llm-jp-3-13b-finetune

推論の方法

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

model_id = "models/models--llm-jp--llm-jp-3-13b/snapshots/cd3823f4c1fcbb0ad2e2af46036ab1b0ca13192a" adapter_id = "llm-jp-3-13b-finetune"

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

model = AutoModelForCausalLM.from_pretrained( model_id, quantization_config=bnb_config, device_map="auto", token = HF_TOKEN )

Load tokenizer

tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True, token = HF_TOKEN)

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 = "" ※elyza-tasks-100-TV_0.jsonlは適宜入手しおいておくこと

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}) 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')