It basically achieves 90% accuracy at BBC News Articles with only 40 minutes to 1 hour iterations-training.
I suggest directly skipping to nmf_train_optimized function.
alperkaya0/nmf-raw-implementation
Folders and files
| Name | Name | Last commit date | ||
|---|---|---|---|---|
| Name | Name | Last commit date | ||
|---|---|---|---|---|
It basically achieves 90% accuracy at BBC News Articles with only 40 minutes to 1 hour iterations-training.
I suggest directly skipping to nmf_train_optimized function.