codet5p-770m-py-codebleu-64-True-1e-06-0.1

This model is a fine-tuned version of Salesforce/codet5p-770m-py on the mbpp dataset. It achieves the following results on the evaluation set:

• Loss: 0.8108
• Codebleu: 0.0867
• Ngram Match Score: 0.0136
• Weighted Ngram Match Score: 0.0422
• Syntax Match Score: 0.1204
• Dataflow Match Score: 0.0824

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: 1e-06
• train_batch_size: 6
• eval_batch_size: 6
• seed: 42
• gradient_accumulation_steps: 64
• total_train_batch_size: 384
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• num_epochs: 50
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Codebleu	Ngram Match Score	Weighted Ngram Match Score	Syntax Match Score	Dataflow Match Score
0.9614	1.0	1	0.9113	0.0039	0.0000	0.0000	0.0048	0.0049
0.4928	2.0	3	0.9113	0.0039	0.0000	0.0000	0.0048	0.0049
0.4867	3.0	5	0.9112	0.0055	0.0000	0.0000	0.0067	0.0070
0.4813	4.0	7	0.9111	0.0063	0.0000	0.0002	0.0072	0.0084
0.4794	5.0	9	0.9108	0.0065	0.0000	0.0002	0.0072	0.0091
0.4857	6.0	11	0.9106	0.0124	0.0000	0.0012	0.0173	0.0133
0.4835	7.0	13	0.9095	0.0132	0.0000	0.0022	0.0178	0.0147
0.4902	8.0	15	0.9090	0.0199	0.0000	0.0054	0.0246	0.0237
0.4859	9.0	17	0.9053	0.0206	0.0000	0.0057	0.0255	0.0244
0.4787	10.0	19	0.9041	0.0326	0.0002	0.0152	0.0414	0.0363
0.485	11.0	21	0.9031	0.0435	0.0008	0.0199	0.0554	0.0482
0.4756	12.0	23	0.8915	0.0485	0.0020	0.0221	0.0592	0.0559
0.4774	13.0	25	0.8893	0.0629	0.0054	0.0314	0.0804	0.0677
0.4724	14.0	27	0.8859	0.0664	0.0055	0.0287	0.0877	0.0698
0.4755	15.0	29	0.8832	0.0782	0.0093	0.0346	0.1084	0.0761
0.458	16.0	31	0.8618	0.0797	0.0102	0.0357	0.1103	0.0775
0.4549	17.0	33	0.8586	0.0797	0.0105	0.0363	0.1122	0.0754
0.448	18.0	35	0.8560	0.0804	0.0109	0.0363	0.1132	0.0761
0.45	19.0	37	0.8530	0.0805	0.0110	0.0363	0.1132	0.0761
0.4403	20.0	39	0.8499	0.0794	0.0111	0.0363	0.1113	0.0754
0.4373	21.0	41	0.8345	0.0797	0.0118	0.0385	0.1113	0.0754
0.4208	22.0	43	0.8381	0.0816	0.0120	0.0386	0.1132	0.0782
0.4159	23.0	45	0.8337	0.0824	0.0122	0.0386	0.1137	0.0796
0.4157	24.0	47	0.8264	0.0857	0.0136	0.0422	0.1199	0.0803
0.4101	25.0	49	0.8161	0.0850	0.0134	0.0420	0.1190	0.0796
0.3181	25.4	50	0.8108	0.0867	0.0136	0.0422	0.1204	0.0824

Framework versions

• Transformers 4.30.0.dev0
• Pytorch 2.0.1
• Datasets 2.13.1
• Tokenizers 0.13.3