longt5_xl_gov_report_bp_10_continue

This model is a fine-tuned version of /home/co-ou1/rds/hpc-work/transformers/examples/pytorch/summarization/longt5_xl_gov_report_bp_10/checkpoint-477 on the learn3r/gov_report_memsum_oracle dataset. It achieves the following results on the evaluation set:

• Loss: 1.4878
• Rouge1: 71.9439
• Rouge2: 43.7031
• Rougel: 41.8301
• Rougelsum: 69.1853
• Gen Len: 833.0319

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: 4.0

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
0.6226	1.0	68	1.4878	71.9439	43.7031	41.8301	69.1853	833.0319
0.4983	1.99	136	1.5908	70.6191	43.2627	42.581	68.0871	627.5031
0.4175	2.99	204	1.6407	71.6704	43.1655	41.9746	68.992	737.4352
0.3958	3.99	272	1.8739	70.7685	42.5122	41.7454	68.0785	671.4938

Framework versions

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