fce2cb8459c5b8d8e073f3dd3208b9f9

This model is a fine-tuned version of deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B on the contemmcm/trec dataset. It achieves the following results on the evaluation set:

• Loss: 1.6428
• Data Size: 1.0
• Epoch Runtime: 44.6530
• Accuracy: 0.9625
• F1 Macro: 0.9607

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	12.0201	0	1.9207	0.0813	0.0675
No log	1	170	9.2042	0.0078	2.2438	0.3917	0.2990
No log	2	340	4.0873	0.0156	3.9739	0.6813	0.5584
No log	3	510	0.9302	0.0312	6.9575	0.9167	0.8854
No log	4	680	1.0604	0.0625	10.1887	0.9104	0.8001
0.1471	5	850	0.8138	0.125	13.3165	0.9458	0.9257
0.1471	6	1020	0.7508	0.25	20.0347	0.9542	0.9596
0.8721	7	1190	0.8174	0.5	30.1888	0.9625	0.9579
0.6083	8.0	1360	0.7073	1.0	52.3669	0.9646	0.9613
0.4208	9.0	1530	1.0547	1.0	41.1186	0.9667	0.9682
0.2781	10.0	1700	1.2486	1.0	42.5361	0.9667	0.9555
0.1435	11.0	1870	1.0317	1.0	45.0078	0.9708	0.9736
0.0281	12.0	2040	1.6428	1.0	44.6530	0.9625	0.9607

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.0.0
• Tokenizers 0.22.1