metadata

license: apache-2.0
tags:
  - generated_from_trainer
metrics:
  - rouge
  - bleu
model-index:
  - name: Salesforce-codet5-small-CodeXGLUE-CONCODE-adamw
    results: []

Salesforce-codet5-small-CodeXGLUE-CONCODE-adamw

This model is a fine-tuned version of Salesforce/codet5-small on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.7666
Exact Match: 0.163
Rouge1: 0.5716
Rouge2: 0.4046
Rougel: 0.5536
Rougelsum: 0.5614
Bleu: 0.1335

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: 32
eval_batch_size: 32
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_ratio: 0.05
num_epochs: 10
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Exact Match	Rouge1	Rouge2	Rougel	Rougelsum	Bleu
2.3935	0.16	500	0.9724	0.129	0.5286	0.3466	0.5098	0.5153	0.1127
0.8984	0.32	1000	0.8919	0.138	0.5463	0.3714	0.5285	0.5353	0.1200
0.8121	0.48	1500	0.8583	0.1455	0.5529	0.3787	0.5350	0.5426	0.1158
0.7598	0.64	2000	0.8437	0.1485	0.5541	0.3813	0.5355	0.5432	0.1197
0.7289	0.8	2500	0.8189	0.158	0.5597	0.3906	0.5416	0.5501	0.1222
0.7053	0.96	3000	0.8145	0.161	0.5572	0.3888	0.5392	0.5469	0.1222
0.6544	1.12	3500	0.7982	0.1565	0.5606	0.3920	0.5436	0.5517	0.1260
0.6334	1.28	4000	0.7974	0.1585	0.5633	0.3906	0.5448	0.5529	0.1284
0.6236	1.44	4500	0.7943	0.163	0.5639	0.3931	0.5455	0.5542	0.1275
0.6221	1.6	5000	0.7824	0.1655	0.5718	0.4011	0.5537	0.5621	0.1310
0.608	1.76	5500	0.7792	0.163	0.5664	0.3997	0.5490	0.5567	0.1314
0.5956	1.92	6000	0.7785	0.1605	0.5641	0.3981	0.5470	0.5546	0.1294
0.5701	2.08	6500	0.7800	0.157	0.5673	0.3955	0.5489	0.5568	0.1336
0.5378	2.24	7000	0.7720	0.1655	0.5686	0.4000	0.5504	0.5582	0.1308
0.541	2.4	7500	0.7709	0.1625	0.5699	0.3984	0.5511	0.5590	0.1313
0.5359	2.56	8000	0.7673	0.164	0.5697	0.4023	0.5521	0.5601	0.1332
0.5322	2.72	8500	0.7642	0.1665	0.5708	0.4033	0.5527	0.5606	0.1350
0.5387	2.88	9000	0.7622	0.159	0.5672	0.3988	0.5500	0.5573	0.1342
0.514	3.04	9500	0.7700	0.166	0.5722	0.4052	0.5546	0.5618	0.1352
0.4895	3.2	10000	0.7676	0.1615	0.5696	0.4016	0.5516	0.5591	0.1359
0.4827	3.36	10500	0.7665	0.162	0.5756	0.4072	0.5577	0.5656	0.1367
0.4814	3.52	11000	0.7700	0.1605	0.5709	0.4026	0.5528	0.5605	0.1334
0.4847	3.68	11500	0.7666	0.163	0.5716	0.4046	0.5536	0.5614	0.1335

Framework versions

Transformers 4.27.1
Pytorch 1.12.1+cu113
Datasets 2.10.1
Tokenizers 0.13.2