Due to the use of excessive clones, performance is currently very poor for the library, at ~8x slower than PyTorch. Clones should be minimized to improve performance. In addition, matmul() should be expedited by using iterators.
Due to the use of excessive clones, performance is currently very poor for the library, at ~8x slower than PyTorch. Clones should be minimized to improve performance. In addition, `matmul()` should be expedited by using iterators.