google-flan-t5-small-e-snli-generation-label_and_explanation-selected-b64

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

• Loss: 1.8703
• Accuracy: 0.8691
• F1: 0.8686
• Bertscore F1: 0.9338
• Rouge1: 0.6063
• Rouge2: 0.3995
• Rougel: 0.5500
• Rougelsum: 0.5521
• Bleu: 0.4012

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: 0.001
• 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: linear
• lr_scheduler_warmup_ratio: 0.05
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Bertscore F1	Rouge1	Rouge2	Rougel	Rougelsum	Bleu
1.4692	0.23	2000	1.7872	0.8212	0.8203	0.9287	0.5787	0.3685	0.5239	0.5257	0.3856
1.2505	0.47	4000	1.8808	0.8263	0.8264	0.9308	0.5870	0.3749	0.5321	0.5337	0.3904
1.2003	0.7	6000	1.8477	0.8475	0.8481	0.9325	0.5984	0.3913	0.5452	0.5469	0.4004
1.1624	0.93	8000	1.8244	0.8599	0.8587	0.9335	0.6029	0.3928	0.5441	0.5457	0.4024
1.1155	1.16	10000	1.8499	0.8695	0.8688	0.9331	0.6083	0.4019	0.5519	0.5540	0.4022
1.0913	1.4	12000	1.8703	0.8691	0.8686	0.9338	0.6063	0.3995	0.5500	0.5521	0.4012

Framework versions

• Transformers 4.27.4
• Pytorch 2.0.0+cu117
• Datasets 2.11.0
• Tokenizers 0.13.2