diff --git a/README.md b/README.md
index 7e029c1..ca819c7 100644
--- a/README.md
+++ b/README.md
@@ -37,7 +37,7 @@ pip3 install torch torchvision torchaudio --extra-index-url https://download.pyt
 
 # install blip and clip-interrogator
 pip install -e git+https://github.com/pharmapsychotic/BLIP.git@lib#egg=blip
-pip install clip-interrogator
+pip install clip-interrogator==0.3.2
 ```
 
 You can then use it in your script
diff --git a/clip_interrogator/__init__.py b/clip_interrogator/__init__.py
index 6470b3b..057cf78 100644
--- a/clip_interrogator/__init__.py
+++ b/clip_interrogator/__init__.py
@@ -1,4 +1,4 @@
 from .clip_interrogator import Interrogator, Config
 
-__version__ = '0.3.1'
+__version__ = '0.3.2'
 __author__ = 'pharmapsychotic'
\ No newline at end of file
diff --git a/clip_interrogator/clip_interrogator.py b/clip_interrogator/clip_interrogator.py
index 67fe5b8..a978e43 100644
--- a/clip_interrogator/clip_interrogator.py
+++ b/clip_interrogator/clip_interrogator.py
@@ -9,7 +9,7 @@ import time
 import torch
 
 from dataclasses import dataclass
-from models.blip import blip_decoder
+from models.blip import blip_decoder, BLIP_Decoder
 from PIL import Image
 from torchvision import transforms
 from torchvision.transforms.functional import InterpolationMode
@@ -20,7 +20,7 @@ from typing import List
 @dataclass 
 class Config:
     # models can optionally be passed in directly
-    blip_model = None
+    blip_model: BLIP_Decoder = None
     clip_model = None
     clip_preprocess = None
 
@@ -256,8 +256,6 @@ class LabelTable():
                         if data.get('hash') == hash:
                             self.labels = data['labels']
                             self.embeds = data['embeds']
-                            if self.device == 'cpu':
-                                self.embeds = [e.astype(np.float32) for e in self.embeds]
                     except Exception as e:
                         print(f"Error loading cached table {desc}: {e}")
 
@@ -281,6 +279,9 @@ class LabelTable():
                         "hash": hash, 
                         "model": config.clip_model_name
                     }, f)
+
+        if self.device == 'cpu' or self.device == torch.device('cpu'):
+            self.embeds = [e.astype(np.float32) for e in self.embeds]
     
     def _rank(self, image_features: torch.Tensor, text_embeds: torch.Tensor, top_count: int=1) -> str:
         top_count = min(top_count, len(text_embeds))
diff --git a/run_cli.py b/run_cli.py
index 8d02f69..fdbba04 100755
--- a/run_cli.py
+++ b/run_cli.py
@@ -20,6 +20,7 @@ def inference(ci, image, mode):
 def main():
     parser = argparse.ArgumentParser()
     parser.add_argument('-c', '--clip', default='ViT-L-14/openai', help='name of CLIP model to use')
+    parser.add_argument('-d', '--device', default='auto', help='device to use (auto, cuda or cpu)')
     parser.add_argument('-f', '--folder', help='path to folder of images')
     parser.add_argument('-i', '--image', help='image file or url')
     parser.add_argument('-m', '--mode', default='best', help='best, classic, or fast')
@@ -41,9 +42,12 @@ def main():
         exit(1)
 
     # select device
-    device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
-    if not torch.cuda.is_available():
-        print("CUDA is not available, using CPU. Warning: this will be very slow!")
+    if args.device == 'auto':
+        device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
+        if not torch.cuda.is_available():
+            print("CUDA is not available, using CPU. Warning: this will be very slow!")
+    else:
+        device = torch.device(args.device)
 
     # generate a nice prompt
     config = Config(device=device, clip_model_name=args.clip)
diff --git a/setup.py b/setup.py
index 12f2d02..19525b5 100644
--- a/setup.py
+++ b/setup.py
@@ -5,7 +5,7 @@ from setuptools import setup, find_packages
 
 setup(
     name="clip-interrogator",
-    version="0.3.1",
+    version="0.3.2",
     license='MIT',
     author='pharmapsychotic',
     author_email='me@pharmapsychotic.com',