sst-gpt2

This model is a fine-tuned version of gpt2 on the sst dataset. It achieves the following results on the evaluation set:

• Loss: 0.0218
• Mse: 0.0218

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: 16
• eval_batch_size: 32
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Mse
0.2144	1.0	534	0.0300	0.0301
0.0251	2.0	1068	0.0249	0.0250
0.0181	3.0	1602	0.0230	0.0230
0.0134	4.0	2136	0.0244	0.0244
0.0096	5.0	2670	0.0228	0.0228
0.0077	6.0	3204	0.0227	0.0227
0.0062	7.0	3738	0.0227	0.0227
0.0052	8.0	4272	0.0229	0.0229
0.0044	9.0	4806	0.0227	0.0226
0.0038	10.0	5340	0.0240	0.0240
0.0034	11.0	5874	0.0221	0.0221
0.0029	12.0	6408	0.0220	0.0220
0.0026	13.0	6942	0.0229	0.0229
0.0024	14.0	7476	0.0217	0.0217
0.002	15.0	8010	0.0225	0.0225
0.0018	16.0	8544	0.0222	0.0222
0.0016	17.0	9078	0.0217	0.0217
0.0015	18.0	9612	0.0218	0.0218
0.0014	19.0	10146	0.0217	0.0217
0.0013	20.0	10680	0.0218	0.0218

Framework versions

• Transformers 4.37.0
• Pytorch 1.13.1+cu117
• Datasets 2.15.0
• Tokenizers 0.15.2