Historique de IBM - Neural Network Modeler

Mis à jour 3 avr. 2018 21:29:41

Date de création 25 mar. 2018 22:14:35

• 
  PALYGAP le 3 avr. 2018 21:29:41Signaler

  Version of the software

  1.1

• 
  PALYGAP le 3 avr. 2018 21:26:56Signaler

  AutoML

  no

  Browser

  no

  Python shell

  no

  - Desktop

  no

• 
  PALYGAP le 3 avr. 2018 21:22:04Signaler

  Custom cost function

  no

  Load Keras Models + Weights

  no

  Load Caffe Models + Weights

  yes

  Model checker

  yes

  Experiments Mangement

  no

• 
  PALYGAP le 3 avr. 2018 21:20:52Signaler

  - LSTM

  yes

  - MLP / Auto-encoders

  yes

  - RNN

  yes

  Use Keras API

  no

  Save Keras HDFS file (.h5)

  no

  Save Keras model .yaml file

  no

  Save Keras model file

  no

  Load Keras model .yaml file

  no

  Load Keras HDFS file (.h5)

  no

  Code generation

  yes

  - Theano

  yes

  - Tensorflow/Keras

  yes

  - Caffe (v1)

  yes

• 
  PALYGAP le 30 mar. 2018 05:31:33Signaler

  DL model training

  no

  Inference available

  no

  Graphical editor provided

  yes

• 
  PALYGAP le 26 mar. 2018 21:07:46Signaler

  Nom

  IBM - Neural Network Modeler

• 
  PALYGAP le 26 mar. 2018 21:05:35Signaler

  Nom

  IBM - Neural Network modeler

  Version of the software

  Version integrated in Watson Studio

• 
  PALYGAP le 25 mar. 2018 22:39:43Signaler

  Web Site

  https://darviz.mybluemix.net/#/dashboard

• 
  PALYGAP le 25 mar. 2018 22:14:36Signaler

  Mode

  70

  Nom

  Deep Learning IDE

  Visibilité

  public

  Editeur

  palygap-1q7klct2

  Version of the software

  Current

  Underlaying Framework

  Theano Tensorflow/Keras Caffe (v1)

  (2 enhanced versions have been release recently)"

  Important

  GPUs

  Depends on the target Framework

  DL MODEL EDITOR

  Yes

  - CNN

  yes

  - LSTM

  Yes

  - MLP / Auto-encoders

  Yes

  - RNN

  Yes

  Pretained Networks

  no

  Use Keras API

  No

  Save Keras HDFS file (.h5)

  No

  Save Keras model .yaml file

  No

  Save Keras model file

  No

  Load Keras model .yaml file

  No

  Load Keras HDFS file (.h5)

  No

  Code generation

  Yes

  - Theano

  Yes

  - Tensorflow/Keras

  Yes

  - Caffe (v1)

  Yes

  Custom cost function

  No

  Load Keras Models + Weights

  No

  Load Caffe Models + Weights

  Yes

  Model checker

  Yes

  Datasets Management

  no

  Experiments Mangement

  No

  - Exp comparison

  N/A

  - Prev. Exp loader

  N/A

  TRAINING

  no

  Tune learning parameter

  N/A

  Live performance monitoring

  N/A

  Results browser

  N/A

  INFERENCE

  no

  View Valid/test set inference

  N/A

  Confusion matrix

  N/A

  Visualize intermediate layer activation

  N/A

  User test int

  N/A

  DEPLOIEMENT

  no

  - Cloud

  "Yes

  - Cloud Trial period

  N/A (no usage limitation)

  - Desktop

  No

  AutoML

  No

  Browser

  No

  Environment management

  no

  Python shell

  No

  Jupyter notebook

  no

  Tutorials

  yes

  Development Activity

  Unkown