AlHfac/llm-jp-3-13b-it

How to Run this Model

基本的にhugging face modelとしてloadすればOK。

elyza-tasks-100-TV_0.jsonl を事前に同じフォルダーに置いてください。

HF_TOKENの入れ替えを忘れないでください

環境準備

!pip install -U bitsandbytes
!pip install -U transformers
!pip install -U accelerate
!pip install -U datasets

結果jsonlを作成ためのコード例

推論結果が　llm-jp-3-13b-it-outputs.jsonl　に作成される

from transformers import (
    AutoModelForCausalLM,
    AutoTokenizer,
    BitsAndBytesConfig,
)
import torch
from tqdm import tqdm
import json

HF_TOKEN = ADD YOUR OWN TOKEN
model_name = "AlHfac/llm-jp-3-13b-it"

# QLoRA config
bnb_config = BitsAndBytesConfig(
    load_in_4bit=True,
    bnb_4bit_quant_type="nf4",
    bnb_4bit_compute_dtype=torch.bfloat16,
    bnb_4bit_use_double_quant=False,
)
# Load model
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    quantization_config=bnb_config,
    device_map="auto",
    token = HF_TOKEN
)

# Load tokenizer
tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True, token = HF_TOKEN)

# Load Questions
datasets = []
with open("./elyza-tasks-100-TV_0.jsonl", "r") as f:
    item = ""
    for line in f:
      line = line.strip()
      item += line
      if item.endswith("}"):
        datasets.append(json.loads(item))
        item = ""

# Generate results using loaded model
results = []
for data in tqdm(datasets):

  input = data["input"]

  prompt = f"""### 指示
  {input}
  ### 回答：
  """

  tokenized_input = tokenizer.encode(prompt, add_special_tokens=False, return_tensors="pt").to(model.device)
  with torch.no_grad():
      outputs = model.generate(
          tokenized_input,
          max_new_tokens=100,
          do_sample=False,
          repetition_penalty=1.2
      )[0]
  output = tokenizer.decode(outputs[tokenized_input.size(1):], skip_special_tokens=True)

  results.append({"task_id": data["task_id"], "input": input, "output": output})


# Generate jsonl
import re
model_name = re.sub(".*/", "", model_name)
with open(f"./{model_name}-outputs.jsonl", 'w', encoding='utf-8') as f:
    for result in results:
        json.dump(result, f, ensure_ascii=False)  # ensure_ascii=False for handling non-ASCII characters
        f.write('\n')

The attention mask and the pad token id were not set. As a consequence, you may observe unexpected behavior. Please pass your input's attention_mask to obtain reliable results. Setting pad_token_id to eos_token_id:2 for open-end generation. ようなlogを無視してもOK

Model Training Information

• Developed by: AlHfac
• License: apache-2.0
• Finetuned from model : llm-jp/llm-jp-3-13b

This llama model was trained 2x faster with Unsloth and Huggingface's TRL library.