Abstract: Techniques are provided for compressing deep neural networks using a structured filter pruning method that is extensible and effective. According to an embodiment, a computer-implemented method comprises determining, by a system operatively coupled to a processor, importance scores for filters of layers of a neural network model previously trained until convergence for an inferencing task on a training dataset. The method further comprises removing, by the system, a subset of the filters from one or more layers of the layers based on the importance scores associated with the subset failing to satisfy a threshold importance score value. The method further comprises converting, by the system, the neural network model into a compressed neural network model with the subset of the filters removed.

Abstract: Trace hydrogen may be removed from a dry gas by passing the dry gas at a temperature from about 0° C. to about 60° C. through at least one layer of a first hopcalite catalyst to produce product gas that is at least substantially free of hydrogen, wherein the first hopcalite catalyst has a molar ratio of copper to manganese of more than 0.55. Advantages include increase hydrogen capacity, lower feed and regeneration temperatures and lower sensitivity to carbon dioxide than equivalent processes using standard hopcalite catalyst having a Cu/Mn molar ratio from 0.45 to 0.55.

Abstract: A door alignment system includes a first alignment member secured to one of a first door or a structure connected to the first door, and a second alignment member secured to the other of the first door or the structure. The first alignment member is configured to couple to the second alignment member when the first door is in a closed position relative to the structure in order to prevent the first door from binding in relation to the structure. The first alignment member is configured to be separated from the second alignment member when the first door is in an open position relative to the structure.

Abstract: A door alignment system includes a first alignment member secured to one of a first door or a structure connected to the first door, and a second alignment member secured to the other of the first door or the structure. The first alignment member is configured to couple to the second alignment member when the first door is in a closed position relative to the structure in order to prevent the first door from binding in relation to the structure. The first alignment member is configured to be separated from the second alignment member when the first door is in an open position relative to the structure.