metadata

language:
  - ja
tags:
  - merge
  - mergekit
  - lazymergekit
  - SakanaAI/EvoLLM-JP-A-v1-7B
  - stabilityai/japanese-stablelm-base-gamma-7b
base_model:
  - SakanaAI/EvoLLM-JP-A-v1-7B
  - stabilityai/japanese-stablelm-base-gamma-7b

Hinoki-Sak-Sta-slerp-7B

Hinoki-Sak-Sta-slerp-7B is a merge of the following models using the LazyMergekit of Maxime Labonne powered by MergeKit of Arcee AI:

SakanaAI/EvoLLM-JP-A-v1-7B (Base model)
stabilityai/japanese-stablelm-base-gamma-7b

🧩 Configuration

slices:
  - sources:
      - model: SakanaAI/EvoLLM-JP-A-v1-7B
        layer_range: [0, 32]
      - model: stabilityai/japanese-stablelm-base-gamma-7b
        layer_range: [0, 32]
merge_method: slerp
base_model: SakanaAI/EvoLLM-JP-A-v1-7B
parameters:
  t:
    - filter: self_attn
      value: [0, 0.5, 0.3, 0.7, 1]
    - filter: mlp
      value: [1, 0.5, 0.7, 0.3, 0]
    - value: 0.5
dtype: bfloat16

💻 Usage

from transformers import AutoModelForCausalLM, AutoTokenizer

model_path = "AkimfromParis/Hinoki-Sak-Sta-slerp-7B"
tokenizer = AutoTokenizer.from_pretrained(model_path)
model = AutoModelForCausalLM.from_pretrained(model_path, torch_dtype="auto", device_map="auto")
model.eval()

requests = [
    "大谷翔平選手について教えてください",
]

system_message = "A chat between a curious user and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the user's questions. USER: {user_input} ASSISTANT:"

for req in requests:
    input_req = system_message.format(user_input=req)
    input_ids = tokenizer.encode(input_req, return_tensors="pt").to(device=model.device)
    tokens = model.generate(
        input_ids,
        max_new_tokens=1024,
        do_sample=True,
        pad_token_id=tokenizer.eos_token_id,
    )
    out = tokenizer.decode(tokens[0][len(input_ids[0]):], skip_special_tokens=True)
    print("USER:\n" + req)
    print("ASSISTANT:\n" + out)
    print()

Citation

@article{goddard2024arcee,
  title={Arcee's MergeKit: A Toolkit for Merging Large Language Models},
  author={Goddard, Charles and Siriwardhana, Shamane and Ehghaghi, Malikeh and Meyers, Luke and Karpukhin, Vlad and Benedict, Brian and McQuade, Mark and Solawetz, Jacob},
  journal={arXiv preprint arXiv:2403.13257},
  year={2024}
}

arxiv.org/abs/2403.13257