Qwen3-30B-A3B-Base-medqa-seed-2405

This model is a fine-tuned version of Qwen/Qwen3-30B-A3B-Base on the medqa dataset. It achieves the following results on the evaluation set:

• Loss: 0.0279

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: 0.0002
• train_batch_size: 2
• eval_batch_size: 2
• seed: 2405
• distributed_type: multi-GPU
• num_devices: 128
• gradient_accumulation_steps: 8
• total_train_batch_size: 2048
• total_eval_batch_size: 256
• optimizer: Use OptimizerNames.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: 0.1
• num_epochs: 3.0

Training results

Training Loss	Epoch	Step	Validation Loss
2.0528	0.4651	10	1.8944
0.316	0.9302	20	0.1930
0.0322	1.3721	30	0.0318
0.0229	1.8372	40	0.0289
0.0206	2.2791	50	0.0280
0.0238	2.7442	60	0.0280

Framework versions

• PEFT 0.15.2
• Transformers 4.52.3
• Pytorch 2.7.0+cu126
• Datasets 3.6.0
• Tokenizers 0.21.1