-
Notifications
You must be signed in to change notification settings - Fork 1
Expand file tree
/
Copy pathTODO.txt
More file actions
54 lines (50 loc) · 4.22 KB
/
TODO.txt
File metadata and controls
54 lines (50 loc) · 4.22 KB
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
HIGH:
1. Modularize and share code b/n dream and deepdream # DONE
2. Implement class for ImageToVideo converter # DONE
3. Implement optional backward loop for video # DONE
4. Implement video framing for deep dream # DONE
5. Handle unmatched dim sizes for dream # DONE
6. Implement validators as decorators for dream and deepdream # DONE
7. Remove redundant or unnecessary type or numpy-tensor (and, vice versa) conversions
8. Add more comments. Duhh
9. Optimize loop of simple dream (seems to slowed down. YIKES!) # DONE
10. Optimize video making by insteading of buffering (in sys memory), save octave by octave # DONE
then concatenate the video (.mp4) files. N.b., keep the buffering method to support # ^^^^
CPU-only runtimes # ^^^^
11. Adaptive frames to interpolate to ensure 30fps # DONE BUT UNNEEDED
12. Modularize deepdream # DONE
13. Consider the mirror_video option on the duration by dividing the duration # DONE
by half (as mirroring doubles) # ^^^^
14. Support custom models and layers # ^^^^
15. Support gif creation # DONE
16. Legacify deep_dream_simple & deep_dream_octaved # DONE
17. Dynamic image preprocessor based on the model # DONE
18. Implement class-based dream and deepdream or instead use polymorphism for a compromise # DONE
19. Decouple Config to DreamConfig and OutputConfig
20. Yank broken versions in PyPI
21. Add option of audio in media outputs # DONE
22. Constrain package versions # DONE
23. Implement class based dream integration tests # DONE
======================================================================================================
MEDIUM:
1. Compare different pre-trained models in generation
2. Optional seeded generations # DONE
3. Remove color artifacts in generated shallow dream # DONE
4. Add option to choose layer settings
5. Simplify imports # PARTIAL
6. Remove redundancy in ImageToVideoConverter by using the Factory design pattern # UNNEEDED
7. Support single channel images via channel (C) repetition, # DONE
since pre-trained models expect RGB-Cs # ^^^^
8. Pass max_dim to get_image # UNNEEDED
9. Fix pathing support in save video # DONE
10. Ensure Dream behavior is as expected
11. Reduce verbosity of clip concatenator # DONE
12. Implement masking
13. If grayscale image is passed ensure output is grayscale as well?
14. Modularize asset (flight.wav) # UNNEEDED
======================================================================================================
LOW:
1. Optimize cicd pipeline
2. Update README # DONE
3. Implement progress bar (TQDM) in deep dream # UNNEEDED
4. Add license # DONE