falcon-7b-ft-codeAlpaca_20k-v2
This model is a fine-tuned version of ybelkada/falcon-7b-sharded-bf16 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.7367
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: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: constant
- lr_scheduler_warmup_ratio: 0.03
- training_steps: 550
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 0.7985 | 0.35 | 100 | 0.7680 |
| 0.7259 | 0.71 | 200 | 0.7499 |
| 0.6691 | 1.06 | 300 | 0.7480 |
| 0.6873 | 1.42 | 400 | 0.7423 |
| 0.5799 | 1.77 | 500 | 0.7367 |
Framework versions
- Transformers 4.31.0.dev0
- Pytorch 2.0.1+cu118
- Datasets 2.13.0
- Tokenizers 0.13.3