e600418560
clean
2023-05-19 00:07:04 +03:00
30a6c65626
simplify
2023-05-18 23:47:51 +03:00
45fb6dc514
clean
2023-05-18 22:10:04 +03:00
b18d6576ba
Revert "del root.initial_info stupid param"
...
This reverts commit e74d7976f8
2023-05-18 22:08:55 +03:00
e74d7976f8
del root.initial_info stupid param
2023-05-18 22:07:08 +03:00
7184abd917
clean
2023-05-18 22:04:07 +03:00
005ff1d2a9
clean
2023-05-18 21:54:53 +03:00
28fe31ce19
claen
2023-05-18 21:42:42 +03:00
badd32f79e
Update args.py
2023-05-18 21:24:43 +03:00
5c3cb5a9cd
clean
2023-05-18 21:21:39 +03:00
703bbfc47c
clean code
2023-05-18 21:16:42 +03:00
48b0ebf097
fix preview of generated videos
2023-05-18 21:08:29 +03:00
43c50b36f9
fix git commit not showing in UI
2023-05-18 15:05:56 +03:00
2173bebd77
fix image not showing on right side of ui after complete gen
2023-05-18 14:25:20 +03:00
eb736f03ff
Update ui_right.py
2023-05-18 14:08:23 +03:00
dbdd1f3bb5
Update settings.py
2023-05-18 14:05:57 +03:00
ce48f78767
Update settings.py
2023-05-18 14:05:38 +03:00
53867ac85a
shorten some long code
2023-05-18 14:04:46 +03:00
ee2bb0c7f2
Update render.py
2023-05-18 13:54:00 +03:00
5666ae8cf3
Update ui_settings.py
2023-05-18 13:53:10 +03:00
3c9007f526
clean
2023-05-18 02:53:07 +03:00
141c67e94b
del debug mode for now
2023-05-18 00:58:23 +03:00
4e2f01fb18
part 2
2023-05-18 00:54:36 +03:00
dd62962fb2
law and order
2023-05-18 00:52:12 +03:00
9b42364159
separate deforum tqdm from settings.py
2023-05-18 00:51:06 +03:00
1d25281935
Update settings.py
2023-05-18 00:47:30 +03:00
f115befa13
Update settings.py
2023-05-18 00:44:50 +03:00
abdccec7ff
Update render_modes.py
2023-05-18 00:41:44 +03:00
3bfe4bf8c7
Update render_modes.py
2023-05-18 00:41:22 +03:00
ea237c9f78
Update webui_sd_pipeline.py
2023-05-18 00:39:43 +03:00
e681702281
clean
2023-05-18 00:38:56 +03:00
82a4a7647a
clean
2023-05-18 00:33:23 +03:00
b208c28433
Update args.py
2023-05-18 00:29:55 +03:00
ca90a74511
del some not needed code
2023-05-18 00:27:55 +03:00
2a374145f6
del unused code
2023-05-18 00:24:16 +03:00
1c3ca807ef
clean
2023-05-18 00:19:33 +03:00
f6a7e424a6
Update defaults.py
2023-05-17 22:32:18 +03:00
7b44682a48
make more sense import wise
2023-05-17 22:22:57 +03:00
1e90fd09ef
del unused code
2023-05-17 22:20:11 +03:00
4278dc98d1
CLEAN
2023-05-17 22:13:44 +03:00
5afb32cfa1
clean
2023-05-17 22:05:29 +03:00
415deef52c
clean/simplify return of run_deforum
2023-05-17 22:02:46 +03:00
3d4c9372eb
simplify coee by not using html_log (never actually used it)
2023-05-17 21:57:42 +03:00
e7b38ec005
del not needed plaintext_to_html func
2023-05-17 21:55:12 +03:00
6be2f3e7a4
del unused line
2023-05-17 21:52:51 +03:00
95f0be28e8
minimal cleaning
2023-05-17 21:50:39 +03:00
e95303fc7d
add missing if not then false to opts data get
2023-05-17 21:44:00 +03:00
bf524f1205
del unused import
2023-05-17 21:41:48 +03:00
4bef3e9486
align verrsion to 2.4 (later on need to bump further)
2023-05-17 21:40:31 +03:00
78fa3106d4
add clarification comment
2023-05-17 21:25:54 +03:00
9c06564837
add missing fn inputs outputs for gradio to make it aligned with rest of code
2023-05-17 21:22:09 +03:00
982e86feab
CLEANING
2023-05-17 21:16:08 +03:00
6a8cb2b652
deduplicate some sys/paths extend code
2023-05-17 21:03:11 +03:00
185ebac804
remove stupid p params from where they were never used
2023-05-17 15:14:40 +03:00
b0b3f52a55
Update render.py
2023-05-17 15:12:29 +03:00
ee024d4bfe
Update render.py
2023-05-17 15:12:11 +03:00
c52fa4567f
CLEAN SOME CODE
2023-05-17 15:02:54 +03:00
25843a12b3
del save_sample_per_step param
2023-05-17 14:58:42 +03:00
1acd632f48
del unused n_batch internal param
2023-05-17 14:57:12 +03:00
8979a47ee1
del unused save_settings param
...
