ffmpeg | branch: master | Guo, Yejun | Thu Sep 5
14:00:39 2019 +0800| [24f507301b4774bb1b082c594d1ce8885b5cabd8] | committer:
Pedro Arthur
FATE/dnn: add unit test for dnn conv2d layer
'make fate-dnn-layer-conv2d' to run the test
Signed-off-by: Guo, Yejun
Signed-off-by: Pe
ffmpeg | branch: master | Guo, Yejun | Thu Sep 5
14:00:54 2019 +0800| [b766a13dba383df065869527aec87843d2183bc1] | committer:
Pedro Arthur
FATE/dnn: add unit test for dnn depth_to_space layer
'make fate-dnn-layer-depth2space' to run the test
Signed-off-by: Guo, Yejun
Signed-off
ffmpeg | branch: master | Guo, Yejun | Thu Sep 5
14:00:28 2019 +0800| [5f058dd693c4bebcd6a293da4630441f3540902f] | committer:
Pedro Arthur
libavfilter/dnn: separate conv2d layer from dnn_backend_native.c to a new file
the logic is that one layer in one separated source file to make
the
ffmpeg | branch: master | Guo, Yejun | Thu Sep 5
14:00:46 2019 +0800| [48133fad056e9173a39333fc5f7d4a724a47a028] | committer:
Pedro Arthur
libavfilter/dnn: separate depth_to_space layer from dnn_backend_native.c to a
new file
the logic is that one layer in one separated source file to make
ffmpeg | branch: master | Guo, Yejun | Thu Sep 12
18:25:21 2019 +0800| [ecd026a48d63166d97f074f072d36ade1b5d270b] | committer:
James Zern
avcodec/libvpxenc: add ROI-based encoding support for VP8/VP9
example command line to verify it:
./ffmpeg -i input.stream -vf addroi=0:0:iw/3:ih/3:-0.8
ffmpeg | branch: master | Guo, Yejun | Fri Sep 20
11:55:55 2019 +0800| [9ae42c130ca989578ac014daef2106e54e09fe34] | committer:
Pedro Arthur
FATE/dnn: add unit test for layer maximum
Signed-off-by: Guo, Yejun
Signed-off-by: Pedro Arthur
> http://git.videolan.org/gitweb.cgi/ffmpeg.git
ffmpeg | branch: master | Guo, Yejun | Fri Sep 20
11:56:03 2019 +0800| [75ca94f3cff8e86036010f496b975cf9c5a7ffb1] | committer:
Pedro Arthur
libavfilter/dnn/dnn_backend_native: find the input operand according to input
name
Signed-off-by: Guo, Yejun
Signed-off-by: Pedro Arthur
> h
ffmpeg | branch: master | Guo, Yejun | Fri Sep 20
11:55:48 2019 +0800| [b2683c66b215ee3b67628880b93f7371d21bc946] | committer:
Pedro Arthur
libavfilter/dnn: add layer maximum for native mode.
The reason to add this layer is that it is used by srcnn in vf_sr.
This layer is currently ignored in
ffmpeg | branch: master | Guo, Yejun | Fri Sep 20
11:56:10 2019 +0800| [8f13a557cad23dfbd8fd5ae4c7bf0d3e9ad96d73] | committer:
Pedro Arthur
libavfilter/dnn: support multiple outputs for native mode
Signed-off-by: Guo, Yejun
Signed-off-by: Pedro Arthur
> http://git.videolan.org/gitweb.
ffmpeg | branch: master | Guo, Yejun | Mon Sep 23
10:02:34 2019 +0800| [104d44138be1f668973ed9ed5693a40d2514600c] | committer:
Mark Thompson
libavcodec/libx264: add a flag to output ROI warnings only once.
Signed-off-by: Guo, Yejun
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Guo, Yejun | Mon Sep 23
10:02:35 2019 +0800| [85e338ab0da63cc9f319d657ee2eefe4d6f32094] | committer:
Mark Thompson
libavcodec/libx265: add a flag to output ROI warnings only once.
