diff --git a/paddle/fluid/operators/cast_op.cc b/paddle/fluid/operators/cast_op.cc
index 84660d042c7b12283fabc316d29609f5eddb825d..8d6a498dc941e44688ec8a2b49a6e080608f9b85 100644
--- a/paddle/fluid/operators/cast_op.cc
+++ b/paddle/fluid/operators/cast_op.cc
@@ -89,4 +89,5 @@ REGISTER_OP_CPU_KERNEL(cast, ops::CastOpKernel<CPU, float>,
                        ops::CastOpKernel<CPU, int>,
                        ops::CastOpKernel<CPU, int64_t>,
                        ops::CastOpKernel<CPU, bool>,
+                       ops::CastOpKernel<CPU, uint8_t>,
                        ops::CastOpKernel<CPU, paddle::platform::float16>);
diff --git a/paddle/fluid/operators/cast_op.cu b/paddle/fluid/operators/cast_op.cu
index c486c5850e25fcf4370f02cb145c244743a4cc4b..657d162878c108760585ca9bd58e2fd34bf1fef3 100644
--- a/paddle/fluid/operators/cast_op.cu
+++ b/paddle/fluid/operators/cast_op.cu
@@ -21,5 +21,5 @@ using CastOpKernel =
 
 REGISTER_OP_CUDA_KERNEL(cast, CastOpKernel<float>, CastOpKernel<double>,
                         CastOpKernel<int>, CastOpKernel<int64_t>,
-                        CastOpKernel<bool>,
+                        CastOpKernel<bool>, CastOpKernel<uint8_t>,
                         CastOpKernel<paddle::platform::float16>);
diff --git a/paddle/fluid/pybind/pybind.cc b/paddle/fluid/pybind/pybind.cc
index 50a1c07251b5bc4e7cc27de63f5457d3f94daef5..2295e83343d3340f112d4239d92d6d1163f65032 100644
--- a/paddle/fluid/pybind/pybind.cc
+++ b/paddle/fluid/pybind/pybind.cc
@@ -117,6 +117,7 @@ PYBIND11_PLUGIN(core) {
       .def("set", PyCPUTensorSetFromArray<int64_t>)
       .def("set", PyCPUTensorSetFromArray<bool>)
       .def("set", PyCPUTensorSetFromArray<uint16_t>)
+      .def("set", PyCPUTensorSetFromArray<uint8_t>)
 #ifdef PADDLE_WITH_CUDA
       .def("set", PyCUDATensorSetFromArray<float>)
       .def("set", PyCUDATensorSetFromArray<int>)
@@ -124,12 +125,14 @@ PYBIND11_PLUGIN(core) {
       .def("set", PyCUDATensorSetFromArray<int64_t>)
       .def("set", PyCUDATensorSetFromArray<bool>)
       .def("set", PyCUDATensorSetFromArray<uint16_t>)
+      .def("set", PyCUDATensorSetFromArray<uint8_t>)
       .def("set", PyCUDAPinnedTensorSetFromArray<float>)
       .def("set", PyCUDAPinnedTensorSetFromArray<int>)
       .def("set", PyCUDAPinnedTensorSetFromArray<double>)
       .def("set", PyCUDAPinnedTensorSetFromArray<int64_t>)
       .def("set", PyCUDAPinnedTensorSetFromArray<bool>)
       .def("set", PyCUDAPinnedTensorSetFromArray<uint16_t>)
+      .def("set", PyCUDAPinnedTensorSetFromArray<uint8_t>)
 #endif
       .def("shape", [](Tensor &self) { return vectorize(self.dims()); })
       .def("set_float_element", TensorSetElement<float>)
diff --git a/python/paddle/fluid/data_feeder.py b/python/paddle/fluid/data_feeder.py
index a44e078d0c13717643a6cfc6dd8bff5901ee9c97..7940dabcfb03cc9eb46f678365685a6e99bcceec 100644
--- a/python/paddle/fluid/data_feeder.py
+++ b/python/paddle/fluid/data_feeder.py
@@ -36,9 +36,11 @@ class DataToLoDTensorConverter(object):
             self.dtype = 'float64'
         elif dtype == core.VarDesc.VarType.INT32:
             self.dtype = 'int32'
+        elif dtype == core.VarDesc.VarType.UINT8:
+            self.dtype = 'uint8'
         else:
             raise ValueError("dtype must be any of [int32, float32, int64, "
-                             "float64]")
+                             "float64, uint8]")
 
         self.data = []
         self.lod = []