590b981208fb911d23715c82ecc36db5

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

• Loss: 1.2671
• Data Size: 1.0
• Epoch Runtime: 10.6850
• Accuracy: 0.7913
• F1 Macro: 0.7415
• Rouge1: 0.7913
• Rouge2: 0.0
• Rougel: 0.7916
• Rougelsum: 0.7919

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	Accuracy	F1 Macro	Rouge1	Rouge2	Rougel	Rougelsum
No log	0	0	5.3584	0	2.0376	0.3544	0.3197	0.3544	0.0	0.3544	0.3550
No log	1	114	3.3104	0.0078	2.8362	0.6374	0.4274	0.6386	0.0	0.6371	0.6374
No log	2	228	3.1061	0.0156	2.2608	0.6527	0.4121	0.6527	0.0	0.6527	0.6527
No log	3	342	1.1286	0.0312	2.5475	0.6539	0.4882	0.6545	0.0	0.6536	0.6536
0.0617	4	456	0.6998	0.0625	2.9157	0.6421	0.5907	0.6421	0.0	0.6421	0.6418
0.0617	5	570	0.7271	0.125	3.3888	0.6792	0.4535	0.6798	0.0	0.6792	0.6792
0.0617	6	684	0.5975	0.25	4.3413	0.6993	0.5278	0.6993	0.0	0.6993	0.6999
0.1531	7	798	0.5503	0.5	6.3705	0.7476	0.7113	0.7476	0.0	0.7471	0.7476
0.4886	8.0	912	0.4413	1.0	10.6696	0.7919	0.7456	0.7913	0.0	0.7919	0.7925
0.3391	9.0	1026	0.4691	1.0	10.1370	0.7883	0.7302	0.7883	0.0	0.7883	0.7883
0.2394	10.0	1140	0.5529	1.0	10.5125	0.8013	0.7723	0.8013	0.0	0.8013	0.8019
0.1431	11.0	1254	0.6756	1.0	10.3943	0.7866	0.7596	0.7866	0.0	0.7866	0.7877
0.1001	12.0	1368	1.2671	1.0	10.6850	0.7913	0.7415	0.7913	0.0	0.7916	0.7919

Framework versions

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