Signed-off-by: Guo, Yejun
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Guo, Yejun | Wed Oct 9
22:08:11 2019 +0800| [3fd5ac7e92049b4f31026acdb53a762289f71448] | committer:
Pedro Arthur
avfilter/dnn: unify the layer execution function in native mode
Signed-off-by: Guo, Yejun
Signed-off-by: Pedro Arthur
> http://git.videolan.
ffmpeg | branch: master | Guo, Yejun | Wed Oct 9
22:08:04 2019 +0800| [b78dc27bba2cc612643df7e9c84addc142273e71] | committer:
Pedro Arthur
avfilter/dnn: add DLT prefix for enum DNNLayerType to avoid potential conflicts
and also change CONV to DLT_CONV2D for better description
Signed-off-by
ffmpeg | branch: master | Guo, Yejun | Wed Oct 9
22:08:18 2019 +0800| [2558e62713ebc5f3ea22c1a28d8e9cf3249badaf] | committer:
Pedro Arthur
avfilter/dnn: unify the layer load function in native mode
Signed-off-by: Guo, Yejun
Signed-off-by: Pedro Arthur
> http://git.videolan.org/gitweb.
ffmpeg | branch: master | Guo, Yejun | Mon Oct 21
20:38:17 2019 +0800| [f4b3c0e55c84434eb897c2a4a1179cb1d202c52c] | committer:
Pedro Arthur
avfilter/dnn: add a new interface to query dnn model's input info
to support dnn networks more general, we need to know the input info
of the dnn
ffmpeg | branch: master | Guo, Yejun | Mon Oct 21
20:38:10 2019 +0800| [e1b45b85963b5aa9d67e23638ef9b045e7fbd875] | committer:
Pedro Arthur
avfilter/dnn: get the data type of network output from dnn execution result
so, we can make a filter more general to accept different network
models, by
ffmpeg | branch: master | Guo, Yejun | Mon Oct 21
20:38:03 2019 +0800| [dff39ea9f0154ec52b7548b122a4a5332df3c2c6] | committer:
Pedro Arthur
dnn: add tf.nn.conv2d support for native model
Unlike other tf.*.conv2d layers, tf.nn.conv2d does not create many
nodes (within a scope) in the graph, it
ffmpeg | branch: master | Guo, Yejun | Mon Oct 28
21:11:02 2019 +0800| [912ab246f16699b2b2c782ab63884b00f2e39dae] | committer:
Pedro Arthur
avfilter/vf_sr: correct flags since the filter changes frame w/h
If filter changes frame w/h, AVFILTER_FLAG_SUPPORT_TIMELINE_GENERIC
cannot be supported
ffmpeg | branch: master | Guo, Yejun | Thu Oct 31
16:33:02 2019 +0800| [4d980a8ceba9d0b7cc9f34a5f5cd769a4d7c2773] | committer:
Pedro Arthur
avfilter/vf_dnn_processing: add a generic filter for image proccessing with dnn
networks
This filter accepts all the dnn networks which do image
ffmpeg | branch: master | Guo, Yejun | Fri Nov 8
15:51:40 2019 +0800| [f6e942251c59239520da3e541dc2e0d785a94bad] | committer:
Michael Niedermayer
avfilter/vf_dnn_processing: fix fate-source
Signed-off-by: Guo, Yejun
Signed-off-by: Michael Niedermayer
> http://git.videolan.org/gitweb.
ffmpeg | branch: master | Guo, Yejun | Sat Nov 30
12:24:58 2019 +0800| [b864af033d41e1628ea02424b5979eaeac51b743] | committer:
Michael Niedermayer
MAINTAINERS: add myself to libavfilter/dnn
Signed-off-by: Guo, Yejun
Signed-off-by: Michael Niedermayer
> http://git.videolan.org/gitweb.
ffmpeg | branch: master | Guo, Yejun | Fri Nov 22
15:49:57 2019 +0800| [ed9fc2e3c576703d0bb904a93d4e654c1239f4da] | committer:
Pedro Arthur
avfilter/vf_dnn_processing: refine code for better naming
Signed-off-by: Guo, Yejun
Signed-off-by: Pedro Arthur
> http://git.videolan.org/gitweb.
