QAT support for facial landmark detection task

[yrajas]

Pull Request Template

Description

Add quantization aware training support for facial landmark detection task models.

Fixes: # (issue)

Changes

Please delete options that are not relevant.

• [x] New feature (non-breaking change which adds functionality)

Configuration:

• Archai version: 1.0.0
• Environment:
• OS:

Final checks:

• [ ] Follows guidelines of project
• [ ] Self-review of code
• [ ] Commented code in hard-to-understand areas
• [ ] Corresponding changes to the documentation
• [ ] Changes generate no new warnings
• [ ] Dependent changes have already been merged
• [ ] Checked code correction any misspellings

[wchen-github]

Also, can you do me a couple of favors? Looks like I left line #97 in model.py for no reason. Then, on line 74 in model.py, maybe we should take away the devault value. For the context of this task, 1000 makes no sense and we must pass in an explicit num_classes.

[yrajas]

[like] Yoganand Rajasekaran reacted to your message:

---

From: wchen-github @.> Sent: Tuesday, May 2, 2023 7:09:53 PM To: microsoft/archai @.> Cc: Yoganand Rajasekaran @.>; Assign @.> Subject: Re: [microsoft/archai] QAT support for facial landmark detection task (PR #240)

Also, can you do me a couple of favors? Looks like I left line #97https://github.com/microsoft/archai/pull/97 in model.py for no reason. Then, on line 74 in model.py, maybe we should take away the devault value. For the context of this task, 1000 makes no sense and we must pass in an explicit num_classes.

— Reply to this email directly, view it on GitHubhttps://github.com/microsoft/archai/pull/240#issuecomment-1532007948, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AOMSXA42WWESGUJRELMTLS3XEFLYDANCNFSM6AAAAAAXTLF4AM. You are receiving this because you were assigned.Message ID: @.***>