0be52334109f7de89ae483ae7934e6ff

This model is a fine-tuned version of deepseek-ai/DeepSeek-R1-Distill-Qwen-7B on the contemmcm/trec dataset. It achieves the following results on the evaluation set:

• Loss: 1.0305
• Data Size: 1.0
• Epoch Runtime: 162.3387
• Accuracy: 0.9625
• F1 Macro: 0.9522

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	24.0991	0	2.1170	0.1146	0.0434
No log	1	170	25.0678	0.0078	3.2572	0.2896	0.1368
No log	2	340	4.9327	0.0156	12.5939	0.5979	0.5061
No log	3	510	1.9932	0.0312	26.1874	0.8396	0.7274
No log	4	680	2.2692	0.0625	37.0808	0.8187	0.7867
0.2962	5	850	1.0584	0.125	51.6924	0.9417	0.9259
0.2962	6	1020	1.3186	0.25	72.8640	0.9229	0.9305
1.1754	7	1190	0.8004	0.5	111.3242	0.9563	0.9492
0.7022	8.0	1360	0.9660	1.0	178.2920	0.9583	0.9631
0.382	9.0	1530	1.1431	1.0	153.1476	0.9604	0.9644
0.5167	10.0	1700	0.9932	1.0	164.1829	0.9667	0.9692
0.3482	11.0	1870	1.0305	1.0	162.3387	0.9625	0.9522

Framework versions

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