ffmpeg | branch: master | Guo, Yejun | Fri Nov 22
15:50:04 2019 +0800| [e52070e89c755bffe08ecb9d9ea6be73ca1509cd] | committer:
Pedro Arthur
convert_from_tensorflow.py: add support when kernel size is 1*1 with one
input/output channel (gray image)
Signed-off-by: Guo, Yejun
Signed-off-by
ffmpeg | branch: master | Guo, Yejun | Fri Dec 27
16:34:15 2019 +0800| [04e6f8a143dc8bcec385e94a653b89c67cbaaca1] | committer:
Pedro Arthur
vf_dnn_processing: remove parameter 'fmt'
do not request AVFrame's format in vf_ddn_processing with 'fmt',
but to add anot
ffmpeg | branch: master | Guo, Yejun | Fri Dec 27
16:34:20 2019 +0800| [37d24a6c8fdff897c5e01a8767bfcdc9ddf1f003] | committer:
Pedro Arthur
vf_dnn_processing: add support for more formats gray8 and grayf32
The following is a python script to halve the value of the gray
image. It demos how to
ffmpeg | branch: master | Guo, Yejun | Wed Jan 8
14:43:59 2020 +0800| [95ade711eb4afb8b76a765ef6571e7934ad7f8cc] | committer:
Pedro Arthur
fate/filter-video: add two tests for dnn_processing with frame format rgb24 and
grayf32
Signed-off-by: Guo, Yejun
Signed-off-by: Pedro Arthur
> h
ffmpeg | branch: master | Guo, Yejun | Mon Jan 13
19:37:56 2020 +0800| [4e1ae43b17a50f125e63acc33ea2d25a1965fe69] | committer:
Pedro Arthur
lavfi/dnn_processing: refine code to use function av_image_copy_plane for data
copy
Signed-off-by: Guo, Yejun
Signed-off-by: Pedro Arthur
> h
ffmpeg | branch: master | Guo, Yejun | Wed Jan 29
01:15:28 2020 +0100| [de1b2aa7968343fe70ca924f473d9b53eeb85490] | committer:
Carl Eugen Hoyos
Revert "fate/filter-video: add two tests for dnn_processing with frame format
rgb24 and grayf32"
The tests broke fate without SAMPLES a
ffmpeg | branch: master | Guo, Yejun | Fri Feb 21
19:29:12 2020 +0800| [f9cb7cf424054ea098cb0e329ba49efb41d02b7c] | committer:
Guo, Yejun
avfilter/vf_sr.c: refine code to use AVPixFmtDescriptor.log2_chroma_h/w
Signed-off-by: Guo, Yejun
Reviewed-by: Pedro Arthur
> http://git.videolan.
ffmpeg | branch: master | Guo, Yejun | Thu Feb 20
15:25:57 2020 +0800| [d86a8c056b4a4e607d93cd91cbdc421bf55a022b] | committer:
Guo, Yejun
avfilter/vf_dnn_processing.c: use swscale for uint8<->float32 convert
Signed-off-by: Guo, Yejun
Reviewed-by: Pedro Arthur
> http://git.vid
ffmpeg | branch: master | Guo, Yejun | Fri Feb 21
14:20:48 2020 +0800| [bd50453894182af095c7d7578596e6ff6c58f852] | committer:
Guo, Yejun
avfilter/vf_dnn_processing.c: add planar yuv format support
Only the Y channel is handled by dnn, the UV channels are copied
without changes.
The command
ffmpeg | branch: master | Guo, Yejun | Fri Feb 21
20:40:07 2020 +0800| [e35f96685312c70f7c1cfaadeb966bce1976eb1b] | committer:
Guo, Yejun
avfilter/vf_dnn_processing.c: add frame size change support for planar yuv
format
The Y channel is handled by dnn, and also resized by dnn. The UV
ffmpeg | branch: master | Guo, Yejun | Thu Apr 2
11:18:08 2020 +0800| [e1488700a25f59c3bd7ec709a48cfc69c3debd1b] | committer:
Guo, Yejun
configure: fix build issue of vf_dnn_processing.c when --disable-swscale
vf_dnn_processing.c recently changed to use swscale to trasfer data
between
ffmpeg | branch: master | Guo, Yejun | Sat Mar 21
10:58:11 2020 +0800| [bbc64799dce727aaa73522353eaf01b11ad79755] | committer:
Guo, Yejun
dnn-layer-mathbinary-test: add unit test for subtraction
Signed-off-by: Guo, Yejun
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Guo, Yejun | Fri Mar 20
20:55:38 2020 +0800| [ffa1561608f513b3a5d3d1568f75126f21bce663] | committer:
Guo, Yejun
dnn_backend_native_layer_mathbinary: add sub support
more math binary operations will be added here
Signed-off-by: Guo, Yejun
> http://git.videolan.
