Comments (2)
It's hard to tell why your model is not learning based on the limited info you provided.
A good sanity check would be using the same train set and validation set during your first training run.
You can even use 2-10 images instead of 360.
http://karpathy.github.io/2019/04/25/recipe/
'overfit one batch. Overfit a single batch of only a few examples (e.g. as little as two). To do so we increase the capacity of our model (e.g. add layers or filters) and verify that we can reach the lowest achievable loss (e.g. zero). I also like to visualize in the same plot both the label and the prediction and ensure that they end up aligning perfectly once we reach the minimum loss. If they do not, there is a bug somewhere and we cannot continue to the next stage.'
from detectron2.
thanks for your advice, it is kind of u to provide an atricle here, which i would read.
in fact, accidently, i guess i have solved the problem, through changeing the category_id in my json_file to [1, categories].
the waring we got is in detectron2\detectron2\data\datasets\coco.py line 104, saying that it will apply a mapping
however, when i read the following codes, i find that there is no operation to category_id as it saied.
in addition, in line 437 in the same file, i see that the
"id" field must start with 1 if we want to use the COCO API,
and then, after changing, it works.
from detectron2.
Related Issues (20)
- missing config key error HOT 2
- Please read & provide the following HOT 1
- Detectron2 about rotated object detection HOT 1
- AttributeError: Cannot find field 'gt_masks' in the given Instances! HOT 1
- DensePose的apply_net.py运行dump的选项时候,如何多gpu运行呢? HOT 1
- Encountered freezing during start training at iteration 0 HOT 2
- printing label name and bbox coordinates of predicted images
- Add device argument for multi-backends access & Ascend NPU support HOT 6
- How to convert densepose model to onnx? HOT 1
- 模型跑出来的效果超出预期
- Does this project support FCOS? HOT 1
- C++ and onnx HOT 5
- The ce_loss became negative when I was using the mask2former to do instance segmentation HOT 4
- from . import deform_conv_cuda,ImportError: /home/user/repos/mmdetection/mmdet/ops/dcn/deform_conv_cuda.cpython-37m-x86_64-linux-gnu..so: undefined symbol: _ZN6caffe26detail37_typeMetaDataInstance_preallocated_32EI want to know the problem with this mmdet installation, my CUDA version is 10.1, pytorch version is 1.7.0, mmcv-full1.7.1 installed, mmdet2.10.0, if there is no other than the version is correct, but there is still this problem, it really crashes HOT 1
- Convert detectron2 to other frameworks - different outputs HOT 2
- Please read & provide the following HOT 3
- I can't train the model with batch size : 28 in linux environment but I can get the training results in windows with batch size 28 ! HOT 2
- Help with multi-GPU training HOT 2
- Does not install via requirements.txt but on command line
- New conda release HOT 2
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 detectron2.