-
Notifications
You must be signed in to change notification settings - Fork 6.8k
Conversation
Do you mind provide a scoring script for the performance testing? |
Is the custom |
@ZiyueHuang Great updates in module. Will update soon. @pengzhao-intel The scope of this pull request is really to maintain the example code to a usable but accurate reference. Performance benchmarking will be included in addition to matching the accuracy in the development of new gluon-cv toolkit. |
Thanks for the information. It's fine for us :) |
@ZiyueHuang MutableModule now gone. |
Conflict with an open pr: #11013. This PR will remove all Cython modules completely. No need for windows fix. |
Can we change the backbone network symbol definition to the gluon version(then symbolic)? |
Could you please point a link to the gluon version of backbone network definition? |
So you mean gluon block definition, gluon training and gluon inference, which is basically a new example. Please see https://gluon-cv.mxnet.io/model_zoo/index.html#object-detection for that. |
Get symbol as following: import mxnet as mx
from mxnet import gluon
import mxnet.gluon.model_zoo.vision as vision
net = vision.resnet50_v2(pretrained=True)
net.hybridize()
data = mx.sym.var('data')
sym = net(data) |
Thanks for the example. Unfortunately definition in To preserve the network definition, we could translate the current symbol definition to As mentioned above, |
- python3 compactible - remove cuda operator, cython utility, pycocotools - remove mutablemodule - remove duplicate code
I found it much slower than old version,and the speed is not stable,I don't change any param, does anyone else have the same problem like me? |
It could be caused by varying CPU and disk workload. Training speed is bottlenecked by single thread python data loading. |
This example was heavily criticized by many due to its complicated engineering optimizations. The PR was meant to improve clarity and simplicity. To start simple, use this example. To achieve state of the art, check out other research implementations in MXNet based on the previous heavily engineered version that provide better performance. To get both simplicity and performance, stay tuned for https://github.com/dmlc/gluon-cv. |
do you try to use MKL-DNN backend? |
- python3 compactible - remove cuda operator, cython utility, pycocotools - remove mutablemodule - remove duplicate code
Description
People complain about the Faster R-CNN example a lot.
python3 train.py -h
to see all of them.Checklist
Essentials
Changes
All changes are limited to example/rcnn folder. Basically no new features are added.
Comments
Note that data processing speed of the pure numpy code could be a limitation, especially for crowded dataset like COCO. However, numpy code is concise and accurate, which can serve as future reference to build better implementations.