ffmpeg | branch: master | Guo, Yejun | Fri Mar 20
20:54:07 2020 +0800| [7e4527e8fa1f8f0bba6a7f7cadc407813f17fbd1] | committer:
Guo, Yejun
avfilter/vf_derain.c: put all the calculation in model file.
currently, the model outputs the rain, and so need a subtraction
in filter c code to get the
ffmpeg | branch: master | Guo, Yejun | Sat Mar 21
17:33:26 2020 +0800| [2114c4241891849323fd65e56f1ab6f70375a291] | committer:
Guo, Yejun
avfilter/vf_dnn_processing.c: fix typo for the linesize of dnn data
Signed-off-by: Guo, Yejun
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Guo, Yejun | Sat Apr 11
13:22:24 2020 +0800| [ef79408e975057e30d8f4afceaa89702fd5f27da] | committer:
Guo, Yejun
dnn/native: add native support for 'mul'
it can be tested with model file generated from above python script:
import tensorflow as tf
import n
ffmpeg | branch: master | Guo, Yejun | Sat Apr 11
13:24:36 2020 +0800| [265b5bd324496d6a342506ffa6157df5f2a85353] | committer:
Guo, Yejun
dnn-layer-mathbinary-test: add unit test for 'mul'
Signed-off-by: Guo, Yejun
> http://git.videolan.org/gitweb.cgi/ffmpeg.gi
ffmpeg | branch: master | Guo, Yejun | Sat Apr 11
13:46:47 2020 +0800| [8ce9d88f930cecd55eb73ea5e8ce749090002aa8] | committer:
Guo, Yejun
dnn/native: add native support for divide
it can be tested with model file generated with below python script:
import tensorflow as tf
import numpy as np
ffmpeg | branch: master | Guo, Yejun | Fri Apr 10
21:35:11 2020 +0800| [6aa7e07e7caed7997e40cee8b203ec56b12d7300] | committer:
Guo, Yejun
dnn/native: add native support for 'add'
It can be tested with the model file generated with below python script:
import tensorflow as tf
impor
ffmpeg | branch: master | Guo, Yejun | Sat Apr 11
13:50:32 2020 +0800| [2e38c63630c29ee7024edc4f01d02fb7b641] | committer:
Guo, Yejun
dnn-layer-mathbinary-test: add unit test for divide
Signed-off-by: Guo, Yejun
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Guo, Yejun | Fri Apr 10
22:32:02 2020 +0800| [17006196a6a51d8a34c41518cc1d17d703f61201] | committer:
Guo, Yejun
dnn-layer-mathbinary-test: add unit test for add
Signed-off-by: Guo, Yejun
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Guo, Yejun | Sun Apr 26
15:49:39 2020 +0800| [6fd61234d59783d87f917dcac4d7cf2b89a661c5] | committer:
Guo, Yejun
dnn-layer-mathbinary-test: add unit test for minimum
Signed-off-by: Guo, Yejun
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Guo, Yejun | Sun Apr 26
15:46:38 2020 +0800| [71e28c5422e321640b69ee512eef3e899c746e1e] | committer:
Guo, Yejun
dnn/native: add native support for minimum
it can be tested with model file generated with below python script:
import tensorflow as tf
import numpy as np
ffmpeg | branch: master | Guo Yejun | Wed Jun 10 13:36:11
