Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

The format of fbuf and 3 parts of nvdla #2

Open
xiao777dong opened this issue Aug 13, 2019 · 4 comments
Open

The format of fbuf and 3 parts of nvdla #2

xiao777dong opened this issue Aug 13, 2019 · 4 comments

Comments

@xiao777dong
Copy link

I'm sorry to bother you. I'd like to ask some questions. What format fbuf is and how to generate it? I saw three nvdlas in the file. How to generate each separate part of nvdla? Looking forward to reply. Thank you.

@prasshantg
Copy link
Owner

@xiao777dong this was done through some private tools. NVDLA compiler does not support this yet.

@spaceotter
Copy link

Is there a way to generate these files, now that the compiler has been released as open source?

@prasshantg
Copy link
Owner

@spaceotter you need to break the YOLOv3 model and compile separately excluding unsupported layers to generate 3 loadables and integrate those in Darknet application.

@ku-researcher
Copy link

@prasshantg
I'm sorry to bother you.
Please teach me the specific steps to generate .nvdla files.
Now I succeeded in compiling caffemodel into .nvdla files by nvdla_compiler built from sources in https://github.com/nvdla/sw.
And I was able to run alexnet on firesim by using your alexnet-odla.cfg and alexnet-odla.nvdla,
but I can't run resnet-50 by using my converted .nvdla file.
So, could you tell me whether you compile model by using nvdla_compiler and otherwise
share me the way to convert?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

4 participants