Closed rafzi closed 4 years ago
Hi Rafael, I will need to restore some old functionality (being able to name the model invoking functions (e.g. if you include more than one model in one binary) and passing more than one input/output tensor), but I already integrated your changes into the branch devel. Also I plan to further reduce the code size by using arrays instead of linear code and do some size measurements on riscv32im. I like the more efficient memory allocation!
Thanks for the comments, they should all be addressed now!