Recipe failure from Sparsify in SparseML

neuralmagic / sparseml

Libraries for applying sparsification recipes to neural networks with a few lines of code, enabling faster and smaller models

Apache License 2.0

2.08k stars 148 forks source link

Recipe failure from Sparsify in SparseML #670

Closed markurtz closed 2 years ago

markurtz commented 2 years ago

Hi @tlrmchlsmth , Thanks for the fix. After install new verion:

deepsparse           0.11.2
sparseml             0.11.1
sparsezoo            0.11.0
sparsify             0.11.0

It seems that the old err is solved, but I got another err, please the the attached image. Screen Shot 2022-03-31 at 11 11 31 r

Originally posted by @edwardnguyen1705 in https://github.com/neuralmagic/deepsparse/issues/279#issuecomment-1084055401

markurtz commented 2 years ago

Hi @edwardnguyen1705 we're actively looking into this. In the interim, are there further details you could provide about the model and training pipeline you're using so we can get you a working recipe?

Thanks, Mark

jeanniefinks commented 2 years ago

Hi @edwardnguyen1705 Just checking in to see if you are able to provide more details about your model and training pipeline? Thank you! Jeannie / Neural Magic

edwardnguyen1705 commented 2 years ago

Thanks for asking, I just tried a pre-trained public model: NanoDet-Plus-m-320

markurtz commented 2 years ago

Hi @edwardnguyen1705, we've landed this diff now on main and will be cutting out a hotfix release by early next week, will update here once that is available: https://github.com/neuralmagic/sparsify/pull/110

dbogunowicz commented 2 years ago

The hotfix is now in place. I am fairly confident we can close this for now.