Renaming dims() to sizes() (caffe2/caffe2) - 4/4

Summary: Codemod generated with clangr shard mode, 25 files per diff, for renaming dims() to sizes()

Reviewed By: ezyang

Differential Revision: D10842900

fbshipit-source-id: 8d58ed4d403fb0308a8fa286659f8e830b040bec

8 files changed, 12 insertions, 12 deletions

diff --git a/caffe2/python/pybind_state.h b/caffe2/python/pybind_state.h
index 340fee2eff..3205e0ac08 100644
--- a/caffe2/python/pybind_state.h
+++ b/caffe2/python/pybind_state.h
@@ -128,7 +128,7 @@ class TensorFetcher : public BlobFetcherBase {
         tensor.meta().name(),
         ".");
     std::vector<npy_intp> npy_dims;
-    for (const auto dim : tensor.dims()) {
+    for (const auto dim : tensor.sizes()) {
       npy_dims.push_back(dim);
     }
     result.copied = force_copy || NeedsCopy(&tensor, tensor.meta());
diff --git a/caffe2/python/pybind_state_dlpack.h b/caffe2/python/pybind_state_dlpack.h
index 6db4ae42b8..68383d8f8f 100644
--- a/caffe2/python/pybind_state_dlpack.h
+++ b/caffe2/python/pybind_state_dlpack.h
@@ -57,7 +57,7 @@ class DLPackWrapper {
     dlTensor.ctx = tensor_context;
     dlTensor.ndim = tensor->ndim();
     dlTensor.dtype = tensor_type;
-    dlTensor.shape = const_cast<int64_t*>(&(tensor->dims()[0]));
+    dlTensor.shape = const_cast<int64_t*>(&(tensor->sizes()[0]));
     dlTensor.strides = nullptr;
     dlTensor.byte_offset = 0;
 
diff --git a/caffe2/python/pybind_state_gpu.cc b/caffe2/python/pybind_state_gpu.cc
index 3893be96ff..d5a561a7dc 100644
--- a/caffe2/python/pybind_state_gpu.cc
+++ b/caffe2/python/pybind_state_gpu.cc
@@ -143,7 +143,7 @@ void addCUDAObjectMethods(py::module& m) {
           "Copy data from given DLPack tensor into this tensor.")
       .def_property_readonly(
           "_shape",
-          [](const DLPackWrapper<CUDAContext>& t) { return t.tensor->dims(); })
+          [](const DLPackWrapper<CUDAContext>& t) { return t.tensor->sizes(); })
       .def(
           "_reshape",
           [](DLPackWrapper<CUDAContext>* t, std::vector<int64_t> dims) {
diff --git a/caffe2/python/pybind_state_int8.cc b/caffe2/python/pybind_state_int8.cc
index f60e4e3cc5..5f77cf99dc 100644
--- a/caffe2/python/pybind_state_int8.cc
+++ b/caffe2/python/pybind_state_int8.cc
@@ -38,11 +38,11 @@ class Int8TensorFetcher : public BlobFetcherBase {
     const int numpy_type = CaffeToNumpyType(src.t.meta());
     CAFFE_ENFORCE(numpy_type != -1, "Int8Tensor contains unknown type data");
     std::vector<npy_intp> npy_dims;
-    for (const auto dim : src.t.dims()) {
+    for (const auto dim : src.t.sizes()) {
       npy_dims.push_back(dim);
     }
     auto data_array = pybind11::reinterpret_steal<pybind11::object>(
-        PyArray_SimpleNew(src.t.dims().size(), npy_dims.data(), numpy_type));
+        PyArray_SimpleNew(src.t.sizes().size(), npy_dims.data(), numpy_type));
     void* ptr = static_cast<void*>(
         PyArray_DATA(reinterpret_cast<PyArrayObject*>(data_array.ptr())));
     CPUContext context;
diff --git a/caffe2/queue/queue_ops.h b/caffe2/queue/queue_ops.h
index 8e924176a0..dac7caecd3 100644
--- a/caffe2/queue/queue_ops.h
+++ b/caffe2/queue/queue_ops.h
@@ -160,7 +160,7 @@ class SafeDequeueBlobsOp final : public Operator<Context> {
               size,
               " total columns");
 