2020 +0800| [0b3bd001ac1745d9d008a2d195817df57d7d1d14] | committer: Guo, Yejun
dnn_backend_native: check operand index
it fixed the issue in https://trac.ffmpeg.org/ticket/8716
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=com
ffmpeg | branch: master | Guo Yejun | Wed Jun 10 10:59:19
2020 +0800| [fc932195ab0c9c00fa0cd9620c60763d978d495b] | committer: Guo, Yejun
dnn_backend_native.c: refine code for fail case
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=fc932195ab0c9c00fa0cd9620c60763d978d4
ffmpeg | branch: release/4.3 | Guo Yejun | Wed Jun 10
10:59:19 2020 +0800| [5530748bfdf1a4d41d4c92e59f662c94e38a5f94] | committer:
Guo, Yejun
dnn_backend_native.c: refine code for fail case
(cherry-pick from fc932195ab0c9c00fa0cd9620c60763d978d495b)
Signed-off-by: Guo, Yejun
> h
ffmpeg | branch: release/4.3 | Guo Yejun | Wed Jun 10
13:36:11 2020 +0800| [dd273d359e45ab69398ac0dc41206d5f1a9371bf] | committer:
Guo, Yejun
dnn_backend_native: check operand index
it fixed the issue in https://trac.ffmpeg.org/ticket/8716
(cherry-pick from
ffmpeg | branch: master | Guo, Yejun | Mon May 25
15:38:09 2020 +0800| [ff37ebaf30e675227655d9055069471bb45e5ceb] | committer:
Guo, Yejun
dnn: add openvino as one of dnn backend
OpenVINO is a Deep Learning Deployment Toolkit at
https://github.com/openvinotoolkit/openvino, it supports CPU, GPU
ffmpeg | branch: master | Guo, Yejun | Mon May 25
16:16:22 2020 +0800| [9bcf2aa47756c253238903f3e5ef6c96d927cfe6] | committer:
Guo, Yejun
vf_dnn_processing.c: add dnn backend openvino
We can try with the srcnn model from sr filter.
1) get srcnn.pb model file, see filter sr
2) convert srcnn.pb
ffmpeg | branch: master | Guo, Yejun | Fri Aug 7
14:32:55 2020 +0800| [0a51abe8ab71c03eff66200265142fc910bcbfe2] | committer:
Guo, Yejun
dnn: add backend options when load the model
different backend might need different options for a better performance,
so, add the parameter into dnn
ffmpeg | branch: master | Guo, Yejun | Thu Aug 13
14:48:11 2020 +0800| [3c05c8a15f2fa5543213f6c8cbe439ad4376137d] | committer:
Guo, Yejun
dnn_backend_tf.c: fix build issue for tensorflow backend
Signed-off-by: Guo, Yejun
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Guo, Yejun | Thu Aug 13
16:19:48 2020 +0800| [0f7a99e37ae52f9ecdc4c81195c14b03f5be3dfd] | committer:
Guo, Yejun
dnn: move output name from DNNModel.set_input_output to DNNModule.execute_model
currently, output is set both at DNNModel.set_input_output and
ffmpeg | branch: master | Guo, Yejun | Mon Aug 24
16:09:59 2020 +0800| [6918e240d706f7390272976d8b8d502afe426a18] | committer:
Guo, Yejun
dnn: add userdata for load model parameter
the userdata will be used for the interaction between AVFrame and DNNData
> http://git.videolan.org/gitweb.
