Comments (9)
I guess, you changed the CLASS_NUM(in the cfgs.py ) to 2, and you have restored the detection model with the Weights offered by us which trained in the pascal VOC.
Some solutions are as following:
Solution 1: You can load ImageNet's pretrained weights to train your data. Our program does not load the weights of Fast-RCNN/cls_fc when loading ImageNet Pretrained weights, so no error is reported.
Solution 2: Use the pre-trained model we provide to load, but you have to write some codes to control it does not load Fast-RCNN/cls_fc, because the shape is VOC trained weights is [2048, 21] while in your model is [2048, 3].
@leetesua
from faster-rcnn_tensorflow.
@yangJirui you are right, i changed the class num to 2. But how can i load imagenet's pretrained weights? Or how can i control not to load fast_rcnn/cls_fc? thanks a lot!
from faster-rcnn_tensorflow.
If you want to train your own dataset, please change the value of VERSION in cfgs.py. If not, model will load the weights which we provide according to the VERSION. @leetesua
from faster-rcnn_tensorflow.
1、If you want to train your own data, please note:
(1) Modify parameters (such as CLASS_NUM, DATASET_NAME, VERSION, etc.) in $PATH_ROOT/libs/configs/cfgs.py
(2) Add category information in $PATH_ROOT/libs/label_name_dict/lable_dict.py
(3) Add data_name to line 76 of $PATH_ROOT/data/io/read_tfrecord.py
from faster-rcnn_tensorflow.
I have changed the VERSION in cfgs from 'FasterRCNN_20180517' to '', so it won't load anything from output/trained_weight, however, when I run train.py again, it stops, with 'restore model' as the last printed line:
also, I am a little bit confused about,,, should i change the trained_weight in output folder or pretrained_weight in data folder ?
from faster-rcnn_tensorflow.
from faster-rcnn_tensorflow.
from faster-rcnn_tensorflow.
There are must something wrong with your tfrecord. @leetesua
from faster-rcnn_tensorflow.
Check your data and xml carefully.
from faster-rcnn_tensorflow.
Related Issues (20)
- 我在tensorflow1.8的环境使用pb的相关代码没有问题,但是我在1.12版本/1.11版本就会出现下面的问题,怎么半? HOT 2
- mAP of Mobilenet_v2 HOT 1
- decreasing the amount of ram used on gpu
- Incompatible shapes
- 关于OHEM HOT 2
- How to use HOT 3
- The precision HOT 1
- Does the roi_pooling func is roi align ? HOT 1
- frcnn_res50 evaluation results of pascal inconsistent with yours HOT 1
- 请问OHEM有用吗 HOT 1
- 运行inference.py,调用与训练模型,检测框和目标名混乱 HOT 1
- how to setting the number of train iteration about rpn and fast rcnn
- How to calculate the F1 score?? HOT 1
- NMS about postfastrcnn HOT 1
- RPN_MAXIMUM_PROPOSAL_TARIN = 2000 HOT 1
- error: python eval.py HOT 4
- Problems with resnet_50 and mobilenet HOT 1
- incompatible shapes
- about voc_eval.py
- about voc_eval.py
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 faster-rcnn_tensorflow.