Hugging Face H5 加载模型错误：在配置文件中找不到模型

Posted 2023-03-29

【中文标题】Hugging Face H5 加载模型错误：在配置文件中找不到模型

【英文标题】：Hugging Face H5 load model error : No model found in config file

【发布时间】：2021-11-19 09:55:22

【问题描述】：

我正在尝试从 Hugging Face 加载模型，并从这里下载了 h5 模型：https://huggingface.co/distilbert-base-uncased-finetuned-sst-2-english/tree/main

from flask import Flask, jsonify, request  # import objects from the Flask model
from keras.models import load_model
from transformers import AutoTokenizer, AutoModelForSequenceClassification,TextClassificationPipeline

model = load_model('./tf_model.h5') # trying to load model here

然后出现错误：

File "C:\D\Learning\Flask\flask-pp-rest\main.py", line 11, in <module>
    model = load_model('./tf_model.h5') File "C:\Users\ndrez\AppData\Local\Programs\Python\Python39\lib\site-packages\keras\saving\save.py",
line 200, in load_model
    return hdf5_format.load_model_from_hdf5(filepath, custom_objects, File
"C:\Users\ndrez\AppData\Local\Programs\Python\Python39\lib\site-packages\keras\saving\hdf5_format.py",
line 176, in load_model_from_hdf5
    raise ValueError('No model found in config file.') ValueError: **No model found in config file.**

有谁知道如何解决这个问题？如果你知道请帮帮我。我会监控这个问题并尝试实施您的解决方案的答案。

【问题讨论】：

你从哪里得到这个 h5 文件或者你是怎么做的？

【参考方案1】：

您可以使用TFAutoModelForSequenceClassification 类加载distilbert-base-uncased-finetuned-sst-2-english 的张量流版本：

from transformers import AutoTokenizer, TFAutoModelForSequenceClassification
tokenizer = AutoTokenizer.from_pretrained("distilbert-base-uncased-finetuned-sst-2-english")
model = TFAutoModelForSequenceClassification.from_pretrained("distilbert-base-uncased-finetuned-sst-2-english")

【参考方案2】：

要加载您指定的模型，代码如下：

from transformers import AutoTokenizer, AutoModelForSequenceClassification
tokenizer = AutoTokenizer.from_pretrained("distilbert-base-uncased-finetuned-sst-2-english")
model = AutoModelForSequenceClassification.from_pretrained("distilbert-base-uncased-finetuned-sst-2-english")

【讨论】：

我试过这个，它需要pytorch，我只想使用tensorflow