diff --git a/src/whisper_streaming/backends.py b/src/whisper_streaming/backends.py
index 99ba762..20522ed 100644
--- a/src/whisper_streaming/backends.py
+++ b/src/whisper_streaming/backends.py
@@ -4,7 +4,7 @@ import logging
 import io
 import soundfile as sf
 import math
-
+import torch
 
 logger = logging.getLogger(__name__)
 
@@ -102,11 +102,13 @@ class FasterWhisperASR(ASRBase):
         else:
             raise ValueError("modelsize or model_dir parameter must be set")
 
-        # this worked fast and reliably on NVIDIA L40
+        device = "cuda" if torch.cuda.is_available() else "cpu"
+        compute_type = "float16" if device == "cuda" else "float32"
+
         model = WhisperModel(
             model_size_or_path,
-            device="cuda",
-            compute_type="float16",
+            device=device,
+            compute_type=compute_type,
             download_root=cache_dir,
         )