flan-t5-base-gen-chat

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

• Loss: 3.2310
• Rouge 1: 9.4633
• Rouge 2: 1.039
• Rouge L: 8.7585
• Avg Len: 12.3194
• Bertscore Prec: 0.8624
• Bertscore Rec: 0.8561
• Bertscore F1: 0.8589

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: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge 1	Rouge 2	Rouge L	Avg Len	Bertscore Prec	Bertscore Rec	Bertscore F1
3.7376	0.5848	200	3.4194	7.4287	0.8472	6.9898	9.809	0.8633	0.8536	0.8581
3.4588	1.1696	400	3.3339	7.1926	0.8083	6.7158	9.6285	0.867	0.8546	0.8604
3.3367	1.7544	600	3.2864	8.7272	1.1167	8.0103	11.9792	0.8638	0.8565	0.8598
3.2902	2.3392	800	3.2639	8.5549	1.1047	7.9038	10.4861	0.8693	0.8558	0.8622
3.2053	2.9240	1000	3.2464	8.715	1.0443	8.0952	10.9375	0.8667	0.8558	0.8609
3.139	3.5088	1200	3.2420	9.1545	1.1215	8.4443	12.0035	0.8651	0.8571	0.8608
3.1294	4.0936	1400	3.2318	9.1819	1.0984	8.5163	11.9028	0.8634	0.8554	0.859
3.0969	4.6784	1600	3.2310	9.4633	1.039	8.7585	12.3194	0.8624	0.8561	0.8589

Framework versions

• Transformers 4.48.3
• Pytorch 2.6.0+cu124
• Datasets 3.4.1
• Tokenizers 0.21.1