Show HN: GPT-2 inference in pure C#, 0 bytes allocated per tokengithub.com1 points·by dev-on-bike·قبل شهرين·1 comments