Nice work, @certik! I hope to talk with you soon about how we might collaborate in this area and what synergies there might be with Inference-Engine. I’m curious what file format you use. Inference-Engine uses a JSON file exported from PyTorch. I’m also investigating ONNX.
1 Like