BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//cfp.devcon.org//devcon8//talk//AU7YLW
BEGIN:VTIMEZONE
TZID:IST
BEGIN:STANDARD
DTSTART:20000101T000000
RRULE:FREQ=YEARLY;BYMONTH=1
TZNAME:IST
TZOFFSETFROM:+0530
TZOFFSETTO:+0530
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
UID:pretalx-devcon8-AU7YLW@cfp.devcon.org
DTSTART;TZID=IST:20261105T144500
DTEND;TZID=IST:20261105T150000
DESCRIPTION:Training loops shouldn't need boilerplate just to see a loss cu
 rve. As part of my GSoC 2026 work with FOSSASIA on Visdom\, I built Visdom
 Logger — a native logging API that streams metrics straight to Visdom's 
 live dashboard. It shipped in Visdom v0.3.0 and is already usable today.\n
 \nVisdomLogger's design keeps Visdom's philosophy intact it receives value
 s through a simple log(name\, value) call and never instruments training i
 tself. That boundary is what makes true autologging possible on top of it 
 — and in this talk I'll show the real thing a Keras callback that hooks 
 into [model.fit](http://model.fit)() and streams loss\, accuracy\, and any
  tracked metric to Visdom automatically\, with zero manual logging calls i
 n the training loop.\n\nIn this lightning talk\, I'll run a live Keras tra
 ining script with no logging boilerplate and watch the dashboard populate 
 in real time\, then briefly show the same pattern working for PyTorch. I'l
 l close with what's next extending true autologging to PyTorch Lightning\,
  XGBoost\, and scikit-learn on a shared backend.
DTSTAMP:20261101T051201Z
LOCATION:Community Led Stage
SUMMARY:Zero-Boilerplate Dashboards: Live Autologging for Keras and PyTorch
  with Visdom - Jayant Parashar
URL:https://cfp.devcon.org/devcon8/talk/AU7YLW/
END:VEVENT
END:VCALENDAR