Add AudioLDM #2232
patrickvonplaten
merged
79 commits into
huggingface:main
from
audioldm
Add AudioLDM
88b77bc8
up
9b353b01
add vocoder
1a3ea27d
Merge branch 'main' into audioldm
22a315d5
start unet
1023f68b
unconditional unet
81bff992
Merge remote-tracking branch 'origin/audioldm' into audioldm
990b6220
clap, vocoder and vae
6aa4fda3
clean-up: conversion scripts
2482b42d
fix: conversion script token_type_ids
9d986c47
clean-up: pipeline docstring
004fed8e
tests: from SD
9feb6ba3
clean-up: cpu offload vocoder instead of safety checker
bf3964c5
feat: adapt tests to audioldm
f200e803
feat: add docs
dd04c2e0
clean-up: amend pipeline docstrings
1c26ca94
clean-up: make style
d32bd7f4
clean-up: make fix-copies
447013e4
fix: add doc path to toctree
08d6a1f6
clean-up: args for conversion script
9597761e
clean-up: paths to checkpoints
10c584d3
fix: use conditional unet
0f154083
clean-up: make style
d99c9e84
fix: type hints for UNet
293f2a40
sanchit-gandhi
commented
on 2023-02-21
sanchit-gandhi
commented
on 2023-02-21
clean-up: docstring for UNet
8b52493a
Merge branch 'main' into audioldm
13f6f3e9
clean-up: make style
00399217
Merge remote-tracking branch 'origin/audioldm' into audioldm
1222a638
clean-up: remove duplicate in docstring
0be07899
clean-up: make style
3f5f8630
clean-up: make fix-copies
3033ac1f
sanchit-gandhi
commented
on 2023-02-21
clean-up: move imports to start in code snippet
dd1882fa
fix: pass cross_attention_dim as a list/tuple to unet
4471f088
clean-up: make fix-copies
e81696ff
fix: update checkpoint path
b8165a1f
fix: unet cross_attention_dim in tests
1a1dc585
williamberman
commented
on 2023-02-24
williamberman
commented
on 2023-02-24
williamberman
commented
on 2023-02-24
williamberman
commented
on 2023-02-24
williamberman
commented
on 2023-02-24
williamberman
commented
on 2023-02-24
williamberman
commented
on 2023-02-24
williamberman
commented
on 2023-02-24
williamberman
commented
on 2023-02-24
williamberman
commented
on 2023-02-24
film embeddings -> class embeddings
3947e375
williamberman
commented
on 2023-02-24
williamberman
commented
on 2023-02-24
williamberman
commented
on 2023-02-24
williamberman
commented
on 2023-02-24
williamberman
commented
on 2023-02-24
williamberman
commented
on 2023-02-24
williamberman
commented
on 2023-02-24
Apply suggestions from code review
1503f757
Merge pull request #1 from williamberman/will/audioldm
cccf556d
fix: unet film embed to use existing args
074f8837
fix: unet tests to use existing args
94dc7617
fix: make style
e66476e2
fix: transformers import and version in init
9f776a28
clean-up: make style
5d6d1f8b
Revert "clean-up: make style"
8b4ea075
clean-up: make style
876f241c
clean-up: use pipeline tester mixin tests where poss
dfc1c859
clean-up: skip attn slicing test
ad809114
Merge branch 'main' into audioldm
4d2b9026
fix: add torch dtype to docs
68cc47e9
fix: remove conversion script out of src
6bc6a75f
fix: remove .detach from 1d waveform
99a33889
fix: reduce default num inf steps
ed9be202
fix: swap height/width -> audio_length_in_s
87755de2
clean-up: make style
42294e59
Merge remote-tracking branch 'origin/audioldm' into audioldm
c62a0703
fix: remove nightly tests
21d64481
fix: imports in conversion script
01f9ade8
clean-up: slim-down to two slow tests
a9faabbf
clean-up: slim-down fast tests
9f26689e
fix: batch consistent tests
7bc812dd
clean-up: make style
f0002f1c
clean-up: remove vae slicing fast test
a0a156a4
clean-up: propagate changes to doc
a01022ab
fix: increase test tol to 1e-2
460231ed
Merge branch 'main' into audioldm
9cb4426f
clean-up: finish docs
c8a74367
Merge remote-tracking branch 'origin/audioldm' into audioldm
01fbbcfe
clean-up: make style
ee672775
Merge branch 'main' into audioldm
56203905
patrickvonplaten
commented
on 2023-03-21
feat: vocoder / VAE compatibility check
d8ab1a19
feat: possibly expand / cut audio waveform
56e3fb9e
fix: pipeline call signature test
e66dfc73
Merge remote-tracking branch 'origin/audioldm' into audioldm
4d7849e6
fix: slow tests output len
7ed071a5
clean-up: make style
b90d564d
Merge branch 'main' into audioldm
ef0e8b35
Merge branch 'main' into audioldm
b0ade430
make style
ef6c8e06
Reviewers
Assignees
Labels
None yet
Milestone
No milestone
Login to write a write a comment.
Login via GitHub