741f73439c2dfa8d05bace436d284154

This model is a fine-tuned version of facebook/opt-125m on the contemmcm/trec dataset. It achieves the following results on the evaluation set:

• Loss: 0.3158
• Data Size: 1.0
• Epoch Runtime: 13.0821
• Accuracy: 0.9479
• F1 Macro: 0.9352

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	1.9194	0	1.2057	0.1292	0.0870
No log	1	170	1.5633	0.0078	1.4298	0.4313	0.2455
No log	2	340	1.2538	0.0156	1.2709	0.5813	0.4299
No log	3	510	0.5704	0.0312	1.6920	0.8125	0.6968
No log	4	680	0.2417	0.0625	1.9839	0.9354	0.9234
0.0372	5	850	0.2498	0.125	2.9048	0.9167	0.9020
0.0372	6	1020	0.1981	0.25	4.3300	0.95	0.9237
0.2046	7	1190	0.2539	0.5	7.2602	0.9062	0.9154
0.1463	8.0	1360	0.1276	1.0	13.4284	0.9688	0.9736
0.0864	9.0	1530	0.1417	1.0	13.1737	0.975	0.9616
0.0671	10.0	1700	0.2472	1.0	13.1137	0.95	0.9559
0.0201	11.0	1870	0.2529	1.0	13.2263	0.9563	0.9429
0.0262	12.0	2040	0.3158	1.0	13.0821	0.9479	0.9352

Framework versions

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