fmcquillan99 commented on issue #392: DL: Improve performance for predict URL: https://github.com/apache/madlib/pull/392#issuecomment-494189650 (1) `model_size` is not computed correctly ``` DROP TABLE IF EXISTS model1, model1_summary; SELECT madlib.madlib_keras_fit('mnist_train_packed', -- source_table 'model1', -- model 'model_arch_library', -- model_arch_table 1, -- model_arch_id $$ loss='categorical_crossentropy', optimizer='adam', metrics=['acc'] $$, -- compile_params NULL, -- fit_params 5, -- num_iterations 0, -- gpus per host NULL, -- validation_table NULL, -- metrics compute frequency 'Frank', -- name 'A test model' -- description ); madlib=# select model_size from model1_summary; model_size ------------ 43 (1 row) ``` ``` model_simple = Sequential() model_simple.add(Dense(100, activation='relu', input_shape=(784,))) model_simple.add(Dense(10, activation='softmax')) model_simple.summary() Layer (type) Output Shape Param # ================================================================= dense_1 (Dense) (None, 100) 78500 _________________________________________________________________ dense_2 (Dense) (None, 10) 1010 ================================================================= Total params: 79,510 Trainable params: 79,510 Non-trainable params: 0 _________________________________________________________________ ``` Should be the size in bytes of the bytea storage of the model weights.
---------------------------------------------------------------- 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
