orm-big-math-digits-v2-correctness

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

• Loss: 0.4828
• Accuracy: 0.768

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-06
• train_batch_size: 1
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 64
• total_train_batch_size: 64
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.05
• training_steps: 300

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.5622	0.08	25	0.5628	0.717
0.4747	0.16	50	0.5106	0.747
0.4651	0.24	75	0.5284	0.73
0.4415	0.32	100	0.5070	0.755
0.4442	0.4	125	0.4878	0.76
0.4253	0.48	150	0.4866	0.757
0.4829	0.56	175	0.5158	0.741
0.4456	0.64	200	0.4799	0.76
0.4249	0.72	225	0.4830	0.766
0.4527	0.8	250	0.4816	0.764
0.4169	0.88	275	0.4833	0.763
0.4743	0.96	300	0.4828	0.768

Framework versions

• Transformers 4.48.3
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0