name: Jebadiah/Aria-tree-coder-7b
models:
  - model: Jebadiah/Aria-coder-7b
    parameters:
      density: 1.0
      weight: 1.0

merge_method: passthrough

layers:
  - source:
      layer_range: [0, 15]  # layers before the middle
    target:
      layer_range: [0, 15]

  - source:
      layer_range: [16, 16]  # the middle layer to duplicate
    target:
      layer_range: [16, 17]  # copy to positions 16 and 17

  - source:
      layer_range: [17, 30]  # layers after the middle (shifted by 1)
    target:
      layer_range: [18, 31]

dtype: float16

💻 Usage

python

!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "Jebadiah/Aria-tree-coder-7b"
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"])