Daschund

Daschund is a merge of the following models using LazyMergekit:

• teknium/OpenHermes-2.5-Mistral-7B
• openchat/openchat-3.5-0106
• andrijdavid/macaroni-7b
• mistralai/Mistral-7B-Instruct-v0.2
• Weyaxi/OpenHermes-2.5-neural-chat-v3-3-Slerp
• Intel/neural-chat-7b-v3-1
• mlabonne/Beagle14-7B
• mlabonne/NeuralBeagle14-7B

🧩 Configuration

slices:
  - sources:
    - model: teknium/OpenHermes-2.5-Mistral-7B 
      layer_range: [0, 4]
  - sources:
    - model: openchat/openchat-3.5-0106
      layer_range: [4, 8]
  - sources:
    - model: andrijdavid/macaroni-7b
      layer_range: [8, 12]
  - sources:
    - model: mistralai/Mistral-7B-Instruct-v0.2 
      layer_range: [12, 16]
  - sources:
    - model: Weyaxi/OpenHermes-2.5-neural-chat-v3-3-Slerp
      layer_range: [16, 20]
  - sources:
    - model: Intel/neural-chat-7b-v3-1
      layer_range: [20, 24]
  - sources:
    - model: mlabonne/Beagle14-7B
      layer_range: [24, 28]
  - sources:
    - model: mlabonne/NeuralBeagle14-7B
      layer_range: [28, 32]
merge_method: passthrough
dtype: bfloat16

💻 Usage

!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "N8Programs/Daschund"
messages = [{"role": "user", "content": "What is a large language model?"}]

tokenizer = AutoTokenizer.from_pretrained(model)
prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
pipeline = transformers.pipeline(
    "text-generation",
    model=model,
    torch_dtype=torch.float16,
    device_map="auto",
)

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"])