56 lines
1.7 KiB
Markdown
56 lines
1.7 KiB
Markdown
|
---
|
||
|
|
license: apache-2.0
|
||
|
|
tags:
|
||
|
|
- moe
|
||
|
|
- frankenmoe
|
||
|
|
- merge
|
||
|
|
- mergekit
|
||
|
|
- lazymergekit
|
||
|
|
- NeuralNovel/Valor-7B-v0.1
|
||
|
|
- Toten5/Marcoroni-neural-chat-7B-v1
|
||
|
|
base_model:
|
||
|
|
- NeuralNovel/Valor-7B-v0.1
|
||
|
|
- Toten5/Marcoroni-neural-chat-7B-v1
|
||
|
|
---
|
||
|
|
|
||
|
|
# Valor_Macaroni_moe
|
||
|
|
|
||
|
|
Valor_Macaroni_moe is a Mixure of Experts (MoE) made with the following models using [LazyMergekit](https://colab.research.google.com/drive/1obulZ1ROXHjYLn6PPZJwRR6GzgQogxxb?usp=sharing):
|
||
|
|
* [NeuralNovel/Valor-7B-v0.1](https://huggingface.co/NeuralNovel/Valor-7B-v0.1)
|
||
|
|
* [Toten5/Marcoroni-neural-chat-7B-v1](https://huggingface.co/Toten5/Marcoroni-neural-chat-7B-v1)
|
||
|
|
|
||
|
|
## 🧩 Configuration
|
||
|
|
|
||
|
|
```yaml
|
||
|
|
base_model: NeuralNovel/Valor-7B-v0.1
|
||
|
|
gate_mode: cheap_embed
|
||
|
|
experts:
|
||
|
|
- source_model: NeuralNovel/Valor-7B-v0.1
|
||
|
|
positive_prompts: ["What should I do if lost my mobile phone"]
|
||
|
|
- source_model: Toten5/Marcoroni-neural-chat-7B-v1
|
||
|
|
positive_prompts: ["I have 3 apples. I lost 2 out of it. After that my father gave me another 3. How many do I have now?"]
|
||
|
|
```
|
||
|
|
|
||
|
|
## 💻 Usage
|
||
|
|
|
||
|
|
```python
|
||
|
|
!pip install -qU transformers bitsandbytes accelerate
|
||
|
|
|
||
|
|
from transformers import AutoTokenizer
|
||
|
|
import transformers
|
||
|
|
import torch
|
||
|
|
|
||
|
|
model = "Vasanth/Valor_Macaroni_moe"
|
||
|
|
|
||
|
|
tokenizer = AutoTokenizer.from_pretrained(model)
|
||
|
|
pipeline = transformers.pipeline(
|
||
|
|
"text-generation",
|
||
|
|
model=model,
|
||
|
|
model_kwargs={"torch_dtype": torch.float16, "load_in_4bit": True},
|
||
|
|
)
|
||
|
|
|
||
|
|
messages = [{"role": "user", "content": "Explain what a Mixture of Experts is in less than 100 words."}]
|
||
|
|
prompt = pipeline.tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
|
||
|
|
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"])
|
||
|
|
```
|