search

koriavinash1 / DeepBrainSeg

Fully automatic brain tumour segmentation using Deep 3-D convolutional neural networks
MIT License
185 stars 43 forks source link

Fine-tuning pipeline for densenet3D mdodel #13

Closed koriavinash1 closed 4 years ago

koriavinash1 commented 4 years ago

initial version for model fine-tuning for the tiramisu 3D model (Training scripts for tumor segmentation problem).

files added (DeepBrainSeg/tumor/finetuning/*.py) compatibility with API will be updated soon...

This also includes selective hard-mining of patches

pull-assistant[bot] commented 4 years ago
Score: 0.72

Best reviewed: commit by commit

Optimal code review plan (4 warnings)

fine tuning setup
> `...umor/finetune/dataGenerator.py` 53% changes removed in conflict fixes
generateCSV for finetuning
> `...Seg/tumor/finetune/training.py` 52% changes removed in training working
class wise csv
> `...Seg/tumor/finetune/training.py` 81% changes removed in data generator for h...
     training pipeline
training pipeline
> `...Seg/tumor/finetune/training.py` 50% changes removed in data generator for h...
     'generateCSV      generate csv for hard mining and fine tuning      data generator for hard mining      data generator for hard mining      data generator working but slow      traiing import fixes      conflict fixes      training working

Powered by Pull Assistant. Last update b19d3b4 ... b297fbc. Read the comment docs.