diff --git a/examples/audio-classification/requirements.txt b/examples/audio-classification/requirements.txt
index a0595c9971..9ab6959db1 100644
--- a/examples/audio-classification/requirements.txt
+++ b/examples/audio-classification/requirements.txt
@@ -1,4 +1,4 @@
-datasets >= 3.0.2
+datasets >= 3.0.2, <= 3.6.0
 evaluate == 0.4.3
 numba==0.60.0
 librosa == 0.10.2.post1
diff --git a/examples/stable-diffusion/training/requirements.txt b/examples/stable-diffusion/training/requirements.txt
index b8a17c887f..74831e33f5 100644
--- a/examples/stable-diffusion/training/requirements.txt
+++ b/examples/stable-diffusion/training/requirements.txt
@@ -1,5 +1,5 @@
 compel
-datasets
+datasets<=3.6.0
 imagesize == 1.4.1
 opencv-python
 peft==0.10.0
diff --git a/examples/text-generation/requirements.txt b/examples/text-generation/requirements.txt
index ca126840c6..177b49c5d3 100644
--- a/examples/text-generation/requirements.txt
+++ b/examples/text-generation/requirements.txt
@@ -1,5 +1,5 @@
-datasets>=3.0.2
+datasets>=3.0.2, <=3.6.0
 peft == 0.11.1
 sentencepiece
 tiktoken
-blobfile
\ No newline at end of file
+blobfile
diff --git a/examples/text-generation/requirements_lm_eval.txt b/examples/text-generation/requirements_lm_eval.txt
index 9f59cceddc..13999a770b 100644
--- a/examples/text-generation/requirements_lm_eval.txt
+++ b/examples/text-generation/requirements_lm_eval.txt
@@ -1,5 +1,5 @@
 lm-eval==0.4.8
-datasets>=3.0.2
+datasets>=3.0.2, <=3.6.0
 evaluate == 0.4.3
 rouge_score == 0.1.2
 accelerate