Comments (3)
Yes, I think this should work since the NR should lower the loss monotonically.
So, if the loss raises in one step, there has to be something wrong.
Since we are calculating the loss in each step anyways, we can just keep the feature probabilities of the last step in a tf.Variable, so performance cost of this check can be kept very low.
from batchglm.
Yes, I think this should work since the NR should lower the loss monotonically.
This is not true unfortunately, it can overstep in the same way as GD can overstep, it is less likely to do so however. But usually it only oversteps very close to the MLE on these small systems.
Since we are calculating the loss in each step anyways, we can just keep the feature probabilities of the last step in a tf.Variable, so performance cost of this check can be kept very low.
That would be good, can you implement that coming week, please?
from batchglm.
Done in #55, referenced as feature-wise termination.
from batchglm.
Related Issues (20)
- 'Simulator' object has no attribute 'np_clip_param' when following diffxpy tutorial HOT 2
- [TF2] Do we need float64? HOT 5
- Controlling verbosity with an argument and proper logging HOT 1
- expand dims bug with da.test.wald HOT 1
- print(test.model) throws RecursionError HOT 2
- tests: cannot import name 'Simulator' from 'batchglm.api.models' HOT 4
- Make Unit Tests Work
- Model Status Tracker
- multi-dispatch for `init_par`
- Nicer model initialisation HOT 2
- [python 3.10] add switch case
- Move `InputDataGLM` to a static util
- Document what `ModelContainer` does
- Check that get_model_container() works with diffxpy interface
- Stop using multiprocessing for fitting dispersion models earlier HOT 1
- Error when batch_size[1] <= number of genes HOT 2
- New Release Issues
- Make `init_par` a class Method HOT 1
- Investigate Removing `.params`
- Not compatible with Python 3.10 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 batchglm.