c90dba61992ce13b962d6a315b00a581

This model is a fine-tuned version of openai-community/gpt2-large on the nyu-mll/glue [stsb] dataset. It achieves the following results on the evaluation set:

• Loss: 0.4788
• Data Size: 1.0
• Epoch Runtime: 46.6523
• Mse: 0.4790
• Mae: 0.5393
• R2: 0.7857

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: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 32
• total_eval_batch_size: 32
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Data Size	Epoch Runtime	Mse	Mae	R2
No log	0	0	11.8708	0	3.8049	11.8719	3.0847	-4.3107
No log	1	179	7.7273	0.0078	4.4553	7.7277	2.2192	-2.4569
No log	2	358	2.3675	0.0156	5.7047	2.3683	1.2690	-0.0594
No log	3	537	1.4222	0.0312	7.4723	1.4227	0.9663	0.3636
No log	4	716	0.8176	0.0625	10.1735	0.8178	0.7287	0.6342
No log	5	895	0.7726	0.125	13.3945	0.7727	0.7087	0.6544
0.0912	6	1074	0.6098	0.25	19.2932	0.6100	0.6195	0.7271
0.5293	7	1253	0.5783	0.5	28.9126	0.5785	0.5982	0.7412
0.3919	8.0	1432	0.5675	1.0	47.9233	0.5677	0.5867	0.7460
0.2448	9.0	1611	0.6561	1.0	48.6200	0.6563	0.6303	0.7064
0.1739	10.0	1790	0.4839	1.0	46.6943	0.4840	0.5299	0.7835
0.1481	11.0	1969	0.5863	1.0	47.2605	0.5864	0.6054	0.7377
0.1053	12.0	2148	0.5150	1.0	46.5543	0.5151	0.5651	0.7696
0.0902	13.0	2327	0.4750	1.0	47.2309	0.4751	0.5379	0.7875
0.0768	14.0	2506	0.5267	1.0	47.0802	0.5269	0.5799	0.7643
0.0627	15.0	2685	0.4757	1.0	46.9345	0.4759	0.5366	0.7871
0.0626	16.0	2864	0.4664	1.0	47.7827	0.4665	0.5298	0.7913
0.0561	17.0	3043	0.4764	1.0	46.8209	0.4764	0.5332	0.7869
0.0548	18.0	3222	0.5269	1.0	47.2221	0.5271	0.5729	0.7642
0.0445	19.0	3401	0.4676	1.0	47.0163	0.4678	0.5299	0.7908
0.0394	20.0	3580	0.4788	1.0	46.6523	0.4790	0.5393	0.7857

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.3.0
• Tokenizers 0.22.1