baseline-qwen3-4b-grounded_table/README.md

library_name, license, base_model, tags, model-index

library_name	license	base_model	tags	model-index
transformers	other	Qwen/Qwen3-4B	llama-factory full generated_from_trainer	name results sft_grounded_table

sft_grounded_table

This model is a fine-tuned version of Qwen/Qwen3-4B on the sunny_reasoning dataset. It achieves the following results on the evaluation set:

• Loss: 0.0085

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: 4e-05
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• distributed_type: multi-GPU
• num_devices: 2
• gradient_accumulation_steps: 32
• total_train_batch_size: 64
• total_eval_batch_size: 2
• 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: 3.0

Training results

Training Loss	Epoch	Step	Validation Loss
0.007	0.3397	92	0.0101
0.0078	0.6794	184	0.0117
0.0027	1.0185	276	0.0084
0.0084	1.3581	368	0.0087
0.0061	1.6978	460	0.0078
0.0037	2.0369	552	0.0077
0.0027	2.3766	644	0.0085
0.0041	2.7163	736	0.0085

Framework versions

• Transformers 4.56.2
• Pytorch 2.11.0+cu128
• Datasets 3.0.0
• Tokenizers 0.22.2