Currently, quantized RNNs with batch and thread size = 1 is what Nuphar is tested on. But we need to provide the user the ability to compare with other configurations using the benchmark.py script.
This should be incorporated into the tutorial with details around the batch and thread sizes configurations to allow the user to compare different configs.
Currently, quantized RNNs with batch and thread size = 1 is what Nuphar is tested on. But we need to provide the user the ability to compare with other configurations using the benchmark.py script.
This should be incorporated into the tutorial with details around the batch and thread sizes configurations to allow the user to compare different configs.