flan-t5-base-flant5-apple-support

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

• Loss: 2.9676
• Rouge1: 12.7991
• Rouge2: 2.244
• Rougel: 9.8075
• Rougelsum: 11.3618
• Gen Len: 18.9087

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
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
3.2673	1.0	1157	3.0350	12.4094	2.1794	9.5255	10.9739	18.9723
3.1854	2.0	2314	2.9992	12.4579	2.1512	9.5232	11.0049	18.9647
3.1006	3.0	3471	2.9792	12.9794	2.2794	9.9245	11.5019	18.9436
3.0751	4.0	4628	2.9711	12.6779	2.1828	9.6962	11.221	18.9137
3.0532	5.0	5785	2.9676	12.7991	2.244	9.8075	11.3618	18.9087

Framework versions

• Transformers 4.25.1
• Pytorch 1.13.1+cu117
• Datasets 2.8.0
• Tokenizers 0.13.2