Thanks again for creating this great product. I have a couple of questions:
- Can we read in video data as input to our batch transform jobs?
- What is the limit to the processing time per-input for batch transform jobs? How can we increase this?
- If 1) is currently not supported, how do you suggest we achieve this functionality? Is it a good idea to download the video from S3 ourselves and go on from there?
Related Question:
aws/sagemaker-mxnet-inference-toolkit#77
Thanks