jialicheng/unlearn_samsum_t5-base_bad_teaching_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: 90.6081
• Rouge1: 0.0
• Rouge2: 0.0
• Rougel: 0.0
• Rougelsum: 0.0
• Gen Len: 127.0
• Test Rougel: 0.0
• Df Rougel: 0.0
• Unlearn Overall Rougel: 0.5
• Unlearn Time: 6438.0157

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: 32
• eval_batch_size: 8
• 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	0.5	451	78.3206	0.0	0.0	0.0	0.0	127.0	0.5	0.5	-1
No log	1.0	902	83.2608	0.0	0.0	0.0	0.0	127.0	0.5	0.5	-1
0.0	1.5	1353	85.1384	0.0	0.0	0.0	0.0	127.0	0.5	0.5	-1
0.0	2.0	1804	86.5946	0.0	0.0	0.0	0.0	127.0	0.5	0.5	-1
0.0	2.5	2255	87.8266	0.0	0.0	0.0	0.0	127.0	0.5	0.5	-1
0.0	3.0	2706	88.8574	0.0	0.0	0.0	0.0	127.0	0.5	0.5	-1
0.0	3.5	3157	89.6562	0.0	0.0	0.0	0.0	127.0	0.5	0.5	-1
0.0	4.0	3608	90.1704	0.0	0.0	0.0	0.0	127.0	0.5	0.5	-1
0.0	4.5	4059	90.5064	0.0	0.0	0.0	0.0	127.0	0.5	0.5	-1
0.0	4.99	4510	90.6081	0.0	0.0	0.0	0.0	127.0	0.5	0.5	-1

Framework versions

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