I have 2 queries:
1. How can I run multiple inferences for a single train phase? I want to run slightly different inference to produce two sets of predictions using the same trained model. AFAIK, We can't access the saved trained model once the process is completed right?
2. If I submit 2 containers, are they run on two different GPUs?
My code was running on 100% GPU usage, so I was worried if I submit two containers to run at once, both may run on 1 GPU and compete for GPU time and both may get slowed.
So this won't be a problem. Two submitted containers will run on two different GPUs, so it will take the same time?
Created by BISPL Yes that was my query. Thank you. It's impossible. Thank you for your prompt reply.
I think you have misunderstood my question. I want to run multiple inference phases using the same test phase. That is I want to keep my trained model but make some changes to my inference code.
Using the submit button, I can just submit for Test phase 2. How can I change the inference phase code of one of my submitted containers with retaining my trained model?
For example, can I change the inference code of my submission with ID: 17da362f-afa6-4977-abeb-2996b851059b? 1. did you read this (https://www.synapse.org/#!Synapse:syn15569329/discussion/threadId=4370) ?
2. every containers have their own gpu!