aditya369-v2

This model is a fine-tuned version of Qwen/Qwen3-4B on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1339
• Macro F1: 0.7762
• Micro F1: 0.7907
• Exact Match: 0.7268

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: 16
• eval_batch_size: 32
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Use OptimizerNames.PAGED_ADAMW_8BIT with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 371
• num_epochs: 2

Training results

Training Loss	Epoch	Step	Exact Match	Validation Loss	Macro F1	Micro F1
No log	0.0054	10	0.2548	0.4154	0.0683	0.1915
0.3120	0.5385	1000	0.6955	0.1524	0.7235	0.7521
0.2348	1.0770	2000	0.7191	0.1399	0.7714	0.7881
0.2527	1.6155	3000	0.1324	0.7750	0.7935	0.7265
0.2141	2.0	3714	0.1327	0.7810	0.7971	0.7307

Framework versions

• PEFT 0.21.2
• Transformers 5.18.0
• Pytorch 2.11.0+cu128
• Datasets 5.0.1
• Tokenizers 0.23.1