metadata
license: bigcode-openrail-m
library_name: peft
tags:
- generated_from_trainer
base_model: bigcode/starcoderbase-1b
model-index:
- name: peft-lora-starcoder1B-Instruction-ny8
results: []
peft-lora-starcoder1B-Instruction-ny8
This model is a fine-tuned version of bigcode/starcoderbase-1b on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.7359
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: 0.0005
- train_batch_size: 4
- eval_batch_size: 4
- seed: 42
- gradient_accumulation_steps: 4
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 30
- training_steps: 2000
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 0.2429 | 0.05 | 100 | 0.2525 |
| 0.2099 | 0.1 | 200 | 0.2812 |
| 0.0957 | 0.15 | 300 | 0.4394 |
| 0.0277 | 0.2 | 400 | 0.5758 |
| 0.015 | 0.25 | 500 | 0.6307 |
| 0.0144 | 0.3 | 600 | 0.6582 |
| 0.0122 | 0.35 | 700 | 0.6811 |
| 0.0105 | 0.4 | 800 | 0.6984 |
| 0.0116 | 0.45 | 900 | 0.7030 |
| 0.0101 | 0.5 | 1000 | 0.7078 |
| 0.0097 | 0.55 | 1100 | 0.7047 |
| 0.0091 | 0.6 | 1200 | 0.7144 |
| 0.0087 | 0.65 | 1300 | 0.7196 |
| 0.0075 | 0.7 | 1400 | 0.7318 |
| 0.0082 | 0.75 | 1500 | 0.7242 |
| 0.008 | 0.8 | 1600 | 0.7289 |
| 0.0078 | 0.85 | 1700 | 0.7322 |
| 0.0074 | 0.9 | 1800 | 0.7398 |
| 0.0075 | 0.95 | 1900 | 0.7349 |
| 0.0073 | 1.0 | 2000 | 0.7359 |
Framework versions
- PEFT 0.7.1
- Transformers 4.36.2
- Pytorch 2.1.2+cu121
- Datasets 2.16.1
- Tokenizers 0.15.0