Eclipse-mistral-4x7b

Eclipse-mistral-4x7b is a Mixture of Experts (MoE) made with the following models:

• SanjiWatsuki/Kunoichi-DPO-v2-7B
• NousResearch/Hermes-2-Pro-Mistral-7B
• MaziyarPanahi/Calme-7B-Instruct-v0.9
• meta-math/MetaMath-Mistral-7B

🧩 Configuration

base_model: HuggingFaceH4/zephyr-7b-beta
gate_mode: hidden
dtype: bfloat16
experts:
    - source_model: SanjiWatsuki/Kunoichi-DPO-v2-7B
      positive_prompts:
      - "storywriting"
      - "write"
      - "scene"
      - "story"
      - "character"
    - source_model: NousResearch/Hermes-2-Pro-Mistral-7B
      positive_prompts:
      - "chat"
      - "assistant"
      - "tell me"
      - "explain"
      - "I want"
    - source_model: MaziyarPanahi/Calme-7B-Instruct-v0.9
      positive_prompts:
      - "chat"
      - "assistant"
      - "tell me"
      - "explain"
      - "I want"
    - source_model: meta-math/MetaMath-Mistral-7B
      positive_prompts:
      - "reason"
      - "math"
      - "mathematics"
      - "solve"
      - "count"

💻 Usage

!pip install -qU transformers bitsandbytes accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "NotAiLOL/Eclipse-mistral-4x7b"

tokenizer = AutoTokenizer.from_pretrained(model)
pipeline = transformers.pipeline(
    "text-generation",
    model=model,
    model_kwargs={"torch_dtype": torch.bfloat16, "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"])