Paddle

Commit Graph

Author	SHA1	Message	Date
qijun	5017b15468	refactor tensor mutable_data	8 years ago
qijun	9e3a9eb233	merge baidu/develop	8 years ago
qijun	d649dbf442	implement add_op kernel	8 years ago
fengjiayi	57a22db3fa	update PADDLE_ENFORCE message	8 years ago
fengjiayi	34beec0f7a	update tensor.h	8 years ago
fengjiayi	8594d5c314	change int numel_ to size_t numel	8 years ago
fengjiayi	1f97388a34	fix several compile error	8 years ago
fengjiayi	dcfcf6872a	Refactor Tensor::CopyFrom() 1. Add template T which indicates data type to `CopyFrom()`, `Slice()` and `ShareData()` functions. This makes `CopyData()` code much clearer. 2. Add `set_dim()`. 3. `product(DDim)` transforms `DDim` to `vector<int>` first and then calculate its product. That might be quite slow. For `product(dims_)` is frequently used in Tensor, we add a mumber variable `numel_` as a cache of the product result. TODO: refactor `product()` to make it more efficient. 4. Unable Tensor::operator= 5. Remove the limit of POD type, because `float16` and `int8` are not POD type.	8 years ago
fengjiayi	a1dc4311a3	Refactor `Tensor::CopyFrom()`	8 years ago
qijun	bac1426d47	add_op kernel implementation	8 years ago
fengjiayi	69d99d481d	Add Tensor::CopyFrom and Tensor::mutable_data(Place place) 1. Add `Tensor::CopyFrom`. Current version can only support CPU memory copy. The support of GPU will be provided later by `paddle::memory`. The current implementation of `Tensor::CopyFrom` is a little inefficient: Every time `CopyFrom` is called, tensor will re-allocate its memory. However, if we try to check and reuse `placeholder_`, we have to provide a template parameter for `CopyFrom` to indicate the data type. It seems strange for a simple copy function. 2. Add `Tensor::mutable_data(Place place)`, which directly use member variable `dims_` as its dim parameter. This interface is required by `Op::InferShape`.	8 years ago
qijun	37bcc0860a	merge baidu/develop	8 years ago
fengjiayi	2dccab87d4	fix some syntax problems	8 years ago
fengjiayi	0665dc9755	add more test	8 years ago
fengjiayi	ca39515e24	Add several interfaces for Tensor class 1. Add member variable 'DDim dims_' and a getter function 'dims()'. 'dims' is supposed to hold tensor's shape during Op::InferShape. 2. Remove 'mutable_data' which use default Place. User must specify a explicit Place when call 'mutable_data'. 3. A PlaceHolder may be shared by more than one tensor, and some of them may be the others' slices. So we add a new member variable 'offset_' for Tensor, which is used to show the byte offset between PlaceHolder::ptr_ and where tensor's data really begins. 4. Add functions 'ShareDataFrom' and 'Slice' for Tensor. TODO: Tensor needs a 'CopyFrom' function.	8 years ago
qijun	a5eb1d8fab	fix build error	8 years ago
qijun	958511160b	add simple add_op_functor	8 years ago
qijun	1a968b4f64	init	8 years ago
fengjiayi	e48e21da2b	remove unnecessary include	8 years ago
fengjiayi	2d1f95de87	fix a compile error	8 years ago
fengjiayi	d054a5eef8	re-submit	8 years ago
fengjiayi	e146fe836b	fix compile errors and add assert test	8 years ago
fengjiayi	9f408dfb1b	fix some compile error	8 years ago
fengjiayi	696ba1d2e1	init tensor_test.cc	8 years ago
Yi Wang	c263c21f7e	Update copyright informaiton	8 years ago
Yi Wang	2c188a20de	Follow QingQing's suggestion	8 years ago
Yi Wang	e2e0fbd418	Add tesnor.h	8 years ago

27 Commits (87189665d90baffb910f120b40148c1627a63c3b)