b3e36b8ecaa0de9195fc36f8913303b8

This model is a fine-tuned version of distilbert/distilbert-base-uncased-distilled-squad on the contemmcm/trec dataset. It achieves the following results on the evaluation set:

• Loss: 0.2582
• Data Size: 1.0
• Epoch Runtime: 5.4157
• Accuracy: 0.9667
• F1 Macro: 0.9567

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
No log	0	0	1.8271	0	0.7633	0.1021	0.0430
No log	1	170	1.6646	0.0078	0.9938	0.3583	0.1980
No log	2	340	1.5154	0.0156	1.0284	0.5333	0.4079
No log	3	510	1.1134	0.0312	1.1448	0.9021	0.7548
No log	4	680	0.5776	0.0625	1.2973	0.9104	0.7653
0.0547	5	850	0.2355	0.125	1.7117	0.9521	0.8007
0.0547	6	1020	0.2012	0.25	2.1348	0.9542	0.9309
0.2029	7	1190	0.2062	0.5	3.2885	0.9563	0.9313
0.141	8.0	1360	0.1672	1.0	5.5069	0.9625	0.9589
0.0709	9.0	1530	0.1640	1.0	5.4317	0.975	0.9776
0.0595	10.0	1700	0.1848	1.0	5.3818	0.9646	0.9432
0.0368	11.0	1870	0.1740	1.0	5.4152	0.9688	0.9647
0.0257	12.0	2040	0.3008	1.0	5.3703	0.9563	0.9349
0.0137	13.0	2210	0.2582	1.0	5.4157	0.9667	0.9567

Framework versions

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