david/DeepFaceLab_Linux
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

upd readme

Browse Source
This commit is contained in:
iperov
2019-02-03 23:19:56 +04:00
parent 707313eb47
commit f0a5f97995
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files
README.md
+1 -1
View File
@@ -134,7 +134,7 @@ SAE tips:
- how to train extremely obstructed face model with SAE: - how to train extremely obstructed face model with SAE:
First train it with both style powers at 10.0+ value. When artifacts become appearing at ~30-40k epochs, set face style to 0.0 or 0.01 and bg style to 0.1-0.3 and continue training. You can slightly vary theese values during training if something goes wrong. Track changes in preview history. First train it with both style powers at 10.0+ value. When artifacts become appearing at ~30-40k epochs, set face style to 0.0 or 0.01 and bg style to 0.1-0.3 and continue training. You can slightly vary theese values during training if something goes wrong. If the colors do not match, increase styles to 1.0 - 3.0. Experiment on your own. Track changes in preview history.
Improperly matched dst landmarks may significantly reduce fake quality: Improperly matched dst landmarks may significantly reduce fake quality: