sql-code-llama

This model is a fine-tuned version of codellama/CodeLlama-7b-hf on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 0.0810

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: 16
  • eval_batch_size: 8
  • seed: 42
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 64
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 100
  • training_steps: 400
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss
2.033 0.28 20 1.9418
1.3136 0.56 40 0.8477
0.1674 0.83 60 0.1384
0.1276 1.11 80 0.1220
0.1106 1.39 100 0.1046
0.102 1.67 120 0.0946
0.0917 1.94 140 0.0903
0.0895 2.22 160 0.0887
0.0889 2.5 180 0.0872
0.0874 2.78 200 0.0858
0.086 3.06 220 0.0851
0.0861 3.33 240 0.0842
0.085 3.61 260 0.0835
0.0821 3.89 280 0.0830
0.0838 4.17 300 0.0823
0.0816 4.44 320 0.0820
0.0785 4.72 340 0.0815
0.0819 5.0 360 0.0812
0.081 5.28 380 0.0810
0.0765 5.56 400 0.0810

Framework versions

  • Transformers 4.36.2
  • Pytorch 2.0.1+cu117
  • Datasets 2.13.1
  • Tokenizers 0.15.0
Downloads last month
13
Inference API
Unable to determine this model’s pipeline type. Check the docs .

Model tree for diwakar21scout/sql-code-llama

Finetuned
(56)
this model