lmeninato/t5-small-codesearchnet-multilang-4-archive

t5-small-codesearchnet-multilang-python-java-javascript-go

This model is a fine-tuned version of t5-small on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4172
• Bleu: 0.0344
• Rouge1: 0.3447
• Rouge2: 0.2223
• Avg Length: 17.0756

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: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 10
• total_train_batch_size: 80
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Rouge1	Rouge2	Avg Length
No log	1.0	375	0.4953	0.0246	0.2585	0.1659	17.688
2.1147	2.0	750	0.4376	0.0263	0.293	0.1912	17.044
0.4761	3.0	1125	0.4125	0.0317	0.2964	0.1965	17.4636
0.4163	4.0	1500	0.3980	0.0277	0.3095	0.2015	16.3844
0.4163	5.0	1875	0.3868	0.0293	0.325	0.2082	16.6804
0.3808	6.0	2250	0.3815	0.0288	0.3232	0.2089	16.7996
0.352	7.0	2625	0.3720	0.024	0.3225	0.2119	15.8864
0.3295	8.0	3000	0.3695	0.0198	0.3142	0.2118	15.4172
0.3295	9.0	3375	0.3736	0.0305	0.343	0.2206	16.8566
0.3052	10.0	3750	0.3811	0.0316	0.3442	0.2215	16.8834
0.2887	11.0	4125	0.3775	0.023	0.3262	0.2149	15.6226
0.2799	12.0	4500	0.3793	0.034	0.3397	0.2186	16.6344
0.2799	13.0	4875	0.3875	0.0312	0.3389	0.2192	16.7964
0.2517	14.0	5250	0.3833	0.0337	0.3387	0.22	16.9076
0.2409	15.0	5625	0.4172	0.0344	0.3447	0.2223	17.0756

Framework versions

• Transformers 4.28.1
• Pytorch 2.0.0+cu118
• Datasets 2.12.0
• Tokenizers 0.13.3