smangrul's picture
Model save
6fd31a6 verified
metadata
license: bigcode-openrail-m
library_name: peft
tags:
  - generated_from_trainer
base_model: bigcode/starcoder2-3b
model-index:
  - name: starcoder-3b-hugcoder-loftq
    results: []

starcoder-3b-hugcoder-loftq

This model is a fine-tuned version of bigcode/starcoder2-3b on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 0.6305

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: 0.0002
  • train_batch_size: 16
  • eval_batch_size: 16
  • seed: 11
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 64
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_ratio: 0.1
  • training_steps: 2000

Training results

Training Loss Epoch Step Validation Loss
1.2064 0.05 100 0.7876
0.9806 0.1 200 0.7596
0.8669 0.15 300 0.7397
0.8555 0.2 400 0.7204
0.8566 0.25 500 0.7058
0.7643 0.3 600 0.6911
0.7464 0.35 700 0.6801
0.734 0.4 800 0.6697
0.7209 0.45 900 0.6608
0.6477 0.5 1000 0.6537
0.7292 0.55 1100 0.6475
0.6612 0.6 1200 0.6434
0.6808 0.65 1300 0.6393
0.6586 0.7 1400 0.6365
0.662 0.75 1500 0.6333
0.7034 0.8 1600 0.6322
0.683 0.85 1700 0.6309
0.7074 0.9 1800 0.6303
0.6362 0.95 1900 0.6307
0.6617 1.0 2000 0.6305

Framework versions

  • PEFT 0.10.1.dev0
  • Transformers 4.40.0.dev0
  • Pytorch 2.2.1+cu121
  • Datasets 2.18.0
  • Tokenizers 0.15.2