PEFT_qwen2.5_coder_3b_probgate_schema_aware_finetuned_only_answerable_delimeters_eos

This model is a fine-tuned version of Qwen/Qwen2.5-Coder-3B on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0637

The model was trained using the <start_sql> and <end_sql> delimiters and the <|end_of_text> token.

LoRA config:

• task_type='CAUSAL_LM'
• r=128
• target_modules={'embed_tokens', 'up_proj', 'gate_proj', 'down_proj', 'v_proj', 'o_proj', 'q_proj', 'lm_head', 'k_proj'}
• lora_alpha=256
• lora_dropout=0.1

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-05
• train_batch_size: 1
• eval_batch_size: 1
• seed: 42
• 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
• num_epochs: 3
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0	0	1.1647
0.0569	1.0	4674	0.0607
0.0519	2.0	9348	0.0608
0.0486	3.0	14022	0.0637

Framework versions

• PEFT 0.14.0
• Transformers 4.48.0
• Pytorch 2.7.1+cu126
• Datasets 3.2.0
• Tokenizers 0.21.0