train with multi-gpu half test bug fix #99

5 years ago · d1e5716610
parent 1dfc28527f
commit d1e5716610
1 changed files with 1 additions and 1 deletions
--- a/test.py
+++ b/test.py
@ -46,7 +46,7 @@ def test(data,
    else:  # called by train.py
        training = True
        device = next(model.parameters()).device  # get model device
-        half = device.type != 'cpu'  # half precision only supported on CUDA
+        half = device.type != 'cpu' and torch.cuda.device_count() == 1  # half precision only supported on single-GPU
        if half:
            model.half()  # to FP16