we never let the users the option to not save a .txt file, and will never do as it's prune to many support issues and crying
2023-05-17 14:50:13 +03:00
e0eafc1e91
del unused line
2023-05-17 14:46:51 +03:00
19f5354d2b
renaming to make more sense
2023-05-17 14:45:45 +03:00
f783821d8a
C1
2023-05-17 14:39:16 +03:00
7e699be8a5
Merge branch 'enable-prompt-subtitles' of https://github.com/rewbs/deforum-for-automatic1111-webui into pr/709
2023-05-16 03:38:58 +03:00
b24b2e4095
Merge remote-tracking branch 'upstream/automatic1111-webui' into pr/709
2023-05-16 03:38:57 +03:00
d10d0e45f7
Merge branch 'enable-prompt-subtitles' of https://github.com/rewbs/deforum-for-automatic1111-webui into enable-prompt-subtitles
2023-05-16 10:26:18 +10:00
4c08b09bac
Include subtitles for RIFE.
2023-05-16 10:25:32 +10:00
96452ee39c
deduplicate CN code (use more code from CN external api)
2023-05-15 23:44:51 +03:00
1091a3e3a6
fix CN not working in Batch Mode (thanks AndyXR!)
2023-05-15 21:57:28 +03:00
396dc5e9b2
hide cn advanced row when cn is not enabled
2023-05-15 21:50:14 +03:00
fe381cdf2b
Update deforum_controlnet_gradio.py
2023-05-15 17:35:44 +03:00
24c8d22164
call the soft-embedded subs Deforum Data instead of Track 1
2023-05-15 16:48:39 +03:00
fe76ee994a
Merge branch 'enable-prompt-subtitles' of https://github.com/rewbs/deforum-for-automatic1111-webui into pr/709
2023-05-15 16:36:08 +03:00
9912465870
Merge remote-tracking branch 'upstream/automatic1111-webui' into pr/709
2023-05-15 16:36:07 +03:00
80f5e99f52
Update deprecation_utils.py
2023-05-15 16:34:44 +03:00
16ed1f8796
Improve console output. Pass srt path in explicitly so subtitles are successfully added on upscaling and frame interpolation.
2023-05-15 16:02:01 +10:00
dac0121fc9
add switch to deprecated grayish colormatch
2023-05-15 01:23:39 +03:00
b280841d1e
fix frames grayness
2023-05-15 00:50:04 +03:00
8db020c6f7
Merge remote-tracking branch 'upstream/automatic1111-webui' into pr/709
2023-05-14 16:50:11 +03:00
f6cd9fb2b2
Align with CN 14-05
2023-05-14 16:43:09 +03:00
e2ca6c067a
Merge remote-tracking branch 'upstream/automatic1111-webui' into pr/709
2023-05-12 19:12:40 +03:00
1856db88b2
Makes optical flow cadence work with hybrid motion ( #708 )
...
