This is an automated email from the ASF dual-hosted git repository.
zhreshold pushed a commit to branch master
in repository https://gitbox.apache.org/repos/asf/incubator-mxnet.git
The following commit(s) were added to refs/heads/master by this push:
new b4161b0 [MXNET-288] Fix ssd example (#10427)
b4161b0 is described below
commit b4161b00201bd9c372b738e7e168d01e209749a3
Author: ThomasDelteil <thomas.delte...@gmail.com>
AuthorDate: Thu Apr 5 23:54:49 2018 -0700
[MXNET-288] Fix ssd example (#10427)
* Update README.md
* Update README.md
---
example/ssd/README.md | 22 +++++++++++++---------
1 file changed, 13 insertions(+), 9 deletions(-)
diff --git a/example/ssd/README.md b/example/ssd/README.md
index 5759fca..0b97092 100644
--- a/example/ssd/README.md
+++ b/example/ssd/README.md
@@ -69,9 +69,13 @@ insanely slow. Using CUDNN is optional, but highly
recommended.
### Try the demo
* Download the pretrained model:
[`ssd_resnet50_0712.zip`](https://github.com/zhreshold/mxnet-ssd/releases/download/v0.6/resnet50_ssd_512_voc0712_trainval.zip),
and extract to `model/` directory.
+
* Run
```
-# cd /path/to/mxnet-ssd
+# cd /path/to/incubator-mxnet/example/ssd
+# download the test images
+python data/demo/download_demo_images.py
+# run the demo
python demo.py --gpu 0
# play with examples:
python demo.py --epoch 0 --images ./data/demo/dog.jpg --thresh 0.5
@@ -102,20 +106,20 @@ The suggested directory structure is to store `VOC2007`
and `VOC2012` directorie
in the same `VOCdevkit` folder.
* Then link `VOCdevkit` folder to `data/VOCdevkit` by default:
```
-ln -s /path/to/VOCdevkit /path/to/mxnet/example/ssd/data/VOCdevkit
+ln -s /path/to/VOCdevkit /path/to/incubator-mxnet/example/ssd/data/VOCdevkit
```
Use hard link instead of copy could save us a bit disk space.
* Create packed binary file for faster training:
```
-# cd /path/to/mxnet/example/ssd
+# cd /path/to/incubator-mxnet/example/ssd
bash tools/prepare_pascal.sh
# or if you are using windows
python tools/prepare_dataset.py --dataset pascal --year 2007,2012 --set
trainval --target ./data/train.lst
-python tools/prepare_dataset.py --dataset pascal --year 2007 --set test
--target ./data/val.lst --shuffle False
+python tools/prepare_dataset.py --dataset pascal --year 2007 --set test
--target ./data/val.lst --no-shuffle
```
* Start training:
```
-# cd /path/to/mxnet/example/ssd
+# cd /path/to/incubator-mxnet/example/ssd
python train.py
```
* By default, this example will use `batch-size=32` and `learning_rate=0.002`.
@@ -129,23 +133,23 @@ python train.py --gpus 0,1,2,3 --batch-size 32
### Evalute trained model
Make sure you have val.rec as validation dataset. It's the same one as used in
training. Use:
```
-# cd /path/to/mxnet/example/ssd
+# cd /path/to/incubator-mxnet/example/ssd
python evaluate.py --gpus 0,1 --batch-size 128 --epoch 0
```
### Convert model to deploy mode
This simply removes all loss layers, and attach a layer for merging results
and non-maximum suppression.
Useful when loading python symbol is not available.
```
-# cd /path/to/mxnet/example/ssd
+# cd /path/to/incubator-mxnet/example/ssd
python deploy.py --num-class 20
```
### Convert caffe model
-Converter from caffe is available at
`/path/to/mxnet/example/ssd/tools/caffe_converter`
+Converter from caffe is available at
`/path/to/incubator-mxnet/example/ssd/tools/caffe_converter`
This is specifically modified to handle custom layer in caffe-ssd. Usage:
```
-cd /path/to/mxnet/example/ssd/tools/caffe_converter
+cd /path/to/incubator-mxnet/example/ssd/tools/caffe_converter
make
python convert_model.py deploy.prototxt
name_of_pretrained_caffe_model.caffemodel ssd_converted
# you will use this model in deploy mode without loading from python
symbol(layer names inconsistent)
--
To stop receiving notification emails like this one, please contact
zhresh...@apache.org.
