README.md

metadata

license: apache-2.0
tags:
  - generated_from_trainer
metrics:
  - bleu
  - rouge
model-index:
  - name: t5-small-codesearchnet-python
    results: []

t5-small-codesearchnet-python

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

• Loss: 0.0785
• Bleu: 0.035
• Rouge1: 0.6257
• Rouge2: 0.6078
• Avg Length: 16.9954

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	0.0801	0.0358	0.6174	0.6	17.1074
1.6066	2.0	750	0.0674	0.036	0.6249	0.6068	17.0262
0.0584	3.0	1125	0.0632	0.0351	0.6255	0.6075	16.9962
0.0484	4.0	1500	0.0605	0.0351	0.6251	0.6071	17.003
0.0484	5.0	1875	0.0596	0.035	0.6255	0.6075	17.0012
0.0418	6.0	2250	0.0602	0.035	0.6258	0.608	16.9958
0.0377	7.0	2625	0.0593	0.0351	0.6259	0.6079	17.0004
0.033	8.0	3000	0.0618	0.035	0.6257	0.6078	17.0032
0.033	9.0	3375	0.0637	0.035	0.6257	0.6078	16.998
0.028	10.0	3750	0.0645	0.035	0.6257	0.6079	16.9984
0.0255	11.0	4125	0.0650	0.035	0.6255	0.6078	17.0008
0.0226	12.0	4500	0.0748	0.035	0.6254	0.6076	16.9976
0.0226	13.0	4875	0.0714	0.035	0.6256	0.6079	16.9954
0.019	14.0	5250	0.0747	0.0349	0.6253	0.6077	16.994
0.0172	15.0	5625	0.0785	0.035	0.6257	0.6078	16.9954

Framework versions

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