* Revert "Merge pull request #422 from hithereai/sampler_sch_fixes"
This reverts commit 09c99d10ed97ab68b13b0826d76e2d
2023-05-12 19:11:26 +03:00
2617b8b0e8
add default values to opts.data.get (critical!)
2023-05-12 18:59:23 +03:00
5955abb18d
Better output formatting (including error cases). Include whether current frame is a cadence frame or not in the subtitles.
2023-05-12 16:35:40 +10:00
ab05c05850
Allow prompts as subtitles. Optionally embed subtitles into mp4. Fix subtitles with Parseq: use getattr() instead of __dict__ to ensure fallback to non-Parseq controlled params works correctly.
2023-05-12 11:45:33 +10:00
c3f5d8ec56
Revert "add torch2 and xformers opts to midas"
...
This reverts commit 9e1393261c
2023-05-11 17:12:02 +03:00
55298f23e5
fix 3d mode not working with depth warp disabled
2023-05-11 11:13:23 +03:00
8b2a87b6e2
Merge pull request #696 from Phobos97/automatic1111-webui
...
bugfix for wrong mask being used
2023-05-10 20:06:46 +03:00
69e138b7f9
dont run FILM in fp16 if --no-half is used as cmd arg
2023-05-10 19:26:53 +03:00
73ad918c78
unload FILM model from vram when done interpolating +
...
make model loading make more sense, a bit faster now too
2023-05-10 19:20:13 +03:00
713af968d6
set min thresh to 32
2023-05-10 13:22:52 +03:00
c19b4e92e2
use bicubic interpolation for masks
...
instead of nearest by default to reduce pixelation
2023-05-10 13:10:02 +03:00
17df392c25
support word-based-masking for any dimensions
2023-05-10 01:29:39 +02:00
01497b7b43
Merge branch 'automatic1111-webui' into pr/696
2023-05-09 23:10:26 +03:00
9e1393261c
add torch2 and xformers opts to midas
...
3d mode goes brrr....
2023-05-09 22:04:13 +03:00
a678262ec0
remove old commented line
2023-05-09 18:55:47 +03:00
8f1405ce6e
make CN's weight, guidance start/stop whole 1 liner each
...
and add the name schedule to the actual ui param name
2023-05-09 18:54:05 +03:00
a050498053
add info to p2p cfg sch ui param
2023-05-09 18:48:25 +03:00
ccffd7435c
make run tab a lil bit tidier
2023-05-09 18:47:16 +03:00
48144bb9c1
bugfix for wrong mask being used
2023-05-09 00:59:49 +02:00
f39acbf486
fixed frames for 0-based counting in inputframes
...
also removed the stupid switching behavior for max frames when using a mask??? wtf
2023-05-08 13:41:13 -07:00
05d044f394
Merge remote-tracking branch 'upstream/automatic1111-webui' into Frame-Alignment-RBN
2023-05-08 23:36:11 +03:00
ea2d899851
fix
2023-05-08 22:27:43 +03:00
1dfab22195
auto hide/show ddim and ancestral schedule related params in ui
2023-05-08 22:23:52 +03:00
8bfd6b80e7
use frame_idx instead of frame_idx+1 in masks.py
2023-05-08 21:45:22 +03:00
c360360b51
add ancestral_eta_scheduling
...
todo: auto-show/hide in ui
2023-05-08 03:28:20 +03:00
6095dd757c
replace ddim_eta with ddim_eta_schedule
...
turns out ddim_eta is a global general setting param now, not passed directly as a p.ddim_eta any more. So i made it a schedule.
Todo: only show in ui if ddim is selected.
