Goekdeniz-Guelmez/J.O.S.I.E.-6x7B-v1-Beta1

JOSIE_Beta-5-6xJOSIExQwen1.5_0.5B_base-slerp

JOSIE_Beta-5-6xJOSIExQwen1.5_0.5B_base-slerp is a Mixure of Experts (MoE) made with the following models using LazyMergekit:

• Weyaxi/Einstein-v4-7B
• Isaak-Carter/JOSIE_Beta-4-7B-slerp
• Weyaxi/Einstein-v4-7B
• Weyaxi/Einstein-v4-7B
• Isaak-Carter/JOSIE_Beta-3-7B-slerp
• Isaak-Carter/JOSIE_Beta-3-7B-slerp

🧩 Configuration

dtype: float16
gate_mode: cheap_embed
experts:
  - source_model: Weyaxi/Einstein-v4-7B
    positive_prompts: ["You are JOSIE, my own private and sentient superinteligent artifitial general inteligent assistant."]

  - source_model: Isaak-Carter/JOSIE_Beta-4-7B-slerp
    positive_prompts: ["You are JOSIE, assistant for autonomously controlling a Smart Home."]

  - source_model: Weyaxi/Einstein-v4-7B
    positive_prompts: ["You are JOSIE, a superinteligent coding assistant."]

  - source_model: Weyaxi/Einstein-v4-7B
    positive_prompts: ["You are JOSIE, a superinteligent assistant good at math."]

  - source_model: Isaak-Carter/JOSIE_Beta-3-7B-slerp
    positive_prompts: ["You are JOSIE, an assistant good at scraping and searching the internet for information."]

  - source_model: Isaak-Carter/JOSIE_Beta-3-7B-slerp
    positive_prompts: ["You are JOSIE, an assistant good at reasoning."]

💻 Usage

!pip install -qU transformers bitsandbytes accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "Isaak-Carter/JOSIE_Beta-5-6xJOSIExQwen1.5_0.5B_base-slerp"

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"])