.0001
• Rougelsum: 0.0001
• Gen Len: 1.0
• Bleu: 0.0004

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: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 30

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len	Bleu
3.2156	1.0	687	2.6943	0.0	0.0	0.0	0.0	1.0	0.0005
2.7542	2.0	1374	2.5573	0.0	0.0	0.0	0.0	1.0	0.0004
2.4383	3.0	2061	2.4856	0.0	0.0	0.0	0.0	1.0	0.0004
2.3043	4.0	2748	2.4514	0.0	0.0	0.0	0.0	1.0	0.0005
2.2043	5.0	3435	2.4373	0.0001	0.0	0.0001	0.0001	1.0	0.0004
2.0354	6.0	4122	2.4284	0.0003	0.0	0.0003	0.0003	1.0	0.0004
1.9542	7.0	4809	2.4332	0.0	0.0	0.0	0.0	1.0	0.0004
1.8954	8.0	5496	2.4517	0.0001	0.0	0.0001	0.0001	1.0	0.0004

Framework versions

• Transformers 4.25.1
• Pytorch 1.13.0+cu117
• Datasets 2.7.1
• Tokenizers 0.13.2