codet5-small-generate-docstrings-codexglue-python-bs-32

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

• Loss: 2.0595
• Rouge1: 0.3535
• Rouge2: 0.1748
• Rougel: 0.3195
• Rougelsum: 0.336
• Gen Len: 16.1317

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: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
2.6758	1.0	5952	2.2761	0.3458	0.1674	0.312	0.329	16.375
2.4212	2.0	11904	2.1955	0.3471	0.1698	0.3143	0.33	15.3815
2.3395	3.0	17856	2.1480	0.3501	0.1715	0.3159	0.3329	16.4806
2.2889	4.0	23808	2.1198	0.3506	0.1722	0.3167	0.3333	15.974
2.252	5.0	29760	2.0984	0.3524	0.1738	0.3184	0.3351	16.2197
2.226	6.0	35712	2.0851	0.3518	0.1736	0.3182	0.3345	15.9882
2.2047	7.0	41664	2.0746	0.3521	0.1738	0.3185	0.3348	16.2078
2.1894	8.0	47616	2.0654	0.3532	0.1748	0.3194	0.3357	16.1077
2.1798	9.0	53568	2.0617	0.3528	0.1749	0.3192	0.3355	16.0418
2.1732	10.0	59520	2.0595	0.3535	0.1748	0.3195	0.336	16.1317

Framework versions

• Transformers 4.33.0.dev0
• Pytorch 2.0.1+cu118
• Datasets 2.14.4.dev0
• Tokenizers 0.13.3