metadata
license: other
base_model: larryvrh/Yi-34B-200K-Llamafied
tags:
- generated_from_trainer
model-index:
- name: capybara-v4-yi-34b-200k
results: []
capybara-v4-yi-34b-200k
This model is a fine-tuned version of larryvrh/Yi-34B-200K-Llamafied on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.3638
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: 1
- eval_batch_size: 1
- seed: 42
- distributed_type: multi-GPU
- num_devices: 8
- gradient_accumulation_steps: 8
- total_train_batch_size: 64
- total_eval_batch_size: 8
- optimizer: Adam with betas=(0.9,0.95) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_steps: 50
- num_epochs: 3
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 0.7201 | 0.31 | 200 | 0.7801 |
| 0.716 | 0.62 | 400 | 0.7240 |
| 0.6128 | 0.93 | 600 | 0.6696 |
| 0.4111 | 1.24 | 800 | 0.6016 |
| 0.415 | 1.55 | 1000 | 0.5395 |
| 0.3293 | 1.86 | 1200 | 0.4782 |
| 0.3271 | 2.17 | 1400 | 0.4272 |
| 0.2672 | 2.49 | 1600 | 0.3925 |
| 0.2129 | 2.8 | 1800 | 0.3638 |
Framework versions
- Transformers 4.34.1
- Pytorch 2.0.1
- Datasets 2.14.6
- Tokenizers 0.14.1