Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Fine-tuning BEATs on AudioSet-2M multi-label classification #6006

Draft
wants to merge 57 commits into
base: master
Choose a base branch
from
Draft
Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
57 commits
Select commit Hold shift + click to select a range
5d70791
take data normalization via config
Shikhar-S Dec 10, 2024
23914db
add linear decoder for classification tasks
Shikhar-S Dec 10, 2024
225a26f
add 5-fold data prep
Shikhar-S Dec 10, 2024
76f17d5
add config for beats fine-tuning on esc
Shikhar-S Dec 10, 2024
6b8e3d2
[pre-commit.ci] auto fixes from pre-commit.com hooks
pre-commit-ci[bot] Dec 10, 2024
468f817
cleanup
Shikhar-S Dec 10, 2024
6a47f63
add readme template
Shikhar-S Dec 10, 2024
fa5f449
Merge branch 'esc' of github.com:Shikhar-S/espnet into esc
Shikhar-S Dec 10, 2024
02bcb79
add results
Shikhar-S Dec 10, 2024
8d3bb3b
[pre-commit.ci] auto fixes from pre-commit.com hooks
pre-commit-ci[bot] Dec 10, 2024
fd36ebe
restore default slurm config
Shikhar-S Dec 10, 2024
9148500
add files for cls task
Shikhar-S Dec 10, 2024
c167bd3
fix beats test
Shikhar-S Dec 10, 2024
4e07464
Merge branch 'esc' of github.com:Shikhar-S/espnet into esc
Shikhar-S Dec 10, 2024
4756126
add more test for linear decoder
Shikhar-S Dec 11, 2024
56009b9
add roll augmentation
Shikhar-S Dec 11, 2024
6dcc151
Merge branch 'esc' into cls
Shikhar-S Dec 11, 2024
cb80e61
add quantized rolling, separate vocab concerns in linear decoder
Shikhar-S Dec 13, 2024
b897137
add quantized rolling, separate vocab concerns in linear decoder
Shikhar-S Dec 13, 2024
ca9ca1e
add dropout to linear decoder, clean up
Shikhar-S Dec 15, 2024
22a3c82
cleanup, add model links
Shikhar-S Dec 15, 2024
da315fe
clean up decoder
Shikhar-S Dec 15, 2024
e714871
add inferene and training files for classification
Shikhar-S Dec 15, 2024
0e5a217
clean up decoder
Shikhar-S Dec 15, 2024
452291f
add dropout to linear decoder, clean up
Shikhar-S Dec 15, 2024
e3881fc
cleanup, add model links
Shikhar-S Dec 15, 2024
409893d
[pre-commit.ci] auto fixes from pre-commit.com hooks
pre-commit-ci[bot] Dec 15, 2024
16728b9
fix lin dec test
Shikhar-S Dec 15, 2024
6f1a151
fix beats test, the first version had unnecessary dimension always se…
Shikhar-S Dec 15, 2024
bbd2127
Merge branch 'esc' of github.com:Shikhar-S/espnet into esc
Shikhar-S Dec 15, 2024
54b2706
merge esc
Shikhar-S Dec 15, 2024
83fa462
unstable: saving work from babel, initial template for cls task with …
Shikhar-S Dec 25, 2024
c836ede
Merge branch 'espnet:master' into cls
Shikhar-S Dec 27, 2024
77ad502
[pre-commit.ci] auto fixes from pre-commit.com hooks
pre-commit-ci[bot] Dec 27, 2024
d14d6e2
fix cls and setup shells
Shikhar-S Dec 27, 2024
78ae554
add tests for classification metrics
Shikhar-S Dec 28, 2024
bf732ce
merge remote
Shikhar-S Dec 28, 2024
ed25ca7
[pre-commit.ci] auto fixes from pre-commit.com hooks
pre-commit-ci[bot] Dec 28, 2024
7a83600
add scoring scripts
Shikhar-S Dec 29, 2024
6cbd753
add mixup augmentation
Shikhar-S Dec 29, 2024
7447d30
add script to show results
Shikhar-S Dec 29, 2024
6aed560
Merge branch 'cls' of github.com:Shikhar-S/espnet into cls
Shikhar-S Dec 29, 2024
bab37d4
[pre-commit.ci] auto fixes from pre-commit.com hooks
pre-commit-ci[bot] Dec 29, 2024
4a754da
add test for mixup, change mixup rate
Shikhar-S Jan 2, 2025
018e5bd
Merge branch 'cls' of github.com:Shikhar-S/espnet into cls
Shikhar-S Jan 2, 2025
5a3e173
[pre-commit.ci] auto fixes from pre-commit.com hooks
pre-commit-ci[bot] Jan 2, 2025
3c371a8
add recipe for AudioSet-20k
Shikhar-S Jan 2, 2025
635b3ad
Merge branch 'cls' of github.com:Shikhar-S/espnet into cls
Shikhar-S Jan 2, 2025
170b673
[pre-commit.ci] auto fixes from pre-commit.com hooks
pre-commit-ci[bot] Jan 2, 2025
2a6ec9b
tune mixup, use full training data
Shikhar-S Jan 4, 2025
79329f1
Merge branch 'cls' of github.com:Shikhar-S/espnet into cls
Shikhar-S Jan 4, 2025
b1bf3e7
add audio filtering stage
Shikhar-S Jan 4, 2025
5ad1a95
add option to truncate long audio and repeat small audio, fix ci
Shikhar-S Jan 5, 2025
9508135
add packing and hf upload stages
Shikhar-S Jan 5, 2025
9634114
fix ci
Shikhar-S Jan 5, 2025
ad72b4e
add skeleton code for AudioSet-2M finetuning
Shikhar-S Jan 5, 2025
d41e3c7
[pre-commit.ci] auto fixes from pre-commit.com hooks
pre-commit-ci[bot] Jan 5, 2025
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Prev Previous commit
Next Next commit
fix ci
  • Loading branch information
Shikhar-S committed Jan 5, 2025
commit 9634114cd3c35e230f4a9dda752e982512517653
2 changes: 1 addition & 1 deletion egs2/as20k/cls1/cmd.sh
Original file line number Diff line number Diff line change
Expand Up @@ -28,7 +28,7 @@


# Select the backend used by run.sh from "local", "stdout", "sge", "slurm", or "ssh"
cmd_backend='slurm'
cmd_backend='local'

# Local machine, without any Job scheduling system
if [ "${cmd_backend}" = local ]; then
Expand Down
3 changes: 2 additions & 1 deletion espnet2/bin/cls_inference.py
Original file line number Diff line number Diff line change
Expand Up @@ -27,7 +27,8 @@ class Classification:

Examples:
>>> import soundfile
>>> classification_model = Classification("classification_config.yml", "classification_model.pth")
>>> classification_model =
Classification("classification_config.yml", "classification_model.pth")
>>> audio, rate = soundfile.read("speech.wav")
>>> classification_model(audio)
prediction_result (int, or list of ints)
Expand Down
Empty file added espnet2/cls/decoder/__init__.py
Empty file.
Loading