在huggingface中如何使用 google/flan-t5-large

hehui0921

已于 2024-02-12 22:24:22 修改

阅读量381

点赞数 1

分类专栏： huggingface 文章标签： java 服务器 javascript

于 2024-02-12 12:26:36 首次发布

本文链接：https://blog.csdn.net/oHeHui1/article/details/136100477

版权

huggingface 专栏收录该内容

33 篇文章 0 订阅

订阅专栏

参考： https://huggingface.co/google/flan-t5-large
在这里插入图片描述

1：

from huggingface_hub.hf_api import HfFolder
HfFolder.save_token('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')
from transformers import pipeline
from transformers import T5Tokenizer, T5ForConditionalGeneration

tokenizer = T5Tokenizer.from_pretrained("google/flan-t5-large")
model = T5ForConditionalGeneration.from_pretrained("google/flan-t5-large")
input_text = """
Q: Can Geoffrey Hinton have aconversation with George Washington?
Give the rationale before answering
"""
input_ids = tokenizer(input_text, return_tensors="pt").input_ids
outputs = model.generate(input_ids,max_new_tokens=2000)
print(tokenizer.decode(outputs[0]))

2：

from huggingface_hub.hf_api import HfFolder
HfFolder.save_token('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')
from transformers import pipeline
from transformers import T5Tokenizer, T5ForConditionalGeneration

tokenizer = T5Tokenizer.from_pretrained("google/flan-t5-large")
model = T5ForConditionalGeneration.from_pretrained("google/flan-t5-large")

input_text = "answer the result: 1+ 2 ?"
input_ids = tokenizer(input_text, return_tensors="pt").input_ids
outputs = model.generate(input_ids,max_new_tokens=2000)
print(tokenizer.decode(outputs[0]))