flan-t5-small-codesearchnet-python

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

• Loss: 0.0764
• Bleu: 0.0349
• Rouge1: 0.6244
• Rouge2: 0.6055
• Avg Length: 16.9912

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	0.0636	0.0364	0.6253	0.6076	17.029
5.5166	2.0	750	0.0553	0.0351	0.6259	0.6081	16.9996
0.0485	3.0	1125	0.0537	0.0351	0.6258	0.6083	16.99
0.0409	4.0	1500	0.0524	0.0351	0.6258	0.6082	16.9942
0.0409	5.0	1875	0.0524	0.0351	0.6261	0.6086	16.997
0.0345	6.0	2250	0.0526	0.0351	0.6258	0.6081	16.9936
0.0303	7.0	2625	0.0533	0.035	0.6254	0.6076	16.991
0.0256	8.0	3000	0.0566	0.035	0.6257	0.6074	16.9964
0.0256	9.0	3375	0.0592	0.0349	0.6253	0.6074	16.998
0.0205	10.0	3750	0.0612	0.0351	0.6255	0.6073	16.9932
0.0185	11.0	4125	0.0639	0.035	0.6257	0.6079	16.996
0.0157	12.0	4500	0.0698	0.035	0.625	0.6064	16.9944
0.0157	13.0	4875	0.0720	0.035	0.6246	0.6062	16.991
0.0131	14.0	5250	0.0745	0.035	0.6247	0.6062	16.9986
0.0128	15.0	5625	0.0764	0.0349	0.6244	0.6055	16.9912

Framework versions

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