Comments (24)
oook, so the problem is caused by sklearn, the new version broke backwards compatibility by removing the micro avg
metric and replacing it with accuracy
which has less columns in the classification_report ... working on a fix for both
ironically, it was my environment to be outdated, apologies.
from ergo.
yes
from ergo.
So I've inserted the following line at line 30 (original file, not the capture above):
if len(row) > len(measures):
And if you don't care about the json report it seems to be fixing it.
Side note: That'd be great to mention you also need to install python3-tk package somewhere in the doc
from ergo.
could you attach the dataset so I can reproduce it?
Besides:
from ergo.
Ok for the doc, thanks.
The dataset is yours, from the pe-av example.
from ergo.
I just loaded the dataset with no issue whatsoever, maybe some more details about your config? can you post here the output of ergo info
?
from ergo.
Sure. EDIT: Can you tell me what versions of tf/keras you have ?
Also Ubuntu 16.04, Python 3.5.2
from ergo.
Can you try:
sudo pip3 uninstall tensorflow
sudo pip3 install tensorflow==1.12.0
from ergo.
Ok, done. I'm restarting a training.
from ergo.
Hi Tigzy,
It's very odd that you don't have those metrics in the report. That's generated from sklearn. Can you please report the output of this code snippet:
from sklearn.metrics import classification_report
from sklearn import __version__
print (__version__)
a = [ 0,1,0,1]
b = [0,1,1,0]
print (classification_report(a,b))
from ergo.
tf to 1.12 didn't fix, let me get this output @nicochidt
from ergo.
to be honest i'm glad it didn't, it'd be very f weird
from ergo.
Here you go (don't mind my monkey typing :p)
It's not there either...
from ergo.
this is what you should be seeing ...
from ergo.
Maybe I should revert to version 0.20 ? Is that a python module ?
from ergo.
Wtf 0.0 version.... :/
from ergo.
Thanks a lot ! So in theory if I want to continue working on it today I just need to revert sklearn to 0.20 right ?
from ergo.
Oh I see there's a fix already, neat.
from ergo.
or just uninstall ergo-ai
and install this repo from sources
from ergo.
yeah yeah I was heading this way rather.
from ergo.
Was looking at the fix and shouldn't it be a label replacement somewhere as well? Where's that accuracy VS micro avg defined?
from ergo.
i don't understand the question ... the new version just does not present any micro average, but instead there's a new global accuracy metric
from ergo.
The question is: I see there's None returned when unable to parse properly
Isn't that possible to have actual accuracy (the old micro avg) values instead of just handle the parsing issue and ignore rows ?
from ergo.
Oh OK I think I understand. The new accuracy doesn't replace micro AVG, it's a totally other metric that only has f1-score and support, am I right ?
from ergo.
Related Issues (8)
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from ergo.