search

RedaOps / ann-visualizer

A python library for visualizing Artificial Neural Networks (ANN)
MIT License
1.24k stars 216 forks source link

Support for tf.keras #20

Open mmesensol opened 5 years ago

mmesensol commented 5 years ago

Is it possible to add support for models created with tf.keras instead of keras?

Otherwise, do you know of a similar tool with such support?

Thanks

Hemantr05 commented 5 years ago

Yes, It is possible to use tf.keras instead of keras if you're tensorflow to develop a neural network instead of keras.

Just like the above, the layers and models can also be added , tf.keras.models , tf.keras.layers and so on.

On Thu, Nov 15, 2018 at 5:17 AM mmesensol notifications@github.com wrote:

Is it possible to add support for models created with tf.keras instead of keras?

Otherwise, do you know of a similar tool with such support?

Thanks

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/Prodicode/ann-visualizer/issues/20, or mute the thread https://github.com/notifications/unsubscribe-auth/AccUkaxK9wOGsnCmYZK4jn0x8J4M5GU9ks5uvKuBgaJpZM4Yezdk .

mmesensol commented 5 years ago

Thanks. I was able to get it to work by hacking around the ann_visualizer/visualize.py file and replacing keras. by tf.keras. in different ways. It wasn't very straightforward, though.

Still, I got stuck because ANN Visualizer like that my model's first layer is a Flatten. Is there a way to make it work, or is such a limitation by design?

Thanks

Hemantr05 commented 5 years ago

There is an another alternative of visualizing you neural network architecture,

https://keras.io/visualization/ The above link is a part of the keras documentation to visualize the neural network. I'm pretty sure, it can be access using tf.keras as well. You can give it a try

On Fri, Nov 16, 2018 at 3:22 AM mmesensol notifications@github.com wrote:

Thanks. I was able to get it to work by hacking around the ann_visualizer/visualize.py file and replacing keras. by tf.keras. in different ways. It wasn't very straightforward, though.

Still, I got stuck because ANN Visualizer like that my model's first layer is a Flatten. Is there a way to make it work, or is such a limitation by design?

Thanks

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/Prodicode/ann-visualizer/issues/20#issuecomment-439203556, or mute the thread https://github.com/notifications/unsubscribe-auth/AccUkdhIfrHwYtiBMADnjJFXf5yB111Vks5uveISgaJpZM4Yezdk .