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

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

HF_TOKEN = "YourHaggingFaceToken"

base_model_id = "google/gemma-2-9b"
adapter_id = "Ryota2865/gemma-2-9b-finetune5" 

# QLoRA config
bnb_config = BitsAndBytesConfig(
    load_in_4bit=True,
    bnb_4bit_quant_type="nf4",
    bnb_4bit_compute_dtype=torch.bfloat16,
)

# Load model
model = AutoModelForCausalLM.from_pretrained(
    base_model_id,
    quantization_config=bnb_config,
    device_map="auto",
    token = HF_TOKEN
)

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

# 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)

# データセットの読み込み。
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
results = []
for data in tqdm(datasets):

  input = data["input"]
  prompt = f"""### 指示
  {input}
  ### 回答
  """

  input_ids = tokenizer(prompt, return_tensors="pt").to(model.device)
  outputs = model.generate(**input_ids, max_new_tokens=512, do_sample=False, repetition_penalty=1.2,)
  output = tokenizer.decode(outputs[0][input_ids.input_ids.size(1):], skip_special_tokens=True)

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

# jsonlで保存
import re
jsonl_id = re.sub(".*/", "", adapter_id)
with open(f"./{jsonl_id}-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')