mistral7bit-lora-sql

This model is a fine-tuned version of mistralai/Mistral-7B-v0.1 on the generator dataset. It achieves the following results on the evaluation set:

• Loss: 0.3405

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: 0.0003
• train_batch_size: 8
• eval_batch_size: 8
• seed: 1399
• gradient_accumulation_steps: 4
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 100
• training_steps: 500

Training results

Training Loss	Epoch	Step	Validation Loss
0.7523	0.0565	20	0.5037
0.4763	0.1129	40	0.4221
0.4231	0.1694	60	0.3944
0.4051	0.2258	80	0.3833
0.396	0.2823	100	0.3756
0.3885	0.3387	120	0.3717
0.3829	0.3952	140	0.3651
0.3752	0.4517	160	0.3605
0.3752	0.5081	180	0.3565
0.3688	0.5646	200	0.3551
0.3655	0.6210	220	0.3530
0.3596	0.6775	240	0.3472
0.3565	0.7339	260	0.3460
0.3556	0.7904	280	0.3448
0.3553	0.8469	300	0.3440
0.3496	0.9033	320	0.3424
0.3446	0.9598	340	0.3405
0.3328	1.0162	360	0.3402
0.3035	1.0727	380	0.3405

Framework versions

• PEFT 0.12.0
• Transformers 4.44.0
• Pytorch 2.2.2+cu121
• Datasets 2.21.0
• Tokenizers 0.19.1