3ede210cfebe72448f6c40df8f7d34a2

This model is a fine-tuned version of distilbert/distilgpt2 on the nyu-mll/glue [qqp] dataset. It achieves the following results on the evaluation set:

• Loss: 0.3536
• Data Size: 1.0
• Epoch Runtime: 468.2319
• Accuracy: 0.8833
• F1 Macro: 0.8776
• Rouge1: 0.8833
• Rouge2: 0.0
• Rougel: 0.8834
• Rougelsum: 0.8834

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	2.4921	0	20.5944	0.6312	0.3879	0.6310	0.0	0.6311	0.6310
0.7052	1	11370	0.5189	0.0078	25.7257	0.7308	0.7027	0.7307	0.0	0.7309	0.7307
0.4976	2	22740	0.4820	0.0156	27.0800	0.7508	0.7442	0.7509	0.0	0.7508	0.7508
0.461	3	34110	0.4468	0.0312	33.9321	0.7721	0.7348	0.7721	0.0	0.7722	0.7721
0.4139	4	45480	0.4005	0.0625	47.7996	0.8104	0.8032	0.8103	0.0	0.8105	0.8104
0.3849	5	56850	0.3750	0.125	74.9860	0.8211	0.8146	0.8210	0.0	0.8211	0.8211
0.3477	6	68220	0.3379	0.25	128.1065	0.8433	0.8371	0.8433	0.0	0.8432	0.8434
0.2894	7	79590	0.3185	0.5	242.1527	0.8589	0.8530	0.8589	0.0	0.8589	0.8588
0.2851	8.0	90960	0.2862	1.0	460.4774	0.8795	0.8725	0.8795	0.0	0.8795	0.8795
0.2225	9.0	102330	0.2935	1.0	464.4146	0.8794	0.8738	0.8794	0.0	0.8795	0.8794
0.185	10.0	113700	0.2882	1.0	474.8970	0.8885	0.8814	0.8885	0.0	0.8886	0.8885
0.1517	11.0	125070	0.3171	1.0	475.4339	0.8862	0.8801	0.8862	0.0	0.8862	0.8862
0.1442	12.0	136440	0.3536	1.0	468.2319	0.8833	0.8776	0.8833	0.0	0.8834	0.8834

Framework versions

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