ffmpeg | branch: master | Guo, Yejun | Thu Sep 10
22:29:57 2020 +0800| [fce3e3e137843d86411f8868f18e1c3f472de0e5] | committer:
Guo, Yejun
dnn: put DNNModel.set_input and DNNModule.execute_model together
suppose we have a detect and classify filter in the future, the
detect filter generates
ffmpeg | branch: master | Guo, Yejun | Fri Sep 11
22:15:04 2020 +0800| [e71d73b09652f4fc96e512a7d6d4c2ab41860f27] | committer:
Guo, Yejun
dnn: add a new interface DNNModel.get_output
for some cases (for example, super resolution), the DNN model changes
the frame size which impacts the filter
ffmpeg | branch: master | Guo, Yejun | Fri Aug 28
12:51:44 2020 +0800| [2003e32f62d94ba75b59d70632c9f2862b383591] | committer:
Guo, Yejun
dnn: change dnn interface to replace DNNData* with AVFrame*
Currently, every filter needs to provide code to transfer data from
AVFrame* to model input
ffmpeg | branch: master | Guo, Yejun | Mon Oct 12
15:52:26 2020 +0800| [c4a3dbe726150d9217a4d3fed47b012839e33d82] | committer:
Guo, Yejun
dnn_backend_tf.c: add option sess_config for tf backend
TensorFlow C library accepts config for session options to
set different parameters for the
ffmpeg | branch: master | Guo, Yejun | Thu Jan 10
16:53:30 2019 +0800| [1ef4828276e4c568a13b4c57c08aa75470adec9f] | committer:
Derek Buitenhuis
avutil: add ROI (Region Of Interest) data struct and bump version
The encoders such as libx264 support different QPs offset for different MBs,
it
ffmpeg | branch: master | Guo, Yejun | Thu Jan 10
16:54:30 2019 +0800| [aceb9131c16918164279cf0f8e1b5384610e3245] | committer:
Derek Buitenhuis
avcodec/libx264: add support for ROI-based encoding
This patch just enables the path from ffmpeg to libx264,
the more encoders can be added later
ffmpeg | branch: master | Guo, Yejun | Thu Jan 24
00:11:05 2019 +0800| [e0ad7d57415b6c8e32126f1bc21052b2c3698b39] | committer:
Derek Buitenhuis
avcodec/libx265: add support for ROI-based encoding
Signed-off-by: Guo, Yejun
Signed-off-by: Derek Buitenhuis
> http://git.videolan.org/gitweb.
ffmpeg | branch: master | Guo, Yejun | Tue Mar 5
06:09:18 2019 +0800| [d9b2668766e3e924d4ebb3c6531b449874e13666] | committer:
James Almer
configure: use vpx_codec_vp8_dx/cx for libvpx-vp8 checking
Signed-off-by: Guo, Yejun
Signed-off-by: James Almer
> http://git.videolan.org/gitweb.
ffmpeg | branch: master | Guo, Yejun | Tue Mar 5
06:09:11 2019 +0800| [402bf262375dfecd0e90d7acc67c238abe952fc3] | committer:
James Almer
configure: add missing pthreads extralibs dependency for libvpx-vp9
Signed-off-by: Guo, Yejun
Signed-off-by: James Almer
> http://git.videolan.
ffmpeg | branch: release/4.1 | Guo, Yejun | Tue Mar 5
06:09:11 2019 +0800| [9bf40978c65714bd4448d6d445ffe5ac36189ec3] | committer:
Michael Niedermayer
configure: add missing pthreads extralibs dependency for libvpx-vp9
Signed-off-by: Guo, Yejun
Signed-off-by: James Almer
(cherry picked
ffmpeg | branch: release/4.1 | Guo, Yejun | Tue Mar 5
06:09:18 2019 +0800| [83bfd4f3b5b649fc84066495a2886bc819763fd1] | committer:
Michael Niedermayer
configure: use vpx_codec_vp8_dx/cx for libvpx-vp8 checking
Signed-off-by: Guo, Yejun
Signed-off-by: James Almer
(cherry picked from commit
ffmpeg | branch: release/4.0 | Guo, Yejun | Tue Mar 5
06:09:18 2019 +0800| [33651c09407e83b011dab95e15b1519bf48cb32e] | committer:
Michael Niedermayer
configure: use vpx_codec_vp8_dx/cx for libvpx-vp8 checking
Signed-off-by: Guo, Yejun
Signed-off-by: James Almer
(cherry picked from commit
ffmpeg | branch: release/4.0 | Guo, Yejun | Tue Mar 5
06:09:11 2019 +0800| [01209d220b36e42a307233249f917e578ebacc4c] | committer:
Michael Niedermayer
configure: add missing pthreads extralibs dependency for libvpx-vp9
Signed-off-by: Guo, Yejun
Signed-off-by: James Almer
(cherry picked
ffmpeg | branch: master | Guo, Yejun | Thu Apr 25
10:14:25 2019 +0800| [7adfb6132e0823de06f92ecbcb485eeb4260d407] | committer:
Pedro Arthur
libavfilter/dnn: avoid memcpy for tensorflow dnn output
use TF_Tensor's cpu address to avoid extra memcpy.
