꯸ ꬰ ꝛ⼻ↇ comments

Results 6 comments of


                                            ꯸ ꬰ ꝛ⼻ↇ

Set image source after setting decode size to fix #71

1 year later, it'd be cool if this were implemented... I think I'll have to selfbuild and sideload

documentation/1972 - Updated scoring file to reflect current implementation

I wasn't sure if I should remove them or if they were there because they were goals that haven't been implemented yet... Let me know if I should just delete...

documentation/1972 - Updated scoring file to reflect current implementation

I don't really care much about scores either, but it's useful to see at a glance if you're doing something majorly wrong without realizing (e.g. restricted areas). Also super useful...

Conflicting packages `eslint` and `lint-diff`

A temporary workaround is `npm install --force`

GPU for Model Inference + Real time transcription

I think this issue should remain open until someone finds a way of doing this. GPU inference should be much faster than the current method.

GPU for Model Inference + Real time transcription

So I installed CUDA 10.0 (required for `tensorflow-gpu==1.14.0`), and I got it to run on GPU. It raises the following error: Unimplemented: Generic conv implementation only supports NHWC tensor format...