30e7e9f535d84b26ed05320f0208e23e

This model is a fine-tuned version of Qwen/Qwen2.5-0.5B on the contemmcm/trec dataset. It achieves the following results on the evaluation set:

• Loss: 2.6133
• Data Size: 1.0
• Epoch Runtime: 32.1577
• Accuracy: 0.9417
• F1 Macro: 0.9341

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	24.9117	0	2.3755	0.0979	0.0600
No log	1	170	32.4771	0.0078	2.3103	0.3542	0.1750
No log	2	340	13.7967	0.0156	3.1098	0.3833	0.2262
No log	3	510	3.2921	0.0312	4.3444	0.7375	0.5738
No log	4	680	1.4357	0.0625	5.4958	0.8625	0.8626
0.3942	5	850	1.3166	0.125	7.4749	0.8583	0.8146
0.3942	6	1020	1.7033	0.25	11.6296	0.9021	0.9131
1.5578	7	1190	1.1367	0.5	18.6451	0.9062	0.8893
0.9441	8.0	1360	1.0812	1.0	33.8773	0.9542	0.9497
0.6308	9.0	1530	1.4185	1.0	32.5124	0.9333	0.9445
0.4831	10.0	1700	0.8470	1.0	32.1874	0.9542	0.9596
0.4231	11.0	1870	1.5657	1.0	34.3801	0.9458	0.9455
0.1869	12.0	2040	1.1573	1.0	32.6041	0.9521	0.9596
0.1897	13.0	2210	2.4231	1.0	32.4027	0.9375	0.9388
0.1618	14.0	2380	2.6133	1.0	32.1577	0.9417	0.9341

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.2.0
• Tokenizers 0.22.1