2023-05-08 02:57:45 +03:00
d1f860e1ce
remove mp4_path param (not in use)
2023-05-08 01:33:55 +03:00
f72ee00ae6
add new repo name to ext finder func
2023-05-08 01:31:13 +03:00
f4bc6cd176
remove unused import
2023-05-08 00:21:21 +03:00
eea43dfee7
fix frame count error at last frame just before stitching video
2023-05-08 00:20:15 +03:00
38de62d646
shorten func
2023-05-07 13:27:53 +03:00
4b6106b357
fix pers_flip shown in ui if enabled before switching to interp or vid input mode
2023-05-07 13:26:10 +03:00
89221eb639
Let CN in LP mode use init_img if use_init=True on first frame
2023-05-07 13:05:00 +03:00
d82724452f
move depth warp to its own tab, change persp_flip from accord to no accord
2023-05-07 02:43:50 +03:00
02b2f32b35
del unused code
2023-05-07 00:20:32 +03:00
b5bdc5b084
add clarifying comment on some changes with do_not_save_samples param
2023-05-07 00:00:40 +03:00
1e31101d41
fix typo
2023-05-06 23:51:43 +03:00
004be005a6
remove output_format param from code
2023-05-06 23:51:08 +03:00
79610c4f44
remove filename_format param from code
2023-05-06 23:42:29 +03:00
6807cebafa
remove display_samples param from code
2023-05-06 23:41:15 +03:00
149b9bb8f9
remove path_name_modifier and render_steps params from code
2023-05-06 23:34:02 +03:00
6d7bdea73d
delete save_samples function and param from code
2023-05-06 23:27:44 +03:00
a68ac64f0e
Delete __init__.py
...
we don't need it
2023-05-06 23:24:18 +03:00
3e8dd9fbe8
remove from_img2img_instead_of_link param from code
2023-05-06 22:51:07 +03:00
344108a936
remove show_sample_per_step param from code
2023-05-06 22:50:03 +03:00
5010c99470
remove init_c and init_latent params
2023-05-06 22:48:22 +03:00
d55a84b4c6
remove c and f params from code
2023-05-06 22:45:07 +03:00
7a1d1d6fdb
don't save outdir param to setting files
...
we don't use it as it is, but I didn't want to remove it from the code at that stage.
2023-05-06 22:35:30 +03:00
fde93501f4
auto-hide input vid/img fields if loopback is enabled
...
and fix single pic input
2023-05-06 20:31:19 +03:00
fde7553728
fix wrong alignment of frame numbers
2023-05-06 20:22:31 +03:00
c30582ff2e
add loopback mode to deforum's controlnet
2023-05-06 19:44:34 +03:00
b5055fe7d0
fix frame_idx not being passed into prepare_prompt
...
Pointed out in https://github.com/deforum-art/sd-webui-deforum/pull/610
Co-Authored-By: ali42 <7451075+ali42@users.noreply.github.com>
2023-05-06 15:34:30 +03:00
5d346c259a
Merge remote-tracking branch 'upstream/automatic1111-webui' into tests0505
2023-05-05 19:28:39 +03:00
24afbf815a
fix saving setting file with quotes around it (as loading already works)
2023-05-05 16:50:43 +03:00
e3eba88c12
support quotes around file name in load setting file box
2023-05-05 16:31:46 +03:00
707e628df4
final py3d fix
2023-05-05 13:30:30 +03:00
589719e5cb
fix clipseg import, all should work now
2023-05-05 13:27:51 +03:00
17d222ee4c
fix py3d import error (more commits to come)
2023-05-05 13:18:36 +03:00
852ed27753
print depth tensor min max in dev mode
2023-05-05 03:28:41 +03:00
879a7b8433
Merge remote-tracking branch 'upstream/automatic1111-webui' into pr/664
2023-05-04 23:23:53 +03:00
39a40e915f
temporarily remove midas3.1beitlarge
...
since it causes some problems with extreme 3d movement and depth prediction. will return when it's fixed
2023-05-04 23:20:56 +03:00
03f54a3768
FIXED optical flow cadence
...
Properly normalized the optical flow field before warping and after warping based on width and height. Now, because the range of the values are between -1 and 1 (usually much smaller), the flow doesn't get corrupted by the grid_sample for 3D or the warpPerspective for 2D anymore. So, I was able to remove all workarounds and just fix the abs to rel and rel to abs functions.
