merge_method: slerp
dtype: bfloat16
base_model: powermove72/LLama-6B-amt-v0.5-FineTome  # Use as base for tokenizer/alignment
models:
  - model: powermove72/LLama-6B-amt-v0.5-FineTome
    parameters:
      weight: 0.5
  - model: powermove72/LLama-6B-amt-v0.5-Alpagasus
    parameters:
      weight: 0.5
parameters:
  t:
    - filter: self_attn
      value: [0, 0.2, 0.4, 0.6, 0.8, 0.6, 0.4, 1]
    - filter: mlp
      value: [1, 0.8, 0.6, 0.4, 0.2, 0.4, 0.6, 0]
    - value: 0.5

💻 Usage

python

!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "powermove72/LLama-6b-amt-v0.5-FTA"
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"])