TaHiTaHiTa/llm-jp-3-13b-it2

Uploaded model

• Developed by: TaHiTaHiTa
• 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.

llm-jp-3-13b-it2

llm-jp-3-13b-it2 is a large language model fine-tuned specifically for Japanese instruction-response tasks. This model is based on llm-jp/llm-jp-3-13b, developed by the Research and Development Center for National Institute of Informatics. (huggingface.co)

Model Details

• Model Type: Transformer-based language model
• Number of Parameters: Approximately 13 billion
• Tokenizer Used: Hugging Face tokenizer with a Unigram byte fallback model

Training Data

This model was pre-trained using a combination of the following datasets:

• Japanese: Wikipedia, Common Crawl, WARP/PDF, WARP/HTML, Kaken, etc.
• English: Wikipedia, Dolma/CC-head, Dolma/C4, Dolma/Reddit, Dolma/PeS2o, Dolma/Gutenberg, Dolma/Wiki, etc.
• Code: The Stack
• Chinese: Wikipedia
• Korean: Wikipedia

For instruction fine-tuning, the following datasets were used:

• Japanese: ichikara-instruction-004-002, answer-carefully-002, ichikara-instruction-format, AutoMultiTurnByCalm3-22B, ramdom-to-fixed-multiturn-Calm3, wizardlm8x22b-logical-math-coding-sft_additional-ja, Synthetic-JP-EN-Coding-Dataset-567k, etc.
• English: FLAN, etc.

Usage

The following code snippet demonstrates how to load the model and perform text generation:

必要なライブラリをインストール

%%capture !pip install unsloth !pip uninstall unsloth -y && pip install --upgrade --no-cache-dir "unsloth[colab-new] @ git+https://github.com/unslothai/unsloth.git" !pip install -U torch !pip install -U peft

必要なライブラリを読み込み

from unsloth import FastLanguageModel from peft import PeftModel import torch import json from tqdm import tqdm import re

ベースとなるモデルと学習したLoRAのアダプタ（Hugging FaceのIDを指定）。

model_id = "llm-jp/llm-jp-3-13b" adapter_id = ""

Hugging Face Token を指定。

HF_TOKEN = "" #@param {type:"string"}

unslothのFastLanguageModelで元のモデルをロード。

dtype = None # Noneにしておけば自動で設定 load_in_4bit = True # 今回は13Bモデルを扱うためTrue

model, tokenizer = FastLanguageModel.from_pretrained( model_name=model_id, dtype=dtype, load_in_4bit=load_in_4bit, trust_remote_code=True, )

元のモデルにLoRAのアダプタを統合。

model = PeftModel.from_pretrained(model, adapter_id, token = HF_TOKEN)

タスクとなるデータの読み込み。

事前にデータをアップロードしてください。

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 = ""

モデルを用いてタスクの推論。

推論するためにモデルのモードを変更

FastLanguageModel.for_inference(model)

results = [] for dt in tqdm(datasets): input = dt["input"]

prompt = f"""### 指示\n{input}\n### 回答\n"""

inputs = tokenizer([prompt], return_tensors = "pt").to(model.device)

outputs = model.generate(**inputs, max_new_tokens = 512, use_cache = True, do_sample=False, repetition_penalty=1.2) prediction = tokenizer.decode(outputs[0], skip_special_tokens=True).split('\n### 回答')[-1]

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

結果をjsonlで保存。

ここではadapter_idを元にファイル名を決定しているが、ファイル名は任意で問題なし。

json_file_id = re.sub(".*/", "", adapter_id) with open(f"/content/{json_file_id}_output.jsonl", 'w', encoding='utf-8') as f: for result in results: json.dump(result, f, ensure_ascii=False) f.write('\n')