peft-lora-starcoder15B-personal-copilot-A100-40GB-colab
This model is a fine-tuned version of bigcode/starcoder on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.3633
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:
- load_in_8bit: False
- load_in_4bit: True
- 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: nf4
- bnb_4bit_use_double_quant: True
- bnb_4bit_compute_dtype: bfloat16
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0002
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 2
- total_train_batch_size: 16
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: constant
- lr_scheduler_warmup_steps: 30
- training_steps: 2000
Training results
Training Loss | Epoch | Step | Validation Loss |
---|---|---|---|
0.6593 | 0.05 | 100 | 0.5847 |
0.6226 | 0.1 | 200 | 0.5292 |
0.6597 | 0.15 | 300 | 0.4814 |
0.5523 | 0.2 | 400 | 0.4617 |
0.4856 | 0.25 | 500 | 0.4597 |
0.5237 | 0.3 | 600 | 0.4505 |
0.4894 | 0.35 | 700 | 0.4398 |
0.5579 | 0.4 | 800 | 0.4377 |
0.4702 | 0.45 | 900 | 0.4322 |
0.5418 | 0.5 | 1000 | 0.4244 |
0.5159 | 0.55 | 1100 | 0.4133 |
0.524 | 0.6 | 1200 | 0.3977 |
0.4138 | 0.65 | 1300 | 0.3966 |
0.5572 | 0.7 | 1400 | 0.3936 |
0.4146 | 0.75 | 1500 | 0.3904 |
0.7927 | 0.8 | 1600 | 0.3905 |
0.4131 | 0.85 | 1700 | 0.3866 |
0.4552 | 0.9 | 1800 | 0.3881 |
0.3914 | 0.95 | 1900 | 0.3794 |
0.4945 | 1.0 | 2000 | 0.3633 |
Framework versions
- PEFT 0.5.0.dev0
- Transformers 4.32.0.dev0
- Pytorch 2.0.1+cu118
- Datasets 2.14.4
- Tokenizers 0.13.3
- Downloads last month
- 0
Model tree for smangrul/peft-lora-starcoder15B-personal-copilot-A100-40GB-colab
Base model
bigcode/starcoder