|
|
--- |
|
|
license: cc-by-nc-4.0 |
|
|
tags: |
|
|
- merge |
|
|
- mergekit |
|
|
- lazymergekit |
|
|
base_model: |
|
|
- shadowml/WestBeagle-7B |
|
|
- alnrg2arg/test3_sft_16bit |
|
|
--- |
|
|
|
|
|
# FoxBeagle-7B |
|
|
|
|
|
FoxBeagle-7B is a merge of the following models using [LazyMergekit](https://colab.research.google.com/drive/1obulZ1ROXHjYLn6PPZJwRR6GzgQogxxb?usp=sharing): |
|
|
* [shadowml/WestBeagle-7B](https://huggingface.co/shadowml/WestBeagle-7B) |
|
|
* [alnrg2arg/test3_sft_16bit](https://huggingface.co/alnrg2arg/test3_sft_16bit) |
|
|
|
|
|
## 🧩 Configuration |
|
|
|
|
|
```yaml |
|
|
slices: |
|
|
- sources: |
|
|
- model: shadowml/WestBeagle-7B |
|
|
layer_range: [0, 32] |
|
|
- model: alnrg2arg/test3_sft_16bit |
|
|
layer_range: [0, 32] |
|
|
merge_method: slerp |
|
|
base_model: shadowml/WestBeagle-7B |
|
|
parameters: |
|
|
t: |
|
|
- filter: self_attn |
|
|
value: [0, 0.5, 0.3, 0.7, 1] |
|
|
- filter: mlp |
|
|
value: [1, 0.5, 0.7, 0.3, 0] |
|
|
- value: 0.5 |
|
|
dtype: bfloat16 |
|
|
``` |
|
|
|
|
|
## 💻 Usage |
|
|
|
|
|
```python |
|
|
!pip install -qU transformers accelerate |
|
|
|
|
|
from transformers import AutoTokenizer |
|
|
import transformers |
|
|
import torch |
|
|
|
|
|
model = "shadowml/FoxBeagle-7B" |
|
|
messages = [{"role": "user", "content": "What is a large language model?"}] |
|
|
|
|
|
tokenizer = AutoTokenizer.from_pretrained(model) |
|
|
prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True) |
|
|
pipeline = transformers.pipeline( |
|
|
"text-generation", |
|
|
model=model, |
|
|
torch_dtype=torch.float16, |
|
|
device_map="auto", |
|
|
) |
|
|
|
|
|
outputs = pipeline(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95) |
|
|
print(outputs[0]["generated_text"]) |
|
|
``` |
