未验证 提交 a305fff4 编写于 作者: S SunAhong1993 提交者: GitHub

Update README.md

上级 889c859f
...@@ -6,7 +6,7 @@ A demo to show converting caffe models trained on 'imagenet' using caffe2fluid ...@@ -6,7 +6,7 @@ A demo to show converting caffe models trained on 'imagenet' using caffe2fluid
1. Prepare python environment 1. Prepare python environment
2. Download caffe model to "models.caffe/xxx" which contains "xxx.caffemodel" and "xxx.prototxt" 2. Download caffe model to "../../models/xxx" which contains "xxx.caffemodel" and "xxx.prototxt"
3. Convert the Caffe model to Fluid model 3. Convert the Caffe model to Fluid model
- generate fluid code and weight file - generate fluid code and weight file
...@@ -28,14 +28,14 @@ A demo to show converting caffe models trained on 'imagenet' using caffe2fluid ...@@ -28,14 +28,14 @@ A demo to show converting caffe models trained on 'imagenet' using caffe2fluid
5. convert model and do inference together 5. convert model and do inference together
``` ```
bash ./tools/run.sh alexnet ./models.caffe/alexnet ./models/alexnet bash ./tools/run.sh alexnet ../../models ../../models
``` ```
* Assume the Caffe model is stored in '*./models.caffe/alexnet/alexnet.prototxt|caffemodel*' * Assume the Caffe model is stored in '../../models/alexnet.prototxt|caffemodel*'
* converted model will be stored as '*./models/alexnet/alexnet.py|npy*' * converted model will be stored as '../../models/alexnet.py|npy*'
6. test the difference with caffe's results(need pycaffe installed) 6. test the difference with caffe's results(need pycaffe installed)
``` ```
bash ./tools/diff.sh alexnet ../../models/ ../../models bash ./tools/diff.sh alexnet ../../models/ ../../models
``` ```
* Make sure your caffemodel stored in '*./models.caffe/resnet*' * Make sure your caffemodel stored in '../../models/alexnet.prototxt|caffemodel*'
* The results will be stored in '*./results/resnet.paddle|caffe*' * The results will be stored in '*./results/resnet.paddle|caffe*'
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册