Views
No views yet
1from transformers import AutoModelForCausalLM, AutoTokenizer
2
3model = AutoModelForCausalLM.from_pretrained('TheBug95/llama-3.2-1B-MS-MARCO-QLoRA-v3')
4tokenizer = AutoTokenizer.from_pretrained('TheBug95/llama-3.2-1B-MS-MARCO-QLoRA-v3')1def generate_answer(prompt):
2 inputs = tokenizer(prompt, return_tensors='pt').to(device)
3 outputs = model.generate(**inputs, max_length=512, do_sample=True, temperature=0.7)
4 answer = tokenizer.decode(outputs[0], skip_special_tokens=True)
5 return answer
6
7# Ejemplo de uso
8prompt = 'Context:
9[Tu contexto aquí]
10
11Question:
12[Tu pregunta aquí]
13
14Answer:'
15respuesta = generate_answer(prompt)
16print(respuesta)1@misc{bajaj2018msmarcohumangenerated,
2 title={MS MARCO: A Human Generated MAchine Reading COmprehension Dataset},
3 author={Payal Bajaj and Daniel Campos and Nick Craswell and Li Deng and Jianfeng Gao and Xiaodong Liu and Rangan Majumder and Andrew McNamara and Bhaskar Mitra and Tri Nguyen and Mir Rosenberg and Xia Song and Alina Stoica and Saurabh Tiwary and Tong Wang},
4 year={2018},
5 eprint={1611.09268},
6 archivePrefix={arXiv},
7 primaryClass={cs.CL},
8 url={https://arxiv.org/abs/1611.09268},
9}1@misc{dettmers2023qloraefficientfinetuningquantized,
2 title={QLoRA: Efficient Finetuning of Quantized LLMs},
3 author={Tim Dettmers and Artidoro Pagnoni and Ari Holtzman and Luke Zettlemoyer},
4 year={2023},
5 eprint={2305.14314},
6 archivePrefix={arXiv},
7 primaryClass={cs.LG},
8 url={https://arxiv.org/abs/2305.14314},
9}1@misc{hu2021loralowrankadaptationlarge,
2 title={LoRA: Low-Rank Adaptation of Large Language Models},
3 author={Edward J. Hu and Yelong Shen and Phillip Wallis and Zeyuan Allen-Zhu and Yuanzhi Li and Shean Wang and Lu Wang and Weizhu Chen},
4 year={2021},
5 eprint={2106.09685},
6 archivePrefix={arXiv},
7 primaryClass={cs.CL},
8 url={https://arxiv.org/abs/2106.09685},
9}