experiments/STT-tensorflow
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity
r2.0
STT-tensorflow/tensorflow/lite/delegates/gpu/common
History
A. Unique TensorFlower 1c46da48dc Make GPU delegate transformations compatible with -Wmissing-braces 
PiperOrigin-RevId: 261012887
2019-07-31 16:31:25 -07:00
..
testing TFLite GPU Delegate: Implement InvokeInterpreter() function for further testing. 2019-06-24 11:45:01 -07:00
transformations Make GPU delegate transformations compatible with -Wmissing-braces 2019-07-31 16:31:25 -07:00
access_type.h Move AccessType to its own header to be shared by multiple gpu delegates. 2019-05-20 09:31:29 -07:00
BUILD Remove superfluous Dequantize nodes in GPU delegate when executing float16 quantized models. 2019-07-25 07:25:46 -07:00
convert.cc New implementation of the transposed convolution kernel which is faster and easier on the eyes. 2019-07-25 15:39:39 -07:00
convert.h New implementation of the transposed convolution kernel which is faster and easier on the eyes. 2019-07-25 15:39:39 -07:00
data_type.cc
data_type.h
memory_management_test.cc Allow to use naive and equality memory management algorithms, when sizes of 2019-06-20 11:18:08 -07:00
memory_management.cc Allow to use naive and equality memory management algorithms, when sizes of 2019-06-20 11:18:08 -07:00
memory_management.h Allow to use naive and equality memory management algorithms, when sizes of 2019-06-20 11:18:08 -07:00
model_builder_test.cc Remove superfluous Dequantize nodes in GPU delegate when executing float16 quantized models. 2019-07-25 07:25:46 -07:00
model_builder.cc TFLite GPU: Rename SOFT_MAX and SoftMax to SOFTMAX and Softmax, respectively. 2019-07-31 12:03:19 -07:00
model_builder.h Get rid of TensorRefFloat32 specialization. It no longer keeps type in template params, therefore, it is better to use ref explicitly. 2019-06-18 00:43:40 -07:00
model_test.cc Get rid of TensorRefFloat32 specialization. It no longer keeps type in template params, therefore, it is better to use ref explicitly. 2019-06-18 00:43:40 -07:00
model_transformer.cc
model_transformer.h
model.h Added RemoveOneInputOneOutputNode for graph. 2019-04-09 12:21:28 -07:00
operations.cc TFLite GPU: Rename SOFT_MAX and SoftMax to SOFTMAX and Softmax, respectively. 2019-07-31 12:03:19 -07:00
operations.h TFLite GPU: Rename SOFT_MAX and SoftMax to SOFTMAX and Softmax, respectively. 2019-07-31 12:03:19 -07:00
shape_test.cc
shape.cc Optimization of shape.LinearIndex for OHWI layout. 2019-05-08 10:48:35 -07:00
shape.h Allow to use naive and equality memory management algorithms, when sizes of 2019-06-20 11:18:08 -07:00
status.h
tensor.h Get rid of TensorRefFloat32 specialization. It no longer keeps type in template params, therefore, it is better to use ref explicitly. 2019-06-18 12:28:55 -07:00
types.h
util_test.cc
util.h Prepare new api for GPU 2019-06-21 00:35:01 -07:00