Joypop commited on
Commit
213d87c
Β·
verified Β·
1 Parent(s): 30ddab0

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +6 -4
README.md CHANGED
@@ -66,14 +66,16 @@ Download the [OpenImage dataset](https://storage.googleapis.com/openimages/web/i
66
 
67
 
68
  ### Step2: Train NAOSD.
69
- β€˜β€™β€˜
70
  bash scripts/train/train_NAOSD.sh
71
- β€™β€˜β€™
 
72
 
73
  ### Step3: Train GDPO-SR
74
- β€˜β€™β€˜
75
  bash scripts/train/train_GDPOSR.sh
76
- β€™β€˜β€™
 
77
 
78
  ## πŸ”— Citations
79
 
 
66
 
67
 
68
  ### Step2: Train NAOSD.
69
+ ```
70
  bash scripts/train/train_NAOSD.sh
71
+ ```
72
+ The hyperparameters in train_NAOSD.sh can be modified to suit different experimental settings. Besides, after training with NAOSD, you can use GDPOSR/mergelora.py to merge the LoRA into the UNet and VAE as base model for subsequent reinforcement learning training and inference.
73
 
74
  ### Step3: Train GDPO-SR
75
+ ```
76
  bash scripts/train/train_GDPOSR.sh
77
+ ```
78
+ The hyperparameters in train_GDPOSR.sh can be modified to suit different experimental settings. Besides, after training with GDPO-SR, you can use GDPOSR/mergelora.py to merge the LoRA into the UNet for subsequent inference.
79
 
80
  ## πŸ”— Citations
81