Add support for Intel GPU to Fast Neural Style example

fast_neural_style/README.md

@@ -26,8 +26,9 @@ python neural_style/neural_style.py eval --content-image </path/to/content/image
 - `--model`: saved model to be used for stylizing the image (eg: `mosaic.pth`)
 - `--output-image`: path for saving the output image.
 - `--content-scale`: factor for scaling down the content image if memory is an issue (eg: value of 2 will halve the height and width of content-image)
-- `--cuda`: set it to 1 for running on GPU, 0 for CPU.
-- `--mps`: set it to 1 for running on macOS GPU
+- `--cuda 0|1`: set it to 1 for running on GPU, 0 for CPU.
+- `--mps`: use MPS device backend.
+- `--xpu`: use XPU device backend.
 
 Train model
 
@@ -40,8 +41,9 @@ There are several command line arguments, the important ones are listed below
 - `--dataset`: path to training dataset, the path should point to a folder containing another folder with all the training images. I used COCO 2014 Training images dataset [80K/13GB] [(download)](https://cocodataset.org/#download).
 - `--style-image`: path to style-image.
 - `--save-model-dir`: path to folder where trained model will be saved.
-- `--cuda`: set it to 1 for running on GPU, 0 for CPU.
-- `--mps`: set it to 1 for running on macOS GPU
+- `--cuda 0|1`: set it to 1 for running on GPU, 0 for CPU.
+- `--mps`: use MPS device backend.
+- `--xpu`: use XPU device backend.
 
 Refer to `neural_style/neural_style.py` for other command line arguments. For training new models you might have to tune the values of `--content-weight` and `--style-weight`. The mosaic style model shown above was trained with `--content-weight 1e5` and `--style-weight 1e10`. The remaining 3 models were also trained with similar order of weight parameters with slight variation in the `--style-weight` (`5e10` or `1e11`).
 

fast_neural_style/neural_style/neural_style.py

@@ -33,9 +33,13 @@ def train(args):
         device = torch.device("cuda")
     elif args.mps:
         device = torch.device("mps")
+    elif args.xpu:
+        device = torch.device("xpu")
     else:
         device = torch.device("cpu")
 
+    print("Device to use: ", device)
+
     np.random.seed(args.seed)
     torch.manual_seed(args.seed)
 
@@ -126,6 +130,9 @@ def train(args):
 
 def stylize(args):
     device = torch.device("cuda" if args.cuda else "cpu")
+    device = torch.device("xpu" if args.xpu else "cpu")
+
+    print("Device to use: ", device)
 
     content_image = utils.load_image(args.content_image, scale=args.content_scale)
     content_transform = transforms.Compose([
@@ -219,6 +226,10 @@ def main():
                                   help="number of images after which the training loss is logged, default is 500")
     train_arg_parser.add_argument("--checkpoint-interval", type=int, default=2000,
                                   help="number of batches after which a checkpoint of the trained model will be created")
+    train_arg_parser.add_argument('--mps', action='store_true',
+                                  help='enable macOS GPU training')
+    train_arg_parser.add_argument('--xpu', action='store_true',
+                                  help='enable Intel XPU training')
 
     eval_arg_parser = subparsers.add_parser("eval", help="parser for evaluation/stylizing arguments")
     eval_arg_parser.add_argument("--content-image", type=str, required=True,
@@ -233,7 +244,11 @@ def main():
                                  help="set it to 1 for running on cuda, 0 for CPU")
     eval_arg_parser.add_argument("--export_onnx", type=str,
                                  help="export ONNX model to a given file")
-    eval_arg_parser.add_argument('--mps', action='store_true', default=False, help='enable macOS GPU training')
+    eval_arg_parser.add_argument('--mps', action='store_true',
+                                 help='enable macOS GPU evaluation')
+    eval_arg_parser.add_argument('--xpu', action='store_true',
+                                 help='enable Intel XPU evaluation')
+
 
     args = main_arg_parser.parse_args()
 
@@ -245,6 +260,8 @@ def main():
         sys.exit(1)
     if not args.mps and torch.backends.mps.is_available():
         print("WARNING: mps is available, run with --mps to enable macOS GPU")
+    if not args.xpu and torch.xpu.is_available():
+        print("WARNING: XPU is available, run with --xpu to enable Intel XPU")
 
     if args.subcommand == "train":
         check_paths(args)