Exploring Parallelism in FPGA-Based Accelerators for Machine Learning Applications