2023-05-04 12:37:34 -07:00
eeb3003112
Merge remote-tracking branch 'upstream/automatic1111-webui' into DepthV2Tests
2023-05-04 13:16:38 +03:00
92985dc197
clean up
2023-05-04 13:12:35 +03:00
39ed1512ee
feature: add controlnet guidance start/ end schedules
2023-05-04 03:50:03 +03:00
44076c6c2d
bugfix
2023-05-04 00:59:27 +03:00
ad048ff0f3
Add CN sch for guidance start/end - initial commit untested
2023-05-04 00:57:43 +03:00
44d5cafe55
fix model reload after keep in vram set to false (but the model is still in vram)
2023-05-04 00:50:02 +03:00
95919a5fac
update default depth alg in depth.py, had old value in there
2023-05-03 23:58:58 +03:00
c4d69cc13e
set default 3d depth alg to 'Midas-3-Hybrid'
2023-05-03 23:33:16 +03:00
fb2a97057e
shorten and align params order in depth.initialize func
2023-05-03 23:31:30 +03:00
901ff15488
print depth normalization data to cli only in Dev mode
2023-05-03 23:17:24 +03:00
eb71241d72
split depth initialize two 2 funcs + better leres cli print
2023-05-03 22:57:10 +03:00
4b9548fc68
bump version+
...
and credits in info accord
2023-05-03 21:42:33 +03:00
a1de5338bb
add dynamic msg to gradio ui about leres license
2023-05-03 19:11:11 +03:00
530bb80224
update that cli msg again
2023-05-03 18:49:49 +03:00
096729cbea
update adabins error cli msg
2023-05-03 18:49:19 +03:00
692aac9ee6
auto-hide midas/zoe weight param when not in legacy 3d mode(l)
2023-05-03 15:58:46 +03:00
a3c48cabad
add missing import
2023-05-03 15:54:19 +03:00
b8e78908f8
rename legacy modes to have an (old) at their end
2023-05-03 15:30:41 +03:00
e4943b1a51
add comment on midas models in anim.py
2023-05-03 14:31:12 +03:00
26ea16b0ce
updated transform_image_3d_new
...
removed excess options, removed dev code, cleaned up code, ready for production.
- removed depth tensor autocontrast options
- made equalization a part of the normalization process (not an optional arg now)
- removed my mechanism for 0-based normalization. Decided it wasn't any better. Maybe worse... mostly similar.
- cracked open a bottle of champagne on this function's hull
also added a file for the first steps of making an auto-navigation module to be used in the future inside the transform_image_3d function! It generates a rotation matrix based on greatest or least depth in the tensor, instead of using animation keys.
2023-05-03 04:15:54 -07:00
772d4392d1
Merge remote-tracking branch 'upstream/automatic1111-webui' into DepthV2Tests
2023-05-03 14:14:06 +03:00
2091e239b0
feature: control net scheduling of weights
2023-05-03 13:30:12 +03:00
b6c631e6c7
add legacy zoe+adabins mode
2023-05-03 06:53:08 +03:00
9279d8c0f6
print to cli what midas model was actually loaded
2023-05-03 03:49:03 +03:00
95ac505f02
add midas beit large 3.1 version too
2023-05-03 03:46:12 +03:00
2b81316a10
Depth V2 - RBN changes included
2023-05-03 02:38:24 +03:00
ed35a9c946
tidying up - deduplicate some download-model code
2023-05-02 19:43:36 +03:00
6a00408535
remove old adabins code and logic
2023-05-02 19:14:37 +03:00
24c5b79abe
Merge remote-tracking branch 'upstream/automatic1111-webui' into Depth-Revise-V1
2023-05-02 18:29:36 +03:00
87f78d54bd
tiny CN changes to align with their latest changes
2023-05-02 18:28:43 +03:00
28969fa4ce
add auto-download of the LeRes model (for future use)
2023-05-02 18:12:20 +03:00
d38260b13f
add future medvram support to LeRes depth
2023-05-02 12:00:39 +03:00
33e37a3591
more cleaning
2023-05-02 11:52:08 +03:00
0348eb683c
some cleanups
2023-05-02 11:45:52 +03:00
876e594068
reflect change of use_zoe_depth --> depth_algorithm in deprc utils
2023-05-02 11:42:21 +03:00
8bb3a1b7d4
cleanup of not-needed code
2023-05-02 11:06:40 +03:00
3f0d474b28
return kabas changes to prompt and fixed seed
2023-05-02 10:44:21 +03:00
