Views
No views yet
1import torch
2from transformers import AutoTokenizer, AutoModelForCausalLM
3
4# モデルとトークナイザーを読み込み
5model = AutoModelForCausalLM.from_pretrained(
6 "eyepyon/rc3llama-3.2-1b-finetuned",
7 torch_dtype=torch.float16,
8 device_map="auto"
9)
10tokenizer = AutoTokenizer.from_pretrained("eyepyon/rc3llama-3.2-1b-finetuned")
11
12# 推論の実行
13def generate_response(prompt):
14 inputs = tokenizer(prompt, return_tensors="pt")
15 with torch.no_grad():
16 outputs = model.generate(
17 **inputs,
18 max_new_tokens=512,
19 temperature=0.7,
20 do_sample=True,
21 pad_token_id=tokenizer.eos_token_id
22 )
23 response = tokenizer.decode(outputs[0], skip_special_tokens=True)
24 return response[len(prompt):]
25
26# 使用例
27prompt = "Human: こんにちは！\n\nAssistant: "
28response = generate_response(prompt)
29print(response)1# モデルをダウンロード
2ollama pull eyepyon/rc3llama-3.2-1b-finetuned
3
4# チャット開始
5ollama run eyepyon/rc3llama-3.2-1b-finetuned1@misc{rc3llama_3.2_1b_finetuned,
2 title={rc3llama-3.2-1b-finetuned},
3 author={Your Name},
4 year={2025},
5 publisher={Hugging Face},
6 url={https://huggingface.co/eyepyon/rc3llama-3.2-1b-finetuned}
7}