Extractor_Adaptor_Qwen3_Final_SCHEMA_Special

This model is a fine-tuned version of Qwen/Qwen3-0.6B on the web_finetune_train dataset. It achieves the following results on the evaluation set:

• Loss: 0.0385

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: 2
• eval_batch_size: 2
• seed: 42
• distributed_type: multi-GPU
• num_devices: 2
• gradient_accumulation_steps: 8
• total_train_batch_size: 32
• total_eval_batch_size: 4
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED 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: 2.0
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.0474	0.1193	50	0.0482
0.0475	0.2385	100	0.0466
0.0301	0.3578	150	0.0448
0.0418	0.4770	200	0.0438
0.0444	0.5963	250	0.0437
0.0333	0.7156	300	0.0420
0.0297	0.8348	350	0.0409
0.0216	0.9541	400	0.0399
0.0182	1.0716	450	0.0401
0.02	1.1908	500	0.0402
0.0271	1.3101	550	0.0394
0.0244	1.4293	600	0.0394
0.0185	1.5486	650	0.0389
0.0223	1.6679	700	0.0386
0.0186	1.7871	750	0.0385
0.0255	1.9064	800	0.0385

Framework versions

• PEFT 0.15.2
• Transformers 4.57.1
• Pytorch 2.9.0+cu128
• Datasets 3.6.0
• Tokenizers 0.22.1