Skip to content
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.

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

update DFQ/EQ/Evaluate int8 perchannel quant tool #1112

Merged
merged 2 commits into from
Sep 2, 2021

Conversation

BowShotDS
Copy link
Collaborator

Add DFQ and EQ quantization schemes for int8 perchannel quantization.
Add quantitative loss assessment function and corresponding input options.

@codecov
Copy link

codecov bot commented Sep 2, 2021

Codecov Report

Merging #1112 (d1b68d8) into tengine-lite (baf47d3) will increase coverage by 0.32%.
The diff coverage is 1.20%.

❗ Current head d1b68d8 differs from pull request most recent head 120adf9. Consider uploading reports for the commit 120adf9 to get more accurate results
Impacted file tree graph

@@               Coverage Diff                @@
##           tengine-lite    #1112      +/-   ##
================================================
+ Coverage         47.70%   48.03%   +0.32%     
================================================
  Files               419      295     -124     
  Lines             31151    16019   -15132     
================================================
- Hits              14861     7694    -7167     
+ Misses            16290     8325    -7965     
Impacted Files Coverage Δ
tools/quantize/algorithm/quant_dfq.cpp 0.00% <0.00%> (ø)
tools/quantize/algorithm/quant_eq.cpp 0.00% <0.00%> (ø)
tools/quantize/quant_save_graph.cpp 42.93% <0.00%> (-0.25%) ⬇️
tools/quantize/quant_utils.cpp 56.52% <0.00%> (-2.09%) ⬇️
tools/quantize/quant_tool_int8.cpp 45.18% <9.70%> (-19.13%) ⬇️
source/operator/prototype/unsqueeze.c 12.50% <0.00%> (-83.93%) ⬇️
source/operator/prototype/priorbox.c 14.00% <0.00%> (-82.00%) ⬇️
source/operator/prototype/gru.c 17.07% <0.00%> (-80.49%) ⬇️
source/operator/prototype/matmul.c 18.91% <0.00%> (-78.38%) ⬇️
source/operator/prototype/detection_output.c 18.91% <0.00%> (-78.38%) ⬇️
... and 202 more

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update baf47d3...120adf9. Read the comment docs.

Copy link
Contributor

@BUG1989 BUG1989 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

lgtm!

@BUG1989 BUG1989 merged commit 3e71f04 into OAID:tengine-lite Sep 2, 2021
@BowShotDS BowShotDS deleted the update branch September 24, 2021 06:35
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants