~/talk

TensorFlow Lite

on YouTube 34m
Talkat Mobile Era 2018 · 2018-11-01..11-02· [machine-learning]

speaker

name:Natalie Pistunovich
x:nataliepis

resources

video:https://youtu.be/YohUUDyaRDc
slides:none on file

summary

TensorFlow Lite brings on-device inference to Android, iOS and Raspberry Pi with a compact TensorFlow Lite runtime, FlatBuffers model files and NNAPI/OpenGL acceleration coming soon. Use the TOCO converter to quantize models: 25 float models remain 2–355 MB and run 30 ms–8 s with ~72–96% accuracy, while 18 quantized models shrink below ~23 MB and run under ~637 ms with modest accuracy loss. Example apps (Hotdog/Inception) and AutoML-generated models prove the workflow works.

AI-written summary · about

chapters (7)

  1. 0:00Start
  2. 2:47How to ML
  3. 11:51TensorFlow Lite Converter
  4. 12:04Environments
  5. 22:15Float Models
  6. 22:46Quantized Models
  7. 26:25Apps Using TFLite

related talks

~/search
type to search 6,249 conferences, 138,987 speakers and 185,172 talks
~/help
cmdctrl k
search every conference, speaker and series
/
focus the filter search
h
home
c
conferences
s
speakers
e
series
t
light / dark — the status bar button does the same
?
this help
esc
close this, then clear every filter — or just the search, while you are typing in it

Ignored while you are typing. Pages without a filter rail have no search to focus.