3f7d69c9961877061906f84a6ab83765

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

• Loss: 0.3416
• Data Size: 1.0
• Epoch Runtime: 806.2111
• Accuracy: 0.8965
• F1 Macro: 0.8891
• Rouge1: 0.8964
• Rouge2: 0.0
• Rougel: 0.8965
• Rougelsum: 0.8964

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	4.4916	0	33.0682	0.5943	0.4279	0.5943	0.0	0.5944	0.5942
0.8783	1	11370	0.4956	0.0078	40.5802	0.7414	0.7053	0.7414	0.0	0.7415	0.7412
0.4892	2	22740	0.4321	0.0156	44.6986	0.7844	0.7775	0.7843	0.0	0.7844	0.7844
0.436	3	34110	0.3975	0.0312	57.0591	0.8136	0.7956	0.8136	0.0	0.8135	0.8135
0.3809	4	45480	0.3521	0.0625	85.0108	0.8413	0.8321	0.8414	0.0	0.8413	0.8413
0.3387	5	56850	0.3335	0.125	129.3174	0.8488	0.8426	0.8487	0.0	0.8488	0.8488
0.3088	6	68220	0.3078	0.25	224.5243	0.8637	0.8575	0.8637	0.0	0.8637	0.8637
0.2637	7	79590	0.2818	0.5	417.3933	0.8777	0.8717	0.8777	0.0	0.8777	0.8778
0.2632	8.0	90960	0.2594	1.0	800.5520	0.8928	0.8862	0.8928	0.0	0.8928	0.8928
0.1911	9.0	102330	0.2634	1.0	800.5920	0.8926	0.8870	0.8927	0.0	0.8927	0.8927
0.1407	10.0	113700	0.3041	1.0	802.9720	0.8963	0.8893	0.8962	0.0	0.8962	0.8962
0.126	11.0	125070	0.3177	1.0	804.8651	0.8980	0.8917	0.8980	0.0	0.8980	0.8979
0.1013	12.0	136440	0.3416	1.0	806.2111	0.8965	0.8891	0.8964	0.0	0.8965	0.8964

Framework versions

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