flan-t5-base-comma-correction

This model is a fine-tuned version of google/flan-t5-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0492
• Gen Len: 39.5298
• Em Ic: 0.9334
• Em: 0.6384
• Precision: 0.9141
• Recall: 0.8575
• F1: 0.8248
• Levinstein Ratio: 0.9955

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: 2e-05
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 200
• num_epochs: 3.0

Training results

Training Loss	Epoch	Step	Validation Loss	Gen Len	Em Ic	Em	Precision	Recall	F1	Levinstein Ratio
0.042	1.0	2159	0.0500	39.4971	0.9328	0.6256	0.9175	0.8452	0.8165	0.9954
0.0342	2.0	4318	0.0492	39.5158	0.9334	0.6361	0.9159	0.8546	0.8229	0.9955
0.0346	3.0	6477	0.0492	39.5298	0.9334	0.6384	0.9141	0.8575	0.8248	0.9955

Framework versions

• Transformers 4.35.0
• Pytorch 2.1.0+cu118
• Datasets 2.14.6
• Tokenizers 0.14.1