cbdfa5dc3c
Depth Revise Part 1
2023-05-02 04:51:25 +03:00
4b51be881c
add nsfw, nude to default negative prompts
2023-05-01 17:34:36 +03:00
59b8f0da03
Update consistency_check.py
2023-05-01 03:12:44 +03:00
6dd1adc3e4
add comment about usage
2023-05-01 03:03:45 +03:00
40d26a9705
remove commented code as it can be found in the link inside the file
2023-05-01 03:02:48 +03:00
0e9176f0c1
take gradio's ui arg values from da. instead of setting them manually
2023-04-30 21:29:25 +03:00
f976c8d2fd
reorder imports at hybrid_video.py
2023-04-30 21:26:56 +03:00
dde01b3915
remove commented unused imports from render.py
2023-04-30 21:24:49 +03:00
7041917c2e
deduplicate - use a loop gor gradio.change stuff
2023-04-30 21:19:22 +03:00
eec4c7a348
quick fix for progress bar during cadence
...
makes progress advance smoothly during cadence rather than skipping ahead suddenly after cadence completes
left line commented for state.current_image, since I'm not sure what that variable actually does. It doesn't update the preview when I try it, even when turning up the speed of preview updates in auto1111. But, if we decide to make a preview mechanism or need that state var for cadence, it can just be uncommented. It matches the format of the one in the main non-cadence loop.
2023-04-29 23:31:19 -07:00
352a8fdb07
Consistency flow masks & fixed Frames to Video
...
added consistency flow masks
- there is now an option to use flow consistency masks and an attached option for consistency mask blur, defaulted to 2.
- if you save extra frames, it also save consistency masks now
- you can see the effect on the flow in the flow outputs as well
- it doesn't work as great with cadence because you see afterimages, but if you up the blur it can be a little better.
fixed Frames to Video
- made the ffmpeg routine that Frames to Video uses able to take image files other than png. If png, it includes the -vcodec png as normal. But, if anything else it includes vcodec libx264, which works for jpgs. (jpgs don't work if using vcodec png, so I made it switchable). I haven't tested it with other filetypes, but I bet it works with others too. The png vcodec was specific to png.
- also added two more lines of instruction for how to use the file string.
- I also changed a few ransac functions for future use. They work as normal, but now have a switching behavior if passed depth. But, I'm not passing depth to them for now.
- a few minor code var name edits in hybrid video to align code better (mostly changed matrices to M, as is often convention
- commented a bunch of unused imports in render.py
- I'll leave it up to someone else to delete them after it's verified that everything works fine with them commented. I searched and didn't find them in that file. VSCode showed them as gray automatically, but I also verified.
2023-04-29 22:37:49 -07:00
1cd38479b4
support input path strings with quotes around them
...
to support windows built-in "copy-path" option
2023-04-29 20:33:23 +03:00
31aacf342d
don't error when fps field is getting set to '' (nothing)
2023-04-29 19:55:27 +03:00
adcb61eff0
fix vid2frames not working with extract_nth_frame properly
2023-04-29 19:03:58 +03:00
69b607e44a
download adabins from github instead of g-drive
2023-04-29 09:52:37 +03:00
cf55274f33
Update depth.py add missing comma
2023-04-28 13:06:05 +03:00
e4c454e1e9
download adabins directly from google-drive
2023-04-28 12:58:13 +03:00
92db5b0fd7
return mask to UI, but it's broken in CN side of things (probably)
2023-04-27 21:55:52 +03:00
2fdaa80916
update cn related stuff
2023-04-26 14:43:14 +03:00
85d247af11
use better resnet50 model in human masking
2023-04-25 23:46:22 +03:00
hithereai
rewbs
kabachuha
reallybigname
Phobos
reallybigname