Merge pull request #59 from lsisoft/fix_precision_16

Add dtype in axis_pos creation to allow precision=16 training

perceiver_pytorch/perceiver_pytorch.py

@@ -221,13 +221,13 @@ def forward(
         mask = None,
         return_embeddings = False
     ):
-        b, *axis, _, device = *data.shape, data.device
+        b, *axis, _, device, dtype = *data.shape, data.device, data.dtype
         assert len(axis) == self.input_axis, 'input data must have the right number of axis'
 
         if self.fourier_encode_data:
             # calculate fourier encoded positions in the range of [-1, 1], for all axis
 
-            axis_pos = list(map(lambda size: torch.linspace(-1., 1., steps = size, device = device), axis))
+            axis_pos = list(map(lambda size: torch.linspace(-1., 1., steps=size, device=device, dtype=dtype), axis))
             pos = torch.stack(torch.meshgrid(*axis_pos, indexing = 'ij'), dim = -1)
             enc_pos = fourier_encode(pos, self.max_freq, self.num_freq_bands)
             enc_pos = rearrange(enc_pos, '... n d -> ... (n d)')