metadata
license: bigcode-openrail-m
base_model: bigcode/starcoderbase-3b
tags:
- generated_from_trainer
model-index:
- name: outputs
results: []
outputs
This model is a fine-tuned version of bigcode/starcoderbase-3b on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.5539
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: 24
- eval_batch_size: 24
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 2
- training_steps: 1500
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 0.6249 | 0.07 | 50 | 0.7391 |
| 0.367 | 0.14 | 100 | 0.6978 |
| 0.1705 | 0.21 | 150 | 0.6762 |
| 0.0863 | 0.29 | 200 | 0.6702 |
| 0.4031 | 0.36 | 250 | 0.6393 |
| 0.5025 | 0.43 | 300 | 0.6202 |
| 0.4156 | 0.5 | 350 | 0.6114 |
| 0.3885 | 0.57 | 400 | 0.6040 |
| 0.0186 | 0.64 | 450 | 0.6420 |
| 0.4009 | 0.71 | 500 | 0.6055 |
| 0.3831 | 0.79 | 550 | 0.5867 |
| 0.4876 | 0.86 | 600 | 0.5828 |
| 0.3987 | 0.93 | 650 | 0.5813 |
| 0.2669 | 1.0 | 700 | 0.5810 |
| 0.5006 | 0.75 | 750 | 0.5776 |
| 0.4844 | 1.02 | 800 | 0.5748 |
| 0.3621 | 1.07 | 850 | 0.5747 |
| 0.2021 | 1.12 | 900 | 0.5709 |
| 0.1454 | 1.17 | 950 | 0.5720 |
| 0.5536 | 1.22 | 1000 | 0.5725 |
| 0.1201 | 0.7 | 1050 | 0.5649 |
| 0.3127 | 0.73 | 1100 | 0.5599 |
| 0.3316 | 0.77 | 1150 | 0.5582 |
| 0.2635 | 0.8 | 1200 | 0.5580 |
| 0.5157 | 0.83 | 1250 | 0.5573 |
| 0.395 | 0.87 | 1300 | 0.5554 |
| 0.2694 | 0.9 | 1350 | 0.5543 |
| 0.4236 | 0.93 | 1400 | 0.5543 |
| 0.4869 | 0.97 | 1450 | 0.5541 |
| 0.3642 | 1.0 | 1500 | 0.5539 |
Framework versions
- Transformers 4.36.0.dev0
- Pytorch 2.1.0+cu118
- Datasets 2.15.0
- Tokenizers 0.15.0