khannaekta commented on a change in pull request #463: DL: Add asymmetric
cluster check for predict
URL: https://github.com/apache/madlib/pull/463#discussion_r356265527
##########
File path: src/ports/postgres/modules/deep_learning/madlib_keras_predict.py_in
##########
@@ -244,14 +254,14 @@ class PredictBYOM(BasePredict):
def internal_keras_predict(independent_var, model_architecture, model_weights,
is_response, normalizing_const, current_seg_id,
seg_ids,
- images_per_seg, gpus_per_host, segments_per_host,
+ images_per_seg, use_gpus, gpus_per_host,
segments_per_host,
**kwargs):
SD = kwargs['SD']
model_key = 'segment_model_predict'
row_count_key = 'row_count'
try:
- device_name = get_device_name_and_set_cuda_env_predict(gpus_per_host,
- current_seg_id)
+ device_name = get_device_name_and_set_cuda_env(
+ use_gpus, gpus_per_host, current_seg_id)
Review comment:
As we are already validating the gpu configuration prior to calling this
function(for fit/evaluate/predict), the function
`get_device_name_and_set_cuda_env()` should only care about setting device name
(GPU or CPU based on the value of `gpus_per_host`) and should not validate GPU
configuration(based on `use_gpus`).
I think, it's ok to remove `use_gpus` from
`get_device_name_and_set_cuda_env()`, unless I am missing some scenario.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
With regards,
Apache Git Services
