cutelemonlili/Qwen2.5-Math-7B-Instruct_lora_Teacher_forget_RL_data_QwQ-Preview

Teacher_forget_RL_data_QwQ-Preview

This model is a fine-tuned version of Qwen/Qwen2.5-Math-7B-Instruct on the Teacher_forget_RL_data_QwQ-Preview dataset. It achieves the following results on the evaluation set:

• Loss: 0.3734

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: 0.0001
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 4
• total_eval_batch_size: 4
• 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
• lr_scheduler_warmup_ratio: 0.03
• num_epochs: 3.0

Training results

Training Loss	Epoch	Step	Validation Loss
0.3612	0.3300	200	0.4400
0.4758	0.6601	400	0.3989
0.3383	0.9901	600	0.3835
0.265	1.3201	800	0.3786
0.3479	1.6502	1000	0.3753
0.2934	1.9802	1200	0.3716
0.3167	2.3102	1400	0.3741
0.3017	2.6403	1600	0.3737
0.2065	2.9703	1800	0.3734

Framework versions

• PEFT 0.12.0
• Transformers 4.46.1
• Pytorch 2.4.0+cu121
• Datasets 3.1.0
• Tokenizers 0.20.3