jialicheng/unlearn_samsum_t5-base_neggrad_2_42

samsum_42

This model is a fine-tuned version of google/t5-v1_1-base on the samsum dataset. It achieves the following results on the evaluation set:

• Loss: 1.5068
• Rouge1: 48.3063
• Rouge2: 23.9806
• Rougel: 39.8938
• Rougelsum: 44.5518
• Gen Len: 28.0
• Test Rougel: 39.8938
• Df Rougel: 41.3412
• Unlearn Overall Rougel: -0.2237
• Unlearn Time: 833.6055

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: 1e-05
• train_batch_size: 32
• eval_batch_size: 4
• 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	Overall Rougel	Unlearn Overall Rougel	Time
No log	1.0	10	1.4950	48.9548	24.2526	42.6093	45.0242	24.9156	-0.5616	-0.5616	-1
No log	2.0	20	1.4983	48.7024	24.2549	42.095	44.9152	25.5648	-0.3513	-0.3513	-1
No log	3.0	30	1.5019	48.4999	24.1774	41.8905	44.8249	26.4401	-0.2970	-0.2970	-1
No log	4.0	40	1.5055	48.3742	24.0652	41.6664	44.6241	27.6932	-0.3179	-0.3179	-1
No log	5.0	50	1.5068	48.3063	23.9806	41.3412	44.5518	28.0	-0.2237	-0.2237	-1

Framework versions

• Transformers 4.39.3
• Pytorch 2.3.0+cu121
• Datasets 2.19.0
• Tokenizers 0.15.2