Zohaib002/Longformer-Encoder-Decoder-LED

device

This model is a fine-tuned version of pszemraj/led-base-book-summary on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.0247
• Rouge1: 0.6269
• Rouge2: 0.3921
• Rougel: 0.5261
• Rougelsum: 0.5266
• Gen Len: 67.5584

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: 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: 8
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	1.0	274	1.0933	0.5918	0.3356	0.4785	0.4788	72.0547
1.1731	2.0	548	1.0177	0.5985	0.3525	0.4902	0.4906	68.5055
1.1731	3.0	822	0.9976	0.6063	0.3603	0.4982	0.4982	69.7263
0.7216	4.0	1096	0.9922	0.6113	0.3735	0.5081	0.5084	68.1861
0.7216	5.0	1370	0.9957	0.6193	0.3826	0.5216	0.5217	65.4617
0.5252	6.0	1644	1.0127	0.6252	0.3877	0.5231	0.5236	68.0584
0.5252	7.0	1918	1.0221	0.6252	0.3897	0.5246	0.5246	67.5931
0.4079	8.0	2192	1.0247	0.6269	0.3921	0.5261	0.5266	67.5584

Framework versions

• Transformers 4.42.4
• Pytorch 2.3.1+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1