diff --git a/requirements/audio/audio_signal.txt b/requirements/audio/audio_signal.txt
index 6082a2e1..61e688f3 100644
--- a/requirements/audio/audio_signal.txt
+++ b/requirements/audio/audio_signal.txt
@@ -1,5 +1,5 @@
 hyperpyyaml
-librosa
+librosa<=0.9.2
 MinDAEC
 mir_eval>=0.7
 numpy
diff --git a/requirements/audio/audio_tts.txt b/requirements/audio/audio_tts.txt
index 998a0c2c..b1a85faf 100644
--- a/requirements/audio/audio_tts.txt
+++ b/requirements/audio/audio_tts.txt
@@ -3,7 +3,7 @@ greenlet>=1.1.2
 inflect
 jedi>=0.18.1
 kantts
-librosa
+librosa<=0.9.2
 lxml
 matplotlib
 msgpack>=1.0.4
diff --git a/requirements/multi-modal.txt b/requirements/multi-modal.txt
index ef8037b6..49b79f2c 100644
--- a/requirements/multi-modal.txt
+++ b/requirements/multi-modal.txt
@@ -1,7 +1,7 @@
 accelerate
 diffusers>=0.11.1
 ftfy>=6.0.3
-librosa
+librosa<=0.9.2
 opencv-python
 pycocoevalcap>=1.2
 pycocotools>=2.0.4