b66d0c16762c6bd0d744e1c31b678fb0

This model is a fine-tuned version of distilbert/distilgpt2 on the google/boolq dataset. It achieves the following results on the evaluation set:

• Loss: 0.9999
• Data Size: 1.0
• Epoch Runtime: 14.9837
• Accuracy: 0.6998
• F1 Macro: 0.6833
• Rouge1: 0.6998
• Rouge2: 0.0
• Rougel: 0.6998
• Rougelsum: 0.7001

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	5.1151	0	2.2864	0.3790	0.2752	0.3790	0.0	0.3796	0.3793
No log	1	294	0.9175	0.0078	2.9792	0.5153	0.4966	0.5153	0.0	0.5153	0.5150
No log	2	588	0.8855	0.0156	2.6791	0.5904	0.4631	0.5907	0.0	0.5901	0.5895
No log	3	882	0.8422	0.0312	2.9943	0.6085	0.4071	0.6088	0.0	0.6078	0.6081
0.0378	4	1176	0.6656	0.0625	3.2166	0.6060	0.4785	0.6063	0.0	0.6060	0.6057
0.0554	5	1470	0.6731	0.125	3.8802	0.6170	0.3905	0.6170	0.0	0.6164	0.6167
0.097	6	1764	0.6451	0.25	5.3133	0.6345	0.5551	0.6345	0.0	0.6348	0.6345
0.6219	7	2058	0.6136	0.5	8.7357	0.6713	0.6326	0.6713	0.0	0.6708	0.6716
0.5681	8.0	2352	0.5862	1.0	13.8940	0.6857	0.6470	0.6866	0.0	0.6860	0.6860
0.4887	9.0	2646	0.6370	1.0	13.8556	0.6936	0.6452	0.6933	0.0	0.6933	0.6936
0.3573	10.0	2940	0.7260	1.0	13.6806	0.7031	0.6858	0.7028	0.0	0.7031	0.7028
0.3033	11.0	3234	0.8040	1.0	13.7892	0.7065	0.6831	0.7062	0.0	0.7062	0.7065
0.2457	12.0	3528	0.9999	1.0	14.9837	0.6998	0.6833	0.6998	0.0	0.6998	0.7001

Framework versions

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