| Commit message (Expand) | Author | Age |
* | Merge commit for internal changes | Maciek Chociej | 2016-06-22 |
|\ |
|
| * | Improve convolution autotune process. The max batch size VGG model can handle | Xiaoqiang Zheng | 2016-06-21 |
* | | Hard-code libcuda version number to "1". Fixes #2865. | Bob Adolf | 2016-06-14 |
| * | Merge changes from github. | Vijay Vasudevan | 2016-06-11 |
* | | Merge commit for internal changes | Vijay Vasudevan | 2016-06-09 |
|\| |
|
| * | Remove unused MultiOpActivation from stream executor. | A. Unique TensorFlower | 2016-06-08 |
| * | Change a couple of error messages related to getting the NUMA affinity | A. Unique TensorFlower | 2016-06-08 |
* | | Merge commit for internal changes | Vijay Vasudevan | 2016-06-07 |
|\| |
|
| * | Enable fp16 for most of the pooling ops (MaxPool, AvgPool, associated | Benoit Steiner | 2016-06-06 |
| * | Merge changes from github. | Martin Wicke | 2016-06-06 |
* | | cuda 8 api uses cudaDataType for 16bit float type (#2556) | Dr. Kashif Rasul | 2016-06-06 |
* | | cleans up warning/errors tensorflow/stream_executor (#2555) | Fabrizio Milo | 2016-06-03 |
| * | Enable fp16 for most of the pooling ops (MaxPool, AvgPool, associated | A. Unique TensorFlower | 2016-06-03 |
| * | Enable fp16 for most of the pooling ops (MaxPool, AvgPool, associated | Benoit Steiner | 2016-06-03 |
|/ |
|
* | Update copyright for 3p/tf. | A. Unique TensorFlower | 2016-06-02 |
* | Merge changes from github. | Vijay Vasudevan | 2016-05-26 |
* | StreamExecutor add CUDA support for cudnnConvolutionBackwardBias | A. Unique TensorFlower | 2016-05-25 |
* | Add macros for mutex_lock and shared_lock, so compilation fails if | A. Unique TensorFlower | 2016-05-23 |
* | Extend the StreamExecutor interface with ...WithAlgorithm() calls for fp16; | A. Unique TensorFlower | 2016-05-17 |
* | Add Autotune support for Winograd on Cudnn R5. | Xiaoqiang Zheng | 2016-05-16 |
* | In the StreamExecutor, make lack of CUDA 7.5 a non-fatal error for SGEMM | A. Unique TensorFlower | 2016-05-13 |
* | Maintain current cuda context in CUDADriver by checking against a threadlocal | A. Unique TensorFlower | 2016-05-12 |
* | Add fp16 cuDNN convolution support to StreamExecutor. (TensorFlow ops will | A. Unique TensorFlower | 2016-05-11 |
* | Add fp16 matrix multiplication (GEMM) support to StreamExecutor, gated on | A. Unique TensorFlower | 2016-05-11 |
* | Add the autotune version for the backward passes. It is currently | Xiaoqiang Zheng | 2016-05-11 |
* | fp16 support for BiasAdd. Includes support for atomic adds for Eigen::half, | A. Unique TensorFlower | 2016-05-10 |
* | Adding autotune to the Cudnn conv algorithm selection. | Xiaoqiang Zheng | 2016-05-09 |
* | Merge changes from github. | A. Unique TensorFlower | 2016-05-05 |
* | Change CUDAStream::GetOrCreateCompletedEvent from lazy initialization the | A. Unique TensorFlower | 2016-05-03 |
* | Switch CuDNN avg pooling to not use padding when computing the average. | A. Unique TensorFlower | 2016-04-27 |
* | Generalize the stream executor interface to support N-d operations. | A. Unique TensorFlower | 2016-04-27 |
* | Calling Cudnn calls in the current thread, instead of dispatching it into its | Xiaoqiang Zheng | 2016-04-22 |
* | Merge changes from github. | Illia Polosukhin | 2016-04-18 |
* | Support ScratchAllocator in BLAS Batched GEMM | A. Unique TensorFlower | 2016-03-18 |
* | TensorFlow: change cuda-diagnostics to search for so.1 | Vijay Vasudevan | 2016-02-18 |
* | TensorFlow: switch dsoloader path for libcuda to end in .so.1 | Vijay Vasudevan | 2016-02-18 |
* | TensorFlow: fix missing GetCudaVersion() for libcuda.so. h/t to @3XX0. | Vijay Vasudevan | 2016-02-18 |
* | Merge changes from github. | Vijay Vasudevan | 2016-02-17 |
* | Prepare to take another snapshot from stream-executor. | Xiaoqiang Zheng | 2016-02-16 |
* | Deleted unused code | Benoit Steiner | 2016-02-05 |
* | Add assert_same_float_dtype. | A. Unique TensorFlower | 2016-02-02 |
* | Use cc_binary rather than cc_library to reduce size of native library in APK ... | A. Unique TensorFlower | 2016-01-29 |
* | Global search & replace to move to the new location for | Josh Levenberg | 2016-01-26 |
* | Make cuda_dnn work with R4. | Xiaoqiang Zheng | 2016-01-22 |
* | #include "tensorflow/core/platform/mutex.h" | Josh Levenberg | 2016-01-07 |
* | #include third_party/tensorflow/core/platform/macros.h | Josh Levenberg | 2016-01-06 |
* | Added 'logging' import to control_flow_ops which is used in the file but not ... | A. Unique TensorFlower | 2016-01-05 |
* | Fix some lint errors in image_ops_test.py and word2vec_basic.py | Vijay Vasudevan | 2015-12-22 |
* | Adds fft2d and ifft2d. | A. Unique TensorFlower | 2015-12-21 |
* | TensorFlow: Upstream changes from git. | Vijay Vasudevan | 2015-12-09 |