Hi Arne! As one of the developers of AIscm, I think MNIST is ready for you. Currently, most critical things are provided, not only Tensorflow, but also imagick/ffmpeg/opencv...etc. However, for real cases, we still need user friendly APIs. Now, most things are too low level. I have a plan to write a framework based on AIscm, just like Keras to Tensorflow.
I just need more time slots. Best regards. Arne Babenhauserheide <arne_...@web.de> 于 2019年10月24日周四 04:37写道: > Hi Jan, > > Jan Wedekind <j...@wedesoft.de> writes: > > Hi Basile, > > I have implemented Tensorflow bindings for Guile: > http://wedesoft.github.io/aiscm/ > > Let me know if you have any problems installing it. > > Is aiscm already at a stage where I could experiment with using it for > handwriting recognition even though I don’t have prior experience with > setting up machine learning pipelines? > > Does it work with a simple intel GPU or on CPU? > > There’s a project I’ve wanted to start for quite some time now. > > Best wishes, > Arne > -- > Unpolitisch sein > heißt politisch sein > ohne es zu merken >