尝试基本 tensorflow 代码时出现运行时错误

Question

Update:

Whether it is a word, sentence or phrase, The Universal Sentence Encoder will always return vector size of 512. I will like to know why 512 and not something else.

---

The following question was resolved by the answer provided.

I tried the example provided on tensorflow home page:

https://tfhub.dev/google/universal-sentence-encoder/2

I got runtime error like this:

RuntimeError: Exporting/importing meta graphs is not supported when eager execution is enabled. No graph exists when eager execution is enabled.

The code that I tried is:

import tensorflow.compat.v1 as tf
import tensorflow_hub as hub

config = tf.ConfigProto()
session = tf.Session(config=config)

embed = hub.Module("https://tfhub.dev/google/universal-sentence-encoder/2")
embeddings = embed(
    [
        "The quick brown fox jumps over the lazy dog.",
        "I am a sentence for which I would like to get its embedding",
    ]
)

print(session.run(embeddings))

How to run this code correctly?

Answer 1

It's a matter of the tensorflow version you are using.

In Tensorflow 2.0 you should use hub.load() or hub.KerasLayer() .

Answer 2

Based on a discussion from github: https://github.com/tensorflow/hub/issues/350

The below solution worked for me:

import tensorflow.compat.v1 as tf
tf.disable_eager_execution()

The above code disables the eager execution.