metadata
license: apache-2.0
base_model:
- Qwen/Qwen2.5-7B-Instruct
- BioMistral/BioMistral-7B
- LumiOpen/Vigogne-Instruct-7B
tags:
- moe
- frankenmoe
- merge
- mergekit
- lazymergekit
- Qwen/Qwen2.5-7B-Instruct
- BioMistral/BioMistral-7B
- LumiOpen/Vigogne-Instruct-7B
Aion-Pulse
Aion-Pulse is a Mixture of Experts (MoE) made with the following models using LazyMergekit:
🧩 Configuration
base_model: Qwen/Qwen2.5-7B-Instruct
experts:
- source_model: Qwen/Qwen2.5-7B-Instruct
- source_model: BioMistral/BioMistral-7B
- source_model: LumiOpen/Vigogne-Instruct-7B
💻 Usage
!pip install -qU transformers bitsandbytes accelerate
from transformers import AutoTokenizer
import transformers
import torch
model = "Esteban111/Aion-Pulse"
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"])