jialicheng/unlearn_samsum_t5-large_neggrad_8_42

samsum_42

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

• Loss: 380.7704
• Rouge1: 0.2678
• Rouge2: 0.0
• Rougel: 0.2715
• Rougelsum: 0.2684
• Gen Len: 2.0
• Test Rougel: 0.2715
• Df Rougel: 0.259
• Unlearn Overall Rougel: 0.5062
• Unlearn Time: 1071.4580

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: 8
• eval_batch_size: 64
• 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	147	91.4111	2.279	0.0328	2.5438	2.2783	2.7565	0.3691	0.3691	-1
No log	2.0	294	313.6870	0.2756	0.0	0.3062	0.275	2.0	0.4863	0.4863	-1
No log	3.0	441	380.7704	0.2678	0.0	0.259	0.2684	2.0	0.5062	0.5062	-1
No log	4.0	588	414.6086	0.2184	0.0	0.2359	0.2196	2.0	0.4932	0.4932	-1
No log	5.0	735	422.9209	0.2184	0.0	0.2359	0.2196	2.0	0.4932	0.4932	-1

Framework versions

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