|
|
--- |
|
|
language: |
|
|
- en |
|
|
tags: |
|
|
- pytorch |
|
|
- causal-lm |
|
|
- bitnet |
|
|
- layer-skipping |
|
|
- safetensors |
|
|
--- |
|
|
|
|
|
# bitskip2 |
|
|
|
|
|
This is a fine-tuned BitNet model with layer skipping capabilities, converted to SafeTensors format for efficient loading. |
|
|
|
|
|
## Model Details |
|
|
|
|
|
- **Model Type**: BitNet with Layer Skipping |
|
|
- **Base Model**: Unknown |
|
|
- **Architecture**: Unknown |
|
|
- **Format**: SafeTensors (efficient loading) |
|
|
|
|
|
|
|
|
## Usage |
|
|
|
|
|
```python |
|
|
from transformers import AutoModelForCausalLM, AutoTokenizer |
|
|
|
|
|
model = AutoModelForCausalLM.from_pretrained("USERNAME/MODEL_NAME") |
|
|
tokenizer = AutoTokenizer.from_pretrained("USERNAME/MODEL_NAME") |
|
|
|
|
|
# Generate text |
|
|
inputs = tokenizer("Hello, how are you?", return_tensors="pt") |
|
|
outputs = model.generate(**inputs, max_new_tokens=50) |
|
|
print(tokenizer.decode(outputs[0], skip_special_tokens=True)) |
|
|
``` |
|
|
|
|
|
## Training |
|
|
|
|
|
This model was trained using the LayerSkip framework with BitNet architecture and converted to SafeTensors format for efficient deployment. |
|
|
|
|
|
## Model Format |
|
|
|
|
|
This model is saved in SafeTensors format, which provides: |
|
|
- Faster loading times |
|
|
- Memory safety |
|
|
- Better error handling |
|
|
- Cross-platform compatibility |
|
|
|
|
|
## License |
|
|
|
|
|
[Add your license information here] |
|
|
|
