Skip to content

EdgeTPU optimizations #6808

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.

Sign up for GitHub

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

Merged
glenn-jocher merged 9 commits into from
Mar 12, 2022
Merged

EdgeTPU optimizations #6808

Changes from 1 commit
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
3385498
removed transpose op for better edgetpu support
paradigmn Feb 28, 2022
48f9099
fix for training case
paradigmn Feb 28, 2022
e77f15c
Merge branch 'master' into edgetpu-optimizations
glenn-jocher Mar 4, 2022
8b31170
Merge branch 'master' of https://github.com/ultralytics/yolov5 into e…
paradigmn Mar 7, 2022
eb37fb9
enabled experimental new quantizer flag
paradigmn Mar 7, 2022
1c905a2
Merge branch 'edgetpu-optimizations' of github.com:paradigmn/yolov5 i…
paradigmn Mar 7, 2022
55f394b
precalculate add and mul ops at compile time
paradigmn Mar 7, 2022
40d4287
Merge branch 'master' into edgetpu-optimizations
glenn-jocher Mar 11, 2022
68fab69
Merge branch 'master' into edgetpu-optimizations
glenn-jocher Mar 12, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 5 additions & 3 deletions models/tf.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -222,12 +222,14 @@ def call(self, inputs):
x.append(self.m[i](inputs[i]))
# x(bs,20,20,255) to x(bs,3,20,20,85)
ny, nx = self.imgsz[0] // self.stride[i], self.imgsz[1] // self.stride[i]
x[i] = tf.transpose(tf.reshape(x[i], [-1, ny * nx, self.na, self.no]), [0, 2, 1, 3])
x[i] = tf.reshape(x[i], [-1, ny * nx, self.na, self.no])

if not self.training: # inference
y = tf.sigmoid(x[i])
xy = (y[..., 0:2] * 2 - 0.5 + self.grid[i]) * self.stride[i] # xy
wh = (y[..., 2:4] * 2) ** 2 * self.anchor_grid[i]
grid = tf.transpose(self.grid[i], [0, 2, 1, 3])
anchor_grid = tf.transpose(self.anchor_grid[i], [0, 2, 1, 3])
xy = (y[..., 0:2] * 2 - 0.5 + grid) * self.stride[i] # xy
wh = (y[..., 2:4] * 2) ** 2 * anchor_grid
# Normalize xywh to 0-1 to reduce calibration error
xy /= tf.constant([[self.imgsz[1], self.imgsz[0]]], dtype=tf.float32)
wh /= tf.constant([[self.imgsz[1], self.imgsz[0]]], dtype=tf.float32)
Expand Down