Comments (3)
When you load from a checkpoint, it doesn't restart from epoch 1; it will start at the last saved epoch. So if the checkpoint was saved at epoch 3 and you load a model from that checkpoint, training will resume with weights that were saved at epoch 3. However, when the trainer reports its current progress, it will report itself as only being on epoch 1, despite using epoch 3 weights. This mis-reporting was due to a limitation in the beta version of Tensorflow 2.0 which doesn't allow you to specify an epoch starting point.
Once we upgrade to the full release version of Tensorflow 2.0 I will re-evaluate this and see if we're able to fix this discrepancy.
from slideflow.
Okay I won't worry about it for now.
What about restarting from k-fold 1 or 2 and needing to finish out all 3? Is there functionality for that?
from slideflow.
You should be able to do that now. In the train() function, use the argument k_fold_iter to specify which iteration of the k-fold you want to train. Use the checkpoint argument to specify which checkpoint you are loading (usually cp.ckpt in the model folder).
from slideflow.
Related Issues (20)
- [BUG] API doc for Project.extract_tiles is incomplete HOT 1
- Concatenate error with multimodal input HOT 17
- [BUG] Slideflow Studio crashes on M1 Mac, Mac installation bugs HOT 2
- [BUG] Slideflow Studio crashes when doing ROIs HOT 2
- Protobuf dependency conflict HOT 2
- saving models HOT 5
- AUC =0.5 HOT 7
- mil_config model_size argument being silently thrown out by default HOT 3
- [BUG] sf.Heatmap() function incompatible with MIL model HOT 1
- [BUG] build_feature_extractor HOT 5
- TypeError: __init__() got an unexpected keyword argument 'multi_gpu' HOT 3
- MRXS Slide Ramdisk buffer Bug HOT 1
- [BUG] val_strategy='bootstrap' does not create .json file for the bootstrapped split to be referenced later HOT 1
- Generate features from SimCLR model HOT 2
- Saving tiles with the ROI index HOT 3
- [Documentation] Optimization space documentation missing HOT 3
- [BUG] Providing extractor argument to sf.mil.eval.predict_slide throws error HOT 1
- [BUG] Datasets do not pass `min_tiles` filter when training GANs HOT 3
- An approach to consider multiple slides per patient with weak labels. HOT 7
- [BUG] Unable to tile WSIs 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 slideflow.