-          out->Extend(in.dims()[0], kTensorGrowthPct, &context_);
+          out->Extend(in.sizes()[0], kTensorGrowthPct, &context_);
           auto* dst =
               (char*)out->raw_mutable_data() + oldSize * in.meta().itemsize();
           context_.template CopyItems<Context, Context>(
diff --git a/caffe2/queue/rebatching_queue.cc b/caffe2/queue/rebatching_queue.cc
index 6131d974c4..52b304006a 100644
--- a/caffe2/queue/rebatching_queue.cc
+++ b/caffe2/queue/rebatching_queue.cc
@@ -42,7 +42,7 @@ void concat(
       CAFFE_ENFORCE_EQ(inputZero[j].itemsize(), input.itemsize());
       CAFFE_ENFORCE_EQ(inputZero[j].ndim(), input.ndim());
       for (int k = 0; k < input.ndim(); ++k) {
-        CAFFE_ENFORCE_EQ(input.dims()[k], inputZero[j].dims()[k]);
+        CAFFE_ENFORCE_EQ(input.sizes()[k], inputZero[j].dims()[k]);
       }
 
       // Skip empty tensors
@@ -68,7 +68,7 @@ std::vector<std::vector<TensorCPU>> split(
     const std::vector<const TensorCPU*>& inputs) {
   CAFFE_ENFORCE(!inputs.empty());
 
-  const auto outputSize = inputs[0]->dims().at(0);
+  const auto outputSize = inputs[0]->sizes().at(0);
   std::vector<std::vector<TensorCPU>> outputs(outputSize);
 
   for (const auto* inputPtr : inputs) {
@@ -78,10 +78,10 @@ std::vector<std::vector<TensorCPU>> split(
     const auto innerSize = input.size_from_dim(1);
     const auto itemSize = input.meta().itemsize();
 
-    auto outputDims = input.dims().vec();
+    auto outputDims = input.sizes().vec();
     CAFFE_ENFORCE(!outputDims.empty());
     outputDims.erase(outputDims.begin());
-    CAFFE_ENFORCE_EQ(input.dims().at(0), outputSize);
+    CAFFE_ENFORCE_EQ(input.sizes().at(0), outputSize);
 
     for (int i = 0; i < outputSize; ++i) {
       outputs[i].push_back(Tensor(outputDims, CPU));
diff --git a/caffe2/sgd/adagrad_op.h b/caffe2/sgd/adagrad_op.h
index ec88a124d4..de330b2ad0 100644
--- a/caffe2/sgd/adagrad_op.h
+++ b/caffe2/sgd/adagrad_op.h
@@ -250,7 +250,7 @@ class RowWiseSparseAdagradOp final : public Operator<Context> {
 
   bool RunOnDevice() override {
     // Enforce shapes
-    CAFFE_ENFORCE_EQ(Input(PARAM).dims()[0], Input(MOMENT_1).size());
+    CAFFE_ENFORCE_EQ(Input(PARAM).sizes()[0], Input(MOMENT_1).size());
     CAFFE_ENFORCE_EQ(Input(LR).size(), 1);
     CAFFE_ENFORCE_EQ(
         Input(PARAM).size_from_dim(1),
diff --git a/caffe2/sgd/adam_op.h b/caffe2/sgd/adam_op.h
index 699ba7aa5d..0cf2f417cd 100644
--- a/caffe2/sgd/adam_op.h
+++ b/caffe2/sgd/adam_op.h
@@ -333,7 +333,7 @@ class RowWiseSparseAdamOp final : public Operator<Context> {
   bool RunOnDevice() override {
     // Enforce shapes
     CAFFE_ENFORCE_EQ(Input(PARAM).size(), Input(MOMENT_1).size());
-    CAFFE_ENFORCE_EQ(Input(PARAM).dims()[0], Input(MOMENT_2).size());
+    CAFFE_ENFORCE_EQ(Input(PARAM).sizes()[0], Input(MOMENT_2).size());
     CAFFE_ENFORCE_EQ(
         Input(PARAM).size_from_dim(1),
         Input(GRAD).size_from_dim(Input(INDICES).ndim()));