longt5_xl_summ_screen_bp_10

This model is a fine-tuned version of google/long-t5-tglobal-xl on the learn3r/summ_screen_fd_bp dataset. It achieves the following results on the evaluation set:

• Loss: 1.3323
• Rouge1: 22.9554
• Rouge2: 6.4509
• Rougel: 19.7437
• Rougelsum: 20.923
• Gen Len: 497.2456

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: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 32
• total_train_batch_size: 256
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: constant
• num_epochs: 10.0

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
2.4559	0.97	14	2.0707	11.7833	1.6011	11.1858	10.3025	511.0
1.6238	1.95	28	1.5287	19.0489	4.687	16.6504	17.1808	511.0
1.3964	2.99	43	1.3520	21.9994	5.8519	18.9231	19.958	511.0
1.2538	3.97	57	1.3323	22.9554	6.4509	19.7437	20.923	497.2456
1.277	4.94	71	1.5462	14.6326	3.6509	12.4805	13.5001	507.2278
1.0071	5.98	86	1.3604	29.5352	9.9544	22.1073	28.1204	429.7722
0.8685	6.96	100	1.4361	31.0337	10.6724	22.3815	29.6325	451.7840
0.7498	8.0	115	1.5302	28.433	8.4887	21.3588	26.6817	473.8964
0.6226	8.97	129	1.6289	37.251	12.8214	24.8704	36.0027	358.7663
0.5558	9.74	140	1.5811	35.4657	12.0036	24.7787	34.3775	284.0266

Framework versions

• Transformers 4.34.0.dev0
• Pytorch 2.0.1+cu117
• Datasets 2.14.5
• Tokenizers 0.13.3