Sử dụng bộ sưu tập để sắp xếp ngăn nắp các trang
Lưu và phân loại nội dung dựa trên lựa chọn ưu tiên của bạn.
dòng chảy căng:: ôi:: BatchToSpaceND
#include <array_ops.h>
BatchToSpace cho các tensor ND loại T.
Bản tóm tắt
Hoạt động này định hình lại kích thước "lô" 0 thành kích thước M + 1
của hình dạng block_shape + [batch]
, xen kẽ các khối này trở lại lưới được xác định bởi kích thước không gian [1, ..., M]
, để thu được kết quả với cùng thứ hạng với đầu vào. Sau đó, các kích thước không gian của kết quả trung gian này được cắt tùy ý theo crops
để tạo ra đầu ra. Đây là mặt trái của SpaceToBatch. Xem bên dưới để biết mô tả chính xác.
Lập luận:
- phạm vi: Một đối tượng phạm vi
- đầu vào: ND có hình dạng
input_shape = [batch] + spatial_shape + remaining_shape
, trong đó không gian_shape có kích thước M. - block_shape: 1-D với hình dạng
[M]
, tất cả các giá trị phải >= 1. - crop: 2-D có hình dạng
[M, 2]
, tất cả các giá trị phải >= 0. crops[i] = [crop_start, crop_end]
chỉ định số lượng cần cắt từ thứ nguyên đầu vào i + 1
, tương ứng với thứ nguyên không gian i
. Yêu cầu crop_start[i] + crop_end[i] <= block_shape[i] * input_shape[i + 1]
.
Thao tác này tương đương với các bước sau:
- Định hình lại
input
để reshaped
hình dạng: [block_shape[0], ..., block_shape[M-1], batch / prod(block_shape), input_shape[1], ..., input_shape[N-1]] - Cho phép các kích thước của
reshaped
để tạo ra hình dạng permuted
[batch / prod(block_shape),input_shape[1], block_shape[0], ..., input_shape[M], block_shape[M-1],input_shape[M+1], ..., input_shape[N-1]] - Định hình
permuted
để tạo ra reshaped_permuted
[batch / prod(block_shape),input_shape[1] * block_shape[0], ..., input_shape[M] * block_shape[M-1],input_shape[M+1], .. ., input_shape[N-1]] - Cắt phần đầu và phần cuối của kích thước
[1, ..., M]
của reshaped_permuted
theo crops
để tạo ra đầu ra của hình dạng: [batch / prod(block_shape),input_shape[1] * block_shape[0] - crop[0, 0] - crop[0,1], ..., input_shape[M] * block_shape[M-1] - crop[M-1,0] - crop[M-1,1],input_shape[M+1] , ..., input_shape[N-1]]
Một số ví dụ:
(1) Đối với đầu vào sau của hình dạng [4, 1, 1, 1]
, block_shape = [2, 2]
và crops = [[0, 0], [0, 0]]
:
[[[[1]]], [[[2]]], [[[3]]], [[[4]]]]
Tensor đầu ra có dạng [1, 2, 2, 1]
và có giá trị:
x = [[[[1], [2]], [[3], [4]]]]
(2) Đối với đầu vào sau của hình dạng [4, 1, 1, 3]
, block_shape = [2, 2]
và crops = [[0, 0], [0, 0]]
:
[[[[1, 2, 3]]], [[[4, 5, 6]]], [[[7, 8, 9]]], [[[10, 11, 12]]]]
Tensor đầu ra có dạng [1, 2, 2, 3]
và có giá trị:
x = [[[[1, 2, 3], [4, 5, 6]],
[[7, 8, 9], [10, 11, 12]]]]
(3) Đối với đầu vào sau của hình dạng [4, 2, 2, 1]
, block_shape = [2, 2]
và crops = [[0, 0], [0, 0]]
:
x = [[[[1], [3]], [[9], [11]]],
[[[2], [4]], [[10], [12]]],
[[[5], [7]], [[13], [15]]],
[[[6], [8]], [[14], [16]]]]
Tensor đầu ra có dạng [1, 4, 4, 1]
và có giá trị:
x = [[[[1], [2], [3], [4]],
[[5], [6], [7], [8]],
[[9], [10], [11], [12]],
[[13], [14], [15], [16]]]]
(4) Đối với đầu vào sau của hình dạng [8, 1, 3, 1]
, block_shape = [2, 2]
và crops = [[0, 0], [2, 0]]
:
x = [[[[0], [1], [3]]], [[[0], [9], [11]]],
[[[0], [2], [4]]], [[[0], [10], [12]]],
[[[0], [5], [7]]], [[[0], [13], [15]]],
[[[0], [6], [8]]], [[[0], [14], [16]]]]
Tensor đầu ra có dạng [2, 2, 4, 1]
và có giá trị:
x = [[[[1], [2], [3], [4]],
[[5], [6], [7], [8]]],
[[[9], [10], [11], [12]],
[[13], [14], [15], [16]]]]
Trả về:
Thuộc tính công khai
Chức năng công cộng
nút
::tensorflow::Node * node() const
operator::tensorflow::Input() const
toán tử::tenorflow::Đầu ra
operator::tensorflow::Output() const
Trừ phi có lưu ý khác, nội dung của trang này được cấp phép theo Giấy phép ghi nhận tác giả 4.0 của Creative Commons và các mẫu mã lập trình được cấp phép theo Giấy phép Apache 2.0. Để biết thông tin chi tiết, vui lòng tham khảo Chính sách trang web của Google Developers. Java là nhãn hiệu đã đăng ký của Oracle và/hoặc các đơn vị liên kết với Oracle.
