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

how to get all model to generate datasets on my device? #100

Open
cosperypf opened this issue Jan 11, 2023 · 2 comments
Open

how to get all model to generate datasets on my device? #100

cosperypf opened this issue Jan 11, 2023 · 2 comments

Comments

@cosperypf
Copy link

I want to generate my training dataset (including model and corresponding latency) on my own device, I have read the article here (https://github.com/microsoft/nn-Meter/blob/ffd51e32c31026896fe2bda198b49fe5d756f184/docs/dataset.md ), it mentioned there is no model files but only the structure(in .jsonl files) of these model, so I havn't find any way to get the delay data of these model files running on my device.

So how can I get these models?

For example, you mentioned in the article that "it requires hundreds of GB storage to store the full dataset", is it possible to store these data in the network disk and open for download, or is there any other way for me to get my own dataset, so that I can get valuable datasets for me, thanks very much!

@cosperypf
Copy link
Author

I knew that I can generate kernel dataset just like here https://github.com/microsoft/nn-Meter/blob/ffd51e32c31026896fe2bda198b49fe5d756f184/docs/builder/build_kernel_latency_predictor.md, but this is not what I want, I want to get the datasets like here https://github.com/microsoft/nn-Meter/releases/download/v1.0-data/datasets.zip, but the latency should be on my device

@SakuraiYuuta
Copy link

I have the same question, I want to get the models in datasets.zip, and infer them by myself. How can I get the models?

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

2 participants