fix view to reshape

hrdae/models/networks/fb_autoencoder.py

@@ -116,21 +116,21 @@ def forward(
         b, n, s, h = x_1d.size()
         _, _, c, _, w = x_2d_0.size()
 
-        x_1d = x_1d.view(b * n, s, h)
-        x_2d_0 = x_2d_0.view(b * 2, c, h, w)
-        x_1d_0 = x_1d_0.view(b * 2, s, h)
+        x_1d = x_1d.reshape(b * n, s, h)
+        x_2d_0 = x_2d_0.reshape(b * 2, c, h, w)
+        x_1d_0 = x_1d_0.reshape(b * 2, s, h)
 
         # encode
         x_1d = self.encoder_1d(x_1d)
         _, c_, h_ = x_1d.size()
-        x_1d = x_1d.view(b * n, 1 * c_, h_, 1)
+        x_1d = x_1d.reshape(b * n, 1 * c_, h_, 1)
 
         x_2d_0 = self.encoder_2d(x_2d_0)
         _, _, _, w_ = x_2d_0.size()
-        x_2d_0 = x_2d_0.view(b, 2 * c_, h_, w_)
+        x_2d_0 = x_2d_0.reshape(b, 2 * c_, h_, w_)
 
         x_1d_0 = self.encoder_1d(x_1d_0)
-        x_1d_0 = x_1d_0.view(b, 2 * c_, h_, 1)
+        x_1d_0 = x_1d_0.reshape(b, 2 * c_, h_, 1)
 
         # expand
         # (b * n, 1 * c_, h_, 1) -> (b * n, 1 * c_, h_, w_)
@@ -145,7 +145,7 @@ def forward(
 
         # decode
         out = self.decoder_2d(latent)
-        out = out.view(b, n, c, h, w)
+        out = out.reshape(b, n, c, h, w)
 
         if self.activation is not None:
             out = self.activation(out)

hrdae/models/networks/hr_dae.py

@@ -244,12 +244,12 @@ def forward(
         c, cs = self.content_encoder(x_2d_0)
         m = self.motion_encoder(x_1d, x_1d_0)
         b, t, c_, h = m.size()
-        m = m.view(b * t, c_, h)
+        m = m.reshape(b * t, c_, h)
         c = c.repeat(t, 1, 1, 1)
         cs = [c_.repeat(t, 1, 1, 1) for c_ in cs]
         y = self.decoder(m, c, cs[::-1])
         _, c_, h, w = y.size()
-        y = y.view(b, t, c_, h, w)
+        y = y.reshape(b, t, c_, h, w)
         if self.activation is not None:
             y = self.activation(y)
         return y
@@ -293,12 +293,12 @@ def forward(
         c, cs = self.content_encoder(x_3d_0)
         m = self.motion_encoder(x_2d, x_2d_0)
         b, t, c_, h, w = m.size()
-        m = m.view(b * t, c_, h, w)
+        m = m.reshape(b * t, c_, h, w)
         c = c.repeat(t, 1, 1, 1, 1)
         cs = [c_.repeat(t, 1, 1, 1, 1) for c_ in cs]
         y = self.decoder(m, c, cs[::-1])
         _, c_, d, h, w = y.size()
-        y = y.view(b, t, c_, d, h, w)
+        y = y.reshape(b, t, c_, d, h, w)
         if self.activation is not None:
             y = self.activation(y)
         return y

hrdae/models/networks/modules/gru.py

@@ -25,9 +25,9 @@ def forward(
         self, x: Tensor, last_states: Tensor | None = None
     ) -> tuple[Tensor, Tensor]:
         b, t, c, h = x.size()
-        x = x.view(b, t, c * h)
+        x = x.reshape(b, t, c * h)
         y, _last_states = self.rnn(x, last_states)
-        y = y.view(b, t, self.c, h)
+        y = y.reshape(b, t, self.c, h)
         return y, _last_states
 
 
@@ -57,7 +57,7 @@ def forward(
         self, x: Tensor, last_states: Tensor | None = None
     ) -> tuple[Tensor, Tensor]:
         b, t, c, h, w = x.size()
-        x = x.view(b, t, c * h * w)
+        x = x.reshape(b, t, c * h * w)
         y, _last_states = self.rnn(x, last_states)
-        y = y.view(b, t, self.c, h, w)
+        y = y.reshape(b, t, self.c, h, w)
         return y, _last_states

hrdae/models/networks/motion_encoder.py

@@ -166,10 +166,10 @@ def forward(
         x_0: Tensor | None = None,
     ) -> Tensor:
         b, t, c, h = x.size()
-        x = x.view(b * t, c, h)
+        x = x.reshape(b * t, c, h)
         x = self.enc(x)
         _, c, h = x.size()
-        x = x.view(b, t, c, h)
+        x = x.reshape(b, t, c, h)
         if self.debug_show_dim:
             print(f"{self.__class__.__name__}", x.size())
         return x
@@ -201,10 +201,10 @@ def forward(
         x_0: Tensor | None = None,
     ) -> Tensor:
         b, t, c, d, h = x.size()
-        x = x.view(b * t, c, d, h)
+        x = x.reshape(b * t, c, d, h)
         x = self.enc(x)
         _, c, d, h = x.size()
-        x = x.view(b, t, c, d, h)
+        x = x.reshape(b, t, c, d, h)
         if self.debug_show_dim:
             print(f"{self.__class__.__name__}", x.size())
         return x

hrdae/models/networks/r_ae.py

@@ -149,11 +149,11 @@ def forward(
     ) -> Tensor:
         m = self.motion_encoder(x_1d, x_1d_0)
         b, t, c_, h = m.size()
-        m = m.view(b * t, c_, h, 1)
+        m = m.reshape(b * t, c_, h, 1)
         m = interpolate(m, size=self.upsample_size, mode="bilinear", align_corners=True)
         y = self.decoder(m)
         _, c_, h, w = y.size()
-        y = y.view(b, t, c_, h, w)
+        y = y.reshape(b, t, c_, h, w)
         if self.activation is not None:
             y = self.activation(y)
         return y

hrdae/models/networks/r_dae.py

@@ -216,11 +216,11 @@ def forward(
         c = self.content_encoder(x_2d_0)
         m = self.motion_encoder(x_1d, x_1d_0)
         b, t, c_, h = m.size()
-        m = m.view(b * t, c_, h)
+        m = m.reshape(b * t, c_, h)
         c = c.repeat(t, 1, 1, 1)
         y = self.decoder(m, c)
         _, c_, h, w = y.size()
-        y = y.view(b, t, c_, h, w)
+        y = y.reshape(b, t, c_, h, w)
         if self.activation is not None:
             y = self.activation(y)
         return y
@@ -264,7 +264,7 @@ def forward(
         c = self.content_encoder(x_3d_0)
         m = self.motion_encoder(x_2d, x_2d_0)
         b, t, c_, d, h = m.size()
-        m = m.view(b * t, c_, d, h)
+        m = m.reshape(b * t, c_, d, h)
         c = c.repeat(t, 1, 1, 1, 1)
         y = self.decoder(m, c)
         _, c_, d, h, w = y.size()