forked from Tencent/DepthCrafter
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathREADME_4_GUI.txt
55 lines (45 loc) · 1.7 KB
/
README_4_GUI.txt
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
==================================================
DepthCrafter GUI - Parameter Guide
==================================================
DESCRIPTION
-----------
DepthCrafter processes videos into temporally consistent depth maps using advanced video diffusion models. Below is an explanation of the configurable parameters.
==================================================
INPUT/OUTPUT PARAMETERS
==================================================
1. Input Folder
- Directory containing the video files to process.
- Default: ./input_clips
2. Output Folder
- Directory to save the processed depth videos.
- Default: ./depthmap_output
==================================================
PROCESSING PARAMETERS
==================================================
3. Guidance Scale
- Adjusts the balance between model creativity and fidelity to the input.
- Default: 1.0
4. Inference Steps
- Number of steps for depth generation. Higher values improve quality but take longer.
- Default: 5
5. Window Size
- Number of frames processed per batch. Larger values may require more memory.
- Default: 110
6. Maximum Resolution
- Maximum width of the video frames.
- Default: 960
7. Overlap
- Number of overlapping frames between processing batches to ensure smooth transitions.
- Default: 25
8. Seed
- Sets the random seed for reproducibility of results.
- Default: 42
==================================================
RESOURCE MANAGEMENT
==================================================
9. CPU Offload Mode
- Optimizes memory usage during processing.
- Options:
- model: Faster
- sequential: Slower
- Default: model