Closed rohis06 closed 8 months ago
Sorry, can you explain the failure? Does it get stuck at 23%? From your screenshot it just looks like it's in progress.
Yes, that's correct! It essentially terminates at 23%, and the control falls to the following statement in the code.
@swheaton, kindly let me know if you need any other details to debug the issue.
@rohis06 it seems that we just don't handle a 206 partial content response. I don't know if there is a simple resolution to it. If you'd like to look into supporting this mode of operation within eta.core.web.download_file()
that would be awesome and the fastest path to resolution!
Otherwise it seems too niche of a use case to be prioritized for the core team, unfortunately.
I believe you are trying to add the Places dataset to the fiftyone zoo? (Appreciate it!) Perhaps @jacobmarks has ideas for workarounds to this issue with downloading the data?
@swheaton, that makes sense. I'd be glad to explore how to support this mode of operation within eta.core.web.download_file()
!
Yes, I'm attempting to add the Places dataset to the fiftyone Zoo! :) Certainly, I'll reach out to @jacobmarks. One thing I'd like to bring to your attention is that the failure to download the Places dataset doesn't occur consistently. It only happens when the internet speed is <150Mbps. Otherwise, the download proceeds smoothly.
Resolved by https://github.com/voxel51/eta/pull/621
Of late, I have noticed that
eta.core.web.download_file()
is failing while downloading large files. For example:eta.core.web.download_file("http://data.csail.mit.edu/places/places365/test_256.tar", path=<my-path>)
This especially happens when the file is being downloaded in multiple chunks.
Output using eta.core.web.download_file():
Here's the wget output for the same:
Any help here would be appreciated.