Fix the NaN bug
There are currently no pipelines.
To run a merge request pipeline, the jobs in the CI/CD configuration file must be configured to run in merge request pipelines and you must have sufficient permissions in the source project.
Previous version of the PTQ implementation could result in a model producing NaN values.
This was due to a poor memory initialization obtained by doing a multiplication by zero.
Thus, rescaleTensor(tensor, 0)
is now replaced by fillTensor(tensor, 0)
which solves the problem.
To run a merge request pipeline, the jobs in the CI/CD configuration file must be configured to run in merge request pipelines and you must have sufficient permissions in the source project.
Copyright © Eclipse Foundation, Inc. All Rights Reserved. Privacy Policy | Terms of Use | Copyright Agent