Hugging Face's logo

wookidoki
/
autofix10k

PEFT
Safetensors
Generated from Trainer
Model card Files Community
autofix10k / README.md
wookidoki's picture
wookidoki
End of training
8dbfbc1 verified
preview code
|
raw
history blame
4.4 kB
metadata
license: llama2
base_model: codellama/CodeLlama-7b-hf
tags:
  - generated_from_trainer
model-index:
  - name: autofix10k
    results: []
library_name: peft

autofix10k

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: 0.4372

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

The following bitsandbytes quantization config was used during training:

  • quant_method: QuantizationMethod.BITS_AND_BYTES
  • _load_in_8bit: True
  • _load_in_4bit: False
  • llm_int8_threshold: 6.0
  • llm_int8_skip_modules: None
  • llm_int8_enable_fp32_cpu_offload: False
  • llm_int8_has_fp16_weight: False
  • bnb_4bit_quant_type: fp4
  • bnb_4bit_use_double_quant: False
  • bnb_4bit_compute_dtype: float32
  • bnb_4bit_quant_storage: uint8
  • load_in_4bit: False
  • load_in_8bit: True

Training hyperparameters

The following hyperparameters were used during training:

  • learning_rate: 0.0003
  • train_batch_size: 1
  • eval_batch_size: 8
  • seed: 42
  • gradient_accumulation_steps: 8
  • total_train_batch_size: 8
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • num_epochs: 10
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss
0.7922 0.2 20 0.5237
0.5053 0.4 40 0.4857
0.4071 0.6 60 0.4356
0.4297 0.8 80 0.4154
0.5313 1.0 100 0.3827
0.4814 1.2 120 0.3785
0.3739 1.4 140 0.3774
0.3279 1.6 160 0.3761
0.3149 1.8 180 0.3732
0.4086 2.0 200 0.3658
0.3724 2.2 220 0.3664
0.3691 2.4 240 0.3644
0.3065 2.6 260 0.3679
0.2688 2.8 280 0.3767
0.3431 3.0 300 0.3633
0.333 3.2 320 0.3641
0.3052 3.4 340 0.3597
0.2444 3.6 360 0.3779
0.2455 3.8 380 0.3712
0.3078 4.0 400 0.3578
0.2877 4.2 420 0.3650
0.2659 4.4 440 0.3731
0.2496 4.6 460 0.3764
0.218 4.8 480 0.3781
0.219 5.0 500 0.3742
0.2119 5.2 520 0.3808
0.2435 5.4 540 0.3871
0.2331 5.6 560 0.3818
0.1738 5.8 580 0.3758
0.1772 6.0 600 0.3731
0.1607 6.2 620 0.4121
0.1942 6.4 640 0.3943
0.2312 6.6 660 0.3867
0.1528 6.8 680 0.4160
0.1155 7.0 700 0.4100
0.1495 7.2 720 0.4081
0.1674 7.4 740 0.4015
0.1849 7.6 760 0.4075
0.1231 7.8 780 0.4238
0.0905 8.0 800 0.4128
0.1156 8.2 820 0.4278
0.1628 8.4 840 0.4203
0.1545 8.6 860 0.4219
0.1236 8.8 880 0.4294
0.0799 9.0 900 0.4224
0.0991 9.2 920 0.4399
0.1176 9.4 940 0.4350
0.1711 9.6 960 0.4362
0.1106 9.8 980 0.4414
0.0582 10.0 1000 0.4372

Framework versions

  • PEFT 0.4.0
  • Transformers 4.40.0.dev0
  • Pytorch 2.2.0+cu121
  • Datasets 2.17.1
  • Tokenizers 0.15.2