e55b927081ee75728ff75ffe9fd51ed8

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

• Loss: 0.7310
• Data Size: 1.0
• Epoch Runtime: 474.8626
• Accuracy: 0.7783
• F1 Macro: 0.7775
• Rouge1: 0.7780
• Rouge2: 0.0
• Rougel: 0.7784
• Rougelsum: 0.7783

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	3.9277	0	5.5043	0.3542	0.1762	0.3541	0.0	0.3542	0.3539
1.3243	1	12271	0.9965	0.0078	10.4020	0.5034	0.5056	0.5039	0.0	0.5034	0.5036
0.9367	2	24542	0.9056	0.0156	13.6934	0.5969	0.5856	0.5972	0.0	0.5967	0.5973
0.8383	3	36813	0.7913	0.0312	20.3280	0.6473	0.6415	0.6476	0.0	0.6475	0.6476
0.789	4	49084	0.7268	0.0625	35.4271	0.6858	0.6849	0.6859	0.0	0.6860	0.6860
0.6775	5	61355	0.6788	0.125	64.4232	0.7135	0.7128	0.7135	0.0	0.7137	0.7136
0.6607	6	73626	0.6413	0.25	121.3756	0.7334	0.7334	0.7335	0.0	0.7333	0.7331
0.5741	7	85897	0.5919	0.5	248.7705	0.7605	0.7586	0.7605	0.0	0.7604	0.7604
0.5431	8.0	98168	0.5731	1.0	480.2955	0.7761	0.7762	0.7760	0.0	0.7761	0.7765
0.446	9.0	110439	0.5703	1.0	496.2990	0.7799	0.7789	0.7797	0.0	0.7799	0.7802
0.4126	10.0	122710	0.5888	1.0	479.6829	0.7847	0.7835	0.7845	0.0	0.7848	0.7849
0.3776	11.0	134981	0.6292	1.0	473.5134	0.7851	0.7845	0.7850	0.0	0.7851	0.7853
0.3049	12.0	147252	0.6662	1.0	473.7821	0.7849	0.7839	0.7846	0.0	0.7849	0.7850
0.2573	13.0	159523	0.7310	1.0	474.8626	0.7783	0.7775	0.7780	0.0	0.7784	0.7783

Framework versions

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