Linear8bitLt: support device movement after forward() (#1769)

bitsandbytes/nn/modules.py

@@ -679,19 +679,27 @@ class Int8Params(torch.nn.Parameter):
     def to(self, *args, **kwargs):
         device, dtype, non_blocking, convert_to_format = torch._C._nn._parse_to(*args, **kwargs)
 
-        if device is not None and device.type != "meta" and self.data.device.type == "cpu":
-            if device.type != "cpu" or self.data.dtype != torch.int8:
-                return self._quantize(device)
-            elif self.data.dtype == torch.int8 and device.type == "cpu":
-                self.CB = self.data
+        is_quantized = self.data.dtype == torch.int8
 
+        if not is_quantized and device is not None and device.type != "meta" and self.data.device.type == "cpu":
+            # We're moving from a CPU device to a non-meta device.
+            # In this circumstance, we want to quantize if we haven't already.
+            return self._quantize(device)
+
+        # Create a new parameter on the target device.
         new_param = Int8Params(
             super().to(device=device, dtype=dtype, non_blocking=non_blocking),
             requires_grad=self.requires_grad,
             has_fp16_weights=self.has_fp16_weights,
         )
-        new_param.CB = self.CB
-        new_param.SCB = self.SCB
+
+        # If we had already quantized, move the statistics appropriately.
+        if is_quantized and device is not None:
+            if self.CB is not None:
+                new_param.CB = new_param.data
+
+            if self.SCB is not None:
+                new_param.SCB = self.SCB.to(device)
 
         return new_param
 
@@ -1037,6 +1045,21 @@ class Linear8bitLt(nn.Linear):
         self.weight.CB = None
         self.weight.SCB = None
 
+    def to(self, *args, **kwargs):
+        # Call the parent to() method to handle standard parameter/buffer movement
+        result = super().to(*args, **kwargs)
+
+        device, dtype, non_blocking, convert_to_format = torch._C._nn._parse_to(*args, **kwargs)
+
+        # Handle state tensors if needed.
+        if device is not None:
+            if result.state.CB is not None:
+                result.state.CB = result.state.CB.to(device)
+            if result.state.SCB is not None:
+                result.state.SCB = result.state.SCB.to(device)
+
+        return result
+
     def forward(self, x: torch.Tensor):
         self.state.is_training = self.training
         if self.weight.CB is not None:

tests/test_linear8bitlt.py

@@ -293,3 +293,41 @@ def test_linear8bitlt_torch_compile(device, threshold, bias, fullgraph, mode):
             grad_compiled = x.grad.clone()
 
             torch.testing.assert_close(grad_compiled, grad_ref)
+
+
+@pytest.mark.parametrize("device", get_available_devices(no_cpu=True))
+@pytest.mark.skipif(not get_available_devices(no_cpu=True), reason="No accelerator device")
+def test_linear8bitlt_device_movement(device):
+    """Test moving a Linear8bitLt layer between CPU and an accelerator device."""
+
+    # Create a Linear8bitLt layer on CPU
+    layer = bnb.nn.Linear8bitLt(32, 128, bias=False, has_fp16_weights=False)
+    torch.nn.init.xavier_uniform_(layer.weight)
+
+    # Create a sample input.
+    x = torch.randn(4, 32, dtype=torch.float16, device="cpu")
+
+    # Move to the device. This should quantize the weights.
+    layer = layer.to(device)
+    assert layer.weight.data.dtype == torch.int8
+
+    # Call the layer on the accelerator device.
+    out_accelerator = layer(x.to(device))
+
+    # Move back to CPU and call again.
+    layer = layer.to("cpu")
+    out_cpu = layer(x)
+
+    # Move back to the accelerator device and call again.
+    layer = layer.to(device)
+    out_accelerator_2 = layer(x.to(device))
+
+    # Move back to the CPU and call one last time.
+    layer = layer.to("cpu")
+    out_cpu_2 = layer(x)
+
+    # CPU outputs should match both times.
+    torch.testing.assert_close(out_cpu_2, out_cpu, rtol=1e-8, atol=1e-8)
+
+    # Accelerator outputs should match both times.
+    torch.testing.assert_close(out_accelerator_2, out_accelerator, rtol=1e-8, atol=1e-8)