MATH_training_split_distill_large_teacher

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

• Loss: 0.1529

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: 1e-05
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• distributed_type: multi-GPU
• gradient_accumulation_steps: 8
• total_train_batch_size: 8
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• num_epochs: 2

Training results

Training Loss	Epoch	Step	Validation Loss
0.1932	0.1781	100	0.1786
0.201	0.3561	200	0.1736
0.1219	0.5342	300	0.1718
0.1642	0.7122	400	0.1627
0.1437	0.8903	500	0.1561
0.0678	1.0677	600	0.1595
0.0657	1.2457	700	0.1616
0.0602	1.4238	800	0.1589
0.0591	1.6018	900	0.1550
0.0572	1.7799	1000	0.1529
0.0592	1.9579	1100	0.1528

Framework versions

• Transformers 4.49.0
• Pytorch 2.7.0+cu126
• Datasets 3.2.0
• Tokenizers 0.21.0