[TVM Discuss] [Development/RFC] Performing Relay Passes Non-Recursively

2020-03-22 Thread Marisa Kirisame via TVM Discuss
writing trampoline by hand is probably the way to go for those passes. --- [Visit Topic](https://discuss.tvm.ai/t/performing-relay-passes-non-recursively/5696/17) to respond. You are receiving this because you enabled mailing list mode. To unsubscribe from these emails, [click here](htt

[TVM Discuss] [Development/RFC] Performing Relay Passes Non-Recursively

2020-03-22 Thread Marisa Kirisame via TVM Discuss
hmm. after a bit of thinking, doesnt this pass force visit to all children, even though that dont necessarily need it? will making the right hand side a lazy value make more sense? --- [Visit Topic](https://discuss.tvm.ai/t/performing-relay-passes-non-recursively/5696/18) to respond.

[TVM Discuss] [Development] NDArray conversion to numpy array take time (around 350ms) with a simple data

2020-03-22 Thread Tu.Phan via TVM Discuss
Hi, I have built & tuned successfully the mxnet gluoncv model, and the result (classIds, scores, boxes) of the model referenced is NDArray. I want to draw boxes of class Person only, so I need to select these results by the condition. I got a big issue of performance with the tvm.runtime.ndar