Beyonder-4x7b

This model is a Mixure of Experts (MoE) made with mergekit (mixtral branch). It uses the following base models:

🧩 Configuration

base_model: openchat/openchat-3.5-1210
gate_mode: hidden
experts:
  - source_model: openchat/openchat-3.5-1210
    positive_prompts:
    - "chat"
    - "assistant"
    - "tell me"
    - "explain"
    negative_prompts:
    - "storywriting"
    - "mathematics"
    - "reasoning"
    - "code"
    - "programming"
  - source_model: beowolx/CodeNinja-1.0-OpenChat-7B
    positive_prompts:
    - "code"
    - "python"
    - "javascript"
    - "programming"
    - "algorithm"
    negative_prompts:
    - "chat"
    - "assistant"
    - "storywriting"
    - "mathematics"
    - "reasoning"
  - source_model: maywell/PiVoT-0.1-Starling-LM-RP
    positive_prompts:
    - "storywriting"
    - "write"
    - "scene"
    - "story"
    - "character"
    negative_prompts:
    - "chat"
    - "assistant"
    - "code"
    - "programming"
    - "mathematics"
    - "reasoning"
  - source_model: WizardLM/WizardMath-7B-V1.1
    positive_prompts:
    - "reason"
    - "math"
    - "mathematics"
    - "solve"
    - "count"
    negative_prompts:
    - "chat"
    - "assistant"
    - "code"
    - "programming"
    - "storywriting"

💻 Usage

!pip install -qU transformers bitsandbytes accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "mlabonne/Beyonder-4x7b"

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"])

Output:

A Mixture of Experts (MoE) is a neural network architecture that combines the strengths of multiple expert networks to make predictions. It leverages the idea of ensemble learning, where multiple models work together to improve performance. In each MoE, a gating network is used to select the most relevant expert for the input. The final output is a weighted combination of the expert outputs, determined by the gating network's predictions.