Mistral-7B-text-to-sql-flash-attention-2-FAISS

This model is a fine-tuned version of mistralai/Mistral-7B-Instruct-v0.3 on the generator dataset. It achieves the following results on the evaluation set:

• Loss: 0.4583

Model description

Article: https://medium.com/@frankmorales_91352/faiss-powered-semantic-search-meets-fine-tuned-mistral-a-novel-approach-to-text-to-sql-generation-fc633d9c1bc2

Training and evaluation data

Training: https://github.com/frank-morales2020/MLxDL/blob/main/FAISS_FINETUNING.ipynb

Evaluation: https://github.com/frank-morales2020/MLxDL/blob/main/FAISS_Evaluator_Mistral_7B_text_to_sql.ipynb

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0002
• train_batch_size: 3
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 24
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: constant
• lr_scheduler_warmup_ratio: 0.03
• lr_scheduler_warmup_steps: 15
• num_epochs: 3

Training results

Training Loss	Epoch	Step	Validation Loss
1.8409	0.4	10	0.6999
0.616	0.8	20	0.5322
0.4977	1.2	30	0.4910
0.4486	1.6	40	0.4661
0.4313	2.0	50	0.4529
0.36	2.4	60	0.4620
0.3534	2.8	70	0.4583

Framework versions

• PEFT 0.11.1
• Transformers 4.42.3
• Pytorch 2.3.0+cu121
• Datasets 2.20.0
• Tokenizers 0.19.1