Cập nhật lần gần đây nhất: 2025-07-26 UTC.
[null,null,["Cập nhật lần gần đây nhất: 2025-07-26 UTC."],[],[],null,["# tensorflow::ops::BatchToSpaceND Class Reference\n\ntensorflow::ops::BatchToSpaceND\n===============================\n\n`#include \u003carray_ops.h\u003e`\n\n[BatchToSpace](/versions/r1.15/api_docs/cc/class/tensorflow/ops/batch-to-space#classtensorflow_1_1ops_1_1_batch_to_space) for N-D tensors of type T.\n\nSummary\n-------\n\nThis operation reshapes the \"batch\" dimension 0 into `M + 1` dimensions of shape `block_shape + [batch]`, interleaves these blocks back into the grid defined by the spatial dimensions `[1, ..., M]`, to obtain a result with the same rank as the input. The spatial dimensions of this intermediate result are then optionally cropped according to `crops` to produce the output. This is the reverse of SpaceToBatch. See below for a precise description.\n\nArguments:\n\n- scope: A [Scope](/versions/r1.15/api_docs/cc/class/tensorflow/scope#classtensorflow_1_1_scope) object\n- input: N-D with shape `input_shape = [batch] + spatial_shape + remaining_shape`, where spatial_shape has M dimensions.\n- block_shape: 1-D with shape `[M]`, all values must be \\\u003e= 1.\n- crops: 2-D with shape `[M, 2]`, all values must be \\\u003e= 0. `crops[i] = [crop_start, crop_end]` specifies the amount to crop from input dimension `i + 1`, which corresponds to spatial dimension `i`. It is required that `crop_start[i] + crop_end[i] \u003c= block_shape[i] * input_shape[i + 1]`.\n\n\u003cbr /\u003e\n\nThis operation is equivalent to the following steps:\n\n\n1. Reshape `input` to `reshaped` of shape: \\[block_shape\\[0\\], ..., block_shape\\[M-1\\], batch / prod(block_shape), input_shape\\[1\\], ..., input_shape\\[N-1\\]\\]\n2. Permute dimensions of `reshaped` to produce `permuted` of shape \\[batch / prod(block_shape),input_shape\\[1\\], block_shape\\[0\\], ..., input_shape\\[M\\], block_shape\\[M-1\\],input_shape\\[M+1\\], ..., input_shape\\[N-1\\]\\]\n3. Reshape `permuted` to produce `reshaped_permuted` of shape \\[batch / prod(block_shape),input_shape\\[1\\] \\* block_shape\\[0\\], ..., input_shape\\[M\\] \\* block_shape\\[M-1\\],input_shape\\[M+1\\], ..., input_shape\\[N-1\\]\\]\n4. Crop the start and end of dimensions `[1, ..., M]` of `reshaped_permuted` according to `crops` to produce the output of shape: \\[batch / prod(block_shape),input_shape\\[1\\] \\* block_shape\\[0\\] - crops\\[0,0\\] - crops\\[0,1\\], ..., input_shape\\[M\\] \\* block_shape\\[M-1\\] - crops\\[M-1,0\\] - crops\\[M-1,1\\],input_shape\\[M+1\\], ..., input_shape\\[N-1\\]\\]\n\n\u003cbr /\u003e\n\nSome examples:\n\n(1) For the following input of shape `[4, 1, 1, 1]`, `block_shape = [2, 2]`, and `crops = [[0, 0], [0, 0]]`:\n\n\n```text\n[[[[1]]], [[[2]]], [[[3]]], [[[4]]]]\n```\n\n\u003cbr /\u003e\n\nThe output tensor has shape `[1, 2, 2, 1]` and value:\n\n\n```text\nx = [[[[1], [2]], [[3], [4]]]]\n```\n\n\u003cbr /\u003e\n\n(2) For the following input of shape `[4, 1, 1, 3]`, `block_shape = [2, 2]`, and `crops = [[0, 0], [0, 0]]`:\n\n\n```text\n[[[[1, 2, 3]]], [[[4, 5, 6]]], [[[7, 8, 9]]], [[[10, 11, 12]]]]\n```\n\n\u003cbr /\u003e\n\nThe output tensor has shape `[1, 2, 2, 3]` and value:\n\n\n```text\nx = [[[[1, 2, 3], [4, 5, 6]],\n [[7, 8, 9], [10, 11, 12]]]]\n```\n\n\u003cbr /\u003e\n\n(3) For the following input of shape `[4, 2, 2, 1]`, `block_shape = [2, 2]`, and `crops = [[0, 0], [0, 0]]`:\n\n\n```text\nx = [[[[1], [3]], [[9], [11]]],\n [[[2], [4]], [[10], [12]]],\n [[[5], [7]], [[13], [15]]],\n [[[6], [8]], [[14], [16]]]]\n```\n\n\u003cbr /\u003e\n\nThe output tensor has shape `[1, 4, 4, 1]` and value:\n\n\n```text\nx = [[[[1], [2], [3], [4]],\n [[5], [6], [7], [8]],\n [[9], [10], [11], [12]],\n [[13], [14], [15], [16]]]]\n```\n\n\u003cbr /\u003e\n\n(4) For the following input of shape `[8, 1, 3, 1]`, `block_shape = [2, 2]`, and `crops = [[0, 0], [2, 0]]`:\n\n\n```text\nx = [[[[0], [1], [3]]], [[[0], [9], [11]]],\n [[[0], [2], [4]]], [[[0], [10], [12]]],\n [[[0], [5], [7]]], [[[0], [13], [15]]],\n [[[0], [6], [8]]], [[[0], [14], [16]]]]\n```\n\n\u003cbr /\u003e\n\nThe output tensor has shape `[2, 2, 4, 1]` and value:\n\n\n```text\nx = [[[[1], [2], [3], [4]],\n [[5], [6], [7], [8]]],\n [[[9], [10], [11], [12]],\n [[13], [14], [15], [16]]]]\n```\n\n\u003cbr /\u003e\n\nReturns:\n\n- [Output](/versions/r1.15/api_docs/cc/class/tensorflow/output#classtensorflow_1_1_output): The output tensor.\n\n\u003cbr /\u003e\n\n| ### Constructors and Destructors ||\n|---|---|\n| [BatchToSpaceND](#classtensorflow_1_1ops_1_1_batch_to_space_n_d_1ae9fc7cf839b67ec1692eb9dbd13dab3f)`(const ::`[tensorflow::Scope](/versions/r1.15/api_docs/cc/class/tensorflow/scope#classtensorflow_1_1_scope)` & scope, ::`[tensorflow::Input](/versions/r1.15/api_docs/cc/class/tensorflow/input#classtensorflow_1_1_input)` input, ::`[tensorflow::Input](/versions/r1.15/api_docs/cc/class/tensorflow/input#classtensorflow_1_1_input)` block_shape, ::`[tensorflow::Input](/versions/r1.15/api_docs/cc/class/tensorflow/input#classtensorflow_1_1_input)` crops)` ||\n\n| ### Public attributes ||\n|------------------------------------------------------------------------------------------------|----------------------------------------------------------------------------------------------------------|\n| [operation](#classtensorflow_1_1ops_1_1_batch_to_space_n_d_1a1e8d19aed27a8ba75041200ee25a7310) | [Operation](/versions/r1.15/api_docs/cc/class/tensorflow/operation#classtensorflow_1_1_operation) |\n| [output](#classtensorflow_1_1ops_1_1_batch_to_space_n_d_1a2f9a5258c2d37ba9ce71c6ebfe2f754d) | `::`[tensorflow::Output](/versions/r1.15/api_docs/cc/class/tensorflow/output#classtensorflow_1_1_output) |\n\n| ### Public functions ||\n|------------------------------------------------------------------------------------------------------------------------------|------------------------|\n| [node](#classtensorflow_1_1ops_1_1_batch_to_space_n_d_1a8c320b154abac62302b289161e5aa745)`() const ` | `::tensorflow::Node *` |\n| [operator::tensorflow::Input](#classtensorflow_1_1ops_1_1_batch_to_space_n_d_1a94adde19cfddf4d1109cceff401543c8)`() const ` | ` ` ` ` |\n| [operator::tensorflow::Output](#classtensorflow_1_1ops_1_1_batch_to_space_n_d_1a17e07f190557e6565111355cc159b528)`() const ` | ` ` ` ` |\n\nPublic attributes\n-----------------\n\n### operation\n\n```text\nOperation operation\n``` \n\n### output\n\n```text\n::tensorflow::Output output\n``` \n\nPublic functions\n----------------\n\n### BatchToSpaceND\n\n```gdscript\n BatchToSpaceND(\n const ::tensorflow::Scope & scope,\n ::tensorflow::Input input,\n ::tensorflow::Input block_shape,\n ::tensorflow::Input crops\n)\n``` \n\n### node\n\n```gdscript\n::tensorflow::Node * node() const \n``` \n\n### operator::tensorflow::Input\n\n```gdscript\n operator::tensorflow::Input() const \n``` \n\n### operator::tensorflow::Output\n\n```gdscript\n operator::tensorflow::Output() const \n```"]]