tf.function makes invalid assumptions about arguments that are Mapping instances. In general, there are no requirements for Mapping instances to have constructors that accept [(key, value)] initializers, as assumed here.

This leads to cryptic exceptions when used with perfectly valid Mapping s

https://github.com/keras-team/keras/blob/master/keras/engine/training.py#L1071-L1077
There are two validation_steps.

https://github.com/aymericdamien/TensorFlow-Examples/blob/master/tensorflow_v2/notebooks/3_NeuralNetworks/convolutional_network_raw.ipynb

trainable_variables = weights.values() + biases.values() doesn't work.
Also if I write trainable_variables = list(weights.values()) + list(biases.values()), I have to turn on tf.enable_eager_execution(), but the training result is wrong, accuracy is ar

• 原文链接：https://strapi.io/blog/build-a-blog-using-nuxt-strapi-and-apollo
• 简要介绍：前段时间有人推荐了如何使用nest、vue、mongodb，搭建博客。这篇文章是使用Nuxt.js、strapi 和 Apollo 搭建博客。其中Nuxt.js 是Vue的服务端渲染框架、strapi是一个很方便且开源的内容管理平台（这里用来管理数据库里的博客帖子），Apollo 用于 graphQL ，也就是前后端的交互。可以说是上次那篇博客的另一种实现方案。
  另外...strapi现在在公测的beta版本，如果有正式版本发布，掘金翻译计划有没有兴趣翻译一下文档。。。strapi 在GitHub 上有21k 的star，还挺多的。

请完成并勾选一下三项：

• [ x

🐛 Bug

Model I am using (Bert, XLNet....): CamemBERT but this probly applies to all MLMs.

Language I am using the model on (English, Chinese....): French

The problem arise when using:

• my own modified scripts, but I suspect that https://github.com/huggingface/transformers/blob/master/examples/run_lm_finetuning.py is also impacted.

Basically,

Google has started using BERT in its search engine. I imagine it creates embeddings for the query on the search engine, and then find a kind of similarity measure with the potential candidate websites/pages, finally ranking them in search results.

I am curious how do they create embeddings for the documents (the potential candidate websites/pages) if any? Or am I interpreting it wrong?

Alexnet implementation in tensorflow has incomplete architecture where 2 convolution neural layers are missing. This issue is in reference to the python notebook mentioned below.

https://github.com/donnemartin/data-science-ipython-notebooks/blob/master/deep-learning/tensor-flow-examples/notebooks/3_neural_networks/alexnet.ipynb

Thanks for wonderful repository. I inspect model using your provided notebook, such as inspect_model.ipynb.
Now I would like to visualize output from different layers during training time?
Is there any suggestion please

Traceback (most recent call last):
File "/home/ubuntu/Real-Time-Voice-Cloning-master/toolbox/init.py", line 59, in
self.ui.browser_load_button.clicked.connect(lambda: self.load_from_browser())
File "/home/ubuntu/Real-Time-Voice-Cloning-master/toolbox/init.py", line 122, in load_from_browser
self.add_real_utterance(wav, name, speaker_name)
File "/home/ubuntu/Real

transcribe.py has odd directory-scanning behavior which isn't documented

If you point --src to a directory, you get the error:

E Path in --src not existing

Looking at the code logic, the script expects a JSON file with a .catalog file extension. This is (1) not documented, and (2) not a really useful logic. It would be much better to point the script to a dir, and scan f

I have trained a Svc classfier on 300 people dataset. Each class has 40 images.

But accuracy is very poor. Accuracy is 0.010

Looping the process of writing images into the .tfrecords-file works fine, but how do I read multiple images from a .tfrecords-file?

Is there any simple solution? would be great if added to the code.

load_csv line can be since there's columns_to_ignore param's supported

data, labels = load_csv('titanic_dataset.csv', target_column=0, columns_to_ignore=[2, 7], categorical_labels=True, n_classes=2)

and we don't need to do that in preprocess()

def preprocess(passengers):
    for i in range(len(passengers)):
        passengers[i][1] = 1. if passengers[i][1] == 'female' else 0.

I cannot effectively use this program because I am missing the instructions because the link to the Install instructions on github produce a 404 error. Could they be added as a .txt file to future versions or at least can you please fix the link?

I've ran into this issue for a couple hours and I ended up editing the dist library adding two new functions called fetchVideo and bufferToVideo that works pretty much like the fetchImage and bufferToImage functions.

I'll leave it here to help somebody else with the same issue and in case someone wants to include it on future releases.

face-api.js

...
exports.fetchVideo = fetc

"scipy.misc.imsave" was removed in scipy version 1.3
--> pip install scipy==1.2.1

https://docs.scipy.org/doc/scipy/reference/release.1.3.0.html#scipy-interpolate-changes

and functions from scipy.misc (bytescale, fromimage, imfilter, imread, imresize, imrotate, imsave, imshow, toimage) have been removed.

Judging by the logic in https://github.com/horovod/horovod/blob/38e91bee84efbb5b563a4928027a75dc3974633b/setup.py#L1369 it is clear, that before installing Horovod one needs to install the underlying framework(s) (TensorFlow, PyTorch, ...).

This is not mentioned in the installation instructions which made me think, I can install Horovod and then any framework I like (or switch between them) and

We are moving in the direction of having multiple config files like
https://github.com/kubeflow/kubeflow/blob/master/bootstrap/config/kfctl_gcp_iap.0.6.yaml
https://github.com/kubeflow/kubeflow/blob/master/bootstrap/config/kfctl_existing_arrikto.0.6.yaml

that specify different configurations of Kubeflow

We should add unittests that ensure consistency.

i.e. we should add a test that veri

inputs = tf.placeholder(name='foo', shape=[None, 4, 4, 256],dtype=tf.float32)
snt.Conv2DTranspose(output_channels=256, kernel_shape=4, stride=[1, 2], padding="VALID")(inputs)

--> yields shape (?,7,11,256)

tf.layers.conv2d_transpose(inputs, filters=256, kernel_size=4, strides=[1, 2], padding='valid')`

--> yields shape (?,7,10,256)

Why?

Tutorial 19 worked fine with TensorFlow v. 1.4.0 and Keras v. 2.0.8-tf. But I have just upgraded to TensorFlow 1.9 and Keras 2.1.6-tf and now I get an error when trying to load the model with the best hyper-parameters. I think it is a bug in Keras because many other people have experienced variations of the same error-message, see e.g. tensorflow/tensorflow#20073

We wi

Support for storing large tensor values in external files was introduced in #678, but AFAICT is undocumented.

This is a pretty important feature, functionally, but it's also important for end users who may not realise that they need to move around more than just the *.onnx file.

I would suggest it should be documented in IR.md, and perhaps there are other locations from which it could be s