search

microsoft / gated-graph-neural-network-samples

Sample Code for Gated Graph Neural Networks
MIT License
1.03k stars 264 forks source link

eval code for sparse impl #23

Closed sahilsuneja1 closed 5 years ago

sahilsuneja1 commented 5 years ago

Added evaluation code for sparse implementation, along the lines of the eval code for the dense representation case.

python36 chem_tensorflow_sparse.py --restore <model_id>_model_best.pickle --evaluate gives output similar to:

    y_actual
        [[-0.8703278393436317]]
        [[-0.6274553605532948]]
        [[1.672936001814292]]
        [[-0.4367530669782174]]
        [[0.40862621249361225]]
        [[-1.0625337848735348]]
        [[-0.21440870617984975]]
        [[-0.6763567371711069]]
        [[0.07278881719187316]]
        [[1.465301279704437]]
    y_predicted 
        [-0.7689475  -0.6265423   1.3982722  -0.5135421   0.30170625 -0.80062336
         -0.53425217 -0.6803084   0.17987642  1.4061637 ]

Signed-off-by: Sahil Suneja sahilsuneja@gmail.com

msftclas commented 5 years ago

CLA assistant check
All CLA requirements met.

sahilsuneja1 commented 5 years ago

Does this look correct?