led-base-16384-100-MDS

This model is a fine-tuned version of allenai/led-base-16384 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 4.1425
• Rouge1: 16.7324
• Rouge2: 5.8501
• Rougel: 13.908
• Rougelsum: 13.8469
• Gen Len: 20.0

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: 1
• eval_batch_size: 1
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 4
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	1.0	25	3.6187	15.1426	4.2468	13.4488	13.38	20.0
No log	2.0	50	3.9873	13.4341	3.3283	10.2739	10.8229	20.0
No log	3.0	75	4.0264	18.1891	5.3395	15.0797	15.3586	20.0
No log	4.0	100	4.0929	17.0091	5.5336	14.4381	14.5149	19.5
No log	5.0	125	4.1425	16.7324	5.8501	13.908	13.8469	20.0

Framework versions

• Transformers 4.16.2
• Pytorch 1.10.2
• Datasets 1.18.3
• Tokenizers 0.11.0