Uploaded model

  • Developed by: WatariNAKANO
  • 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.

推論用コード

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

# Install Flash Attention 2 for softcapping support
import torch
if torch.cuda.get_device_capability()[0] >= 8:
    !pip install --no-deps packaging ninja einops "flash-attn>=2.6.3"

# Hugging Face Token を指定
HF_TOKEN = "your-token" #@param {type:"string"}

# llm-jp/llm-jp-3-13bを4bit量子化のqLoRA設定でロード。

from unsloth import FastLanguageModel
import torch
max_seq_length = 1024 # unslothではRoPEをサポートしているのでコンテキスト長は自由に設定可能
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_id = "llm-jp/llm-jp-3-13b"
new_model_id = "llm-jp-3-13b-it" #Fine-Tuningしたモデルにつけたい名前、it: Instruction Tuning
# FastLanguageModel インスタンスを作成
model, tokenizer = FastLanguageModel.from_pretrained(
    model_name=model_id,
    dtype=dtype,
    load_in_4bit=load_in_4bit,
    trust_remote_code=True,
)

# SFT用のモデルを用意
model = FastLanguageModel.get_peft_model(
    model,
    r = 32,
    target_modules = ["q_proj", "k_proj", "v_proj", "o_proj",
                      "gate_proj", "up_proj", "down_proj",],
    lora_alpha = 32,
    lora_dropout = 0.05,
    bias = "none",
    use_gradient_checkpointing = "unsloth",
    random_state = 3407,
    use_rslora = False,
    loftq_config = None,
    max_seq_length = max_seq_length,
)

# 学習に用いるデータセットの指定
# CC-BY-NC-SAですのでモデルはライセンスを継承する前提でお使いください。
# https://liat-aip.sakura.ne.jp/wp/llmのための日本語インストラクションデータ作成/llmのための日本語インストラクションデータ-公開/
# 関根聡, 安藤まや, 後藤美知子, 鈴木久美, 河原大輔, 井之上直也, 乾健太郎. ichikara-instruction: LLMのための日本語インストラクションデータの構築. 言語処理学会第30回年次大会(2024)

from datasets import load_dataset

dataset = load_dataset("json", data_files="/content/ichikara-instruction-003-001-1.json") 

# 学習時のプロンプトフォーマットの定義
prompt = """### 指示
{}
### 回答
{}"""


"""
formatting_prompts_func: 各データをプロンプトに合わせた形式に合わせる
"""
EOS_TOKEN = tokenizer.eos_token # トークナイザーのEOSトークン(文末トークン)
def formatting_prompts_func(examples):
    input = examples["text"] # 入力データ
    output = examples["output"] # 出力データ
    text = prompt.format(input, output) + EOS_TOKEN # プロンプトの作成
    return { "formatted_text" : text, } # 新しいフィールド "formatted_text" を返す
pass

# # 各データにフォーマットを適用
dataset = dataset.map(
    formatting_prompts_func,
    num_proc= 4, # 並列処理数を指定
)

dataset


# training_arguments: 学習の設定
from trl import SFTTrainer
from transformers import TrainingArguments
from unsloth import is_bfloat16_supported

trainer = SFTTrainer(
    model = model,
    tokenizer = tokenizer,
    train_dataset=dataset["train"],
    max_seq_length = max_seq_length,
    dataset_text_field="formatted_text",
    packing = False,
    args = TrainingArguments(
        per_device_train_batch_size = 2,
        gradient_accumulation_steps = 4,
        num_train_epochs = 1,
        logging_steps = 10,
        warmup_steps = 10,
        save_steps=100,
        save_total_limit=2,
        max_steps=-1,
        learning_rate = 2e-4,
        fp16 = not is_bfloat16_supported(),
        bf16 = is_bfloat16_supported(),
        group_by_length=True,
        seed = 3407,
        output_dir = "outputs",
        report_to = "none",
    ),
)


#@title 学習実行
trainer_stats = trainer.train()


# データセットの読み込み。
import json
datasets = []
with open("/content//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 = ""


# 学習したモデルを用いてタスクを実行
from tqdm import tqdm

# 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 = 1024, 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})


# jsonlで保存
with open(f"{new_model_id}_output.jsonl", 'w', encoding='utf-8') as f:
    for result in results:
        json.dump(result, f, ensure_ascii=False)
        f.write('\n')


# LoRAアダプタだけ保存
new_model_id = "WatariNAKANO/llm-jp-3-13b-it-2" #Fine-Tuningしたモデルにつけたい名前
model.push_to_hub_merged(
    new_model_id+"_lora",
    tokenizer=tokenizer,
    save_method="lora",
    token=HF_TOKEN,
    private=True
)
Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model is not currently available via any of the supported Inference Providers.
The model cannot be deployed to the HF Inference API: The model has no pipeline_tag.

Model tree for WatariNAKANO/llm-jp-3-13b-it-2_lora

Finetuned
(1119)
this model