nyu-mll/glue
Viewer • Updated • 1.49M • 442k • 523
How to use sncffcns/llm-jp-3-13b-it-20241127_lora with Transformers:
# Use a pipeline as a high-level helper
from transformers import pipeline
pipe = pipeline("text-classification", model="sncffcns/llm-jp-3-13b-it-20241127_lora") # Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("sncffcns/llm-jp-3-13b-it-20241127_lora", device_map="auto")# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("sncffcns/llm-jp-3-13b-it-20241127_lora", device_map="auto")以下の手順に従うことで、Hugging Face上のモデル(llm-jp/llm-jp-3-13b + /sncffcns/llm-jp-3-13b-it-20241127_lora)を用いて入力データ(elyza-tasks-100-TV_0.jsonl)を推論し、その結果を{adapter_id}-outputs.jsonlというファイルに出力することができる。
!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
from unsloth import FastLanguageModel
from peft import PeftModel
import torch
import json
from tqdm import tqdm
import re
model_id = "llm-jp/llm-jp-3-13b"
adapter_id = "sncffcns/llm-jp-3-13b-it-20241127_lora"
# Hugging Face Token を指定する
from google.colab import userdata
HF_TOKEN = userdata.get('HF_TOKEN_WRITE')
# 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のアダプタを統合する
model = PeftModel.from_pretrained(model, adapter_id, token = HF_TOKEN)
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を行う
# 推論するためにモデルのモードを変更する
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})
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')
Base model
llm-jp/llm-jp-3-13b
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-classification", model="sncffcns/llm-jp-3-13b-it-20241127_lora")