Signed-off-by: Guo, Yejun
Signed-o
ffmpeg | branch: master | Guo, Yejun | Thu Apr 25
10:13:49 2019 +0800| [014b6a56f8e4020dd9ec2255266cd0a24c885c0c] | committer:
Pedro Arthur
libavfilter/dnn_backend_tf.c: set layer_add_res for input layer
otherwise, the following check will return error if layer_add_res
is randomly initialized
ffmpeg | branch: master | Guo, Yejun | Thu Apr 25
10:14:33 2019 +0800| [25c1cd909fa6c8b6b778dc24192dc3ec780324b0] | committer:
Pedro Arthur
libavfilter/dnn: support multiple outputs for tensorflow model
some models such as ssd, yolo have more than one output.
the clean up code in this patch
ffmpeg | branch: master | Guo, Yejun | Thu Apr 25
10:14:08 2019 +0800| [05f86f05bb5060492dd3ff22c23628e4e4334a1e] | committer:
Pedro Arthur
libavfilter/dnn: remove limit for the name of DNN model input/output
remove the requirment that the name of DNN model input/output
should be "
ffmpeg | branch: master | Guo, Yejun | Thu Apr 25
10:14:17 2019 +0800| [e2b92896c4ca609c851ea8c1a1bfd5d0918a5269] | committer:
Pedro Arthur
libavfilter/dnn: determine dnn output during execute_model instead of
set_input_output
Currently, within interface set_input_output, the dims/memory of
ffmpeg | branch: master | Guo, Yejun | Thu Apr 25
10:14:01 2019 +0800| [05aec8bb13cc1d698f76c6972a23521a3fba5596] | committer:
Pedro Arthur
libavfilter/vf_sr: refine code to remove keyword 'else'
remove 'else' since there is always 'return' in 'if' sc
ffmpeg | branch: master | Guo, Yejun | Thu Apr 25
10:14:42 2019 +0800| [c636dc9819ebab1a84237cc017a6a3d35ebc9cdc] | committer:
Pedro Arthur
libavfilter/dnn: add more data type support for dnn model input
currently, only float is supported as model input, actually, there
are other data types
ffmpeg | branch: master | Guo, Yejun | Wed Apr 24
21:13:21 2019 +0800| [ae6486c6251039d3a6bb5a90e1d818331cf55edc] | committer:
Alexander Strasser
configure: replace 'pr' with printf since busybox does not support pr
This patch is based on https://trac.ffmpeg.org/ticket/5680 provide
ffmpeg | branch: master | Guo, Yejun | Thu Jun 13
13:30:38 2019 +0800| [50e194e6e126b7fe8e91ecc2e929335d4dc64fcf] | committer:
Pedro Arthur
tools/python: add script to convert TensorFlow model (.pb) to native model
(.model)
For example, given TensorFlow model file espcn.pb,
to generate
ffmpeg | branch: master | Guo, Yejun | Thu Jun 13
13:30:44 2019 +0800| [231d0c819fe49f14f49c240470dd92f43fe2355c] | committer:
Gyan Doshi
doc/filters: update how to generate native model for sr filter
Signed-off-by: Guo, Yejun
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
ffmpeg | branch: master | Guo, Yejun | Tue Jul 16
13:55:45 2019 +0800| [1b9064e3f4ca4cf744f5112c02b31ffd1b44f4c4] | committer:
Pedro Arthur
libavfilter/dnn: move dnn files from libavfilter to libavfilter/dnn
it is expected that there will be more files to support native mode,
so put all the
ffmpeg | branch: master | Guo, Yejun | Mon Jul 29
09:56:33 2019 +0800| [df8db345523f03b2163febe97a4cf07446b28230] | committer:
Pedro Arthur
dnn: add layer pad which is equivalent to tf.pad
the reason to add this layer first is that vf_sr uses it in its
tensorflow model, and the next plan is
ffmpeg | branch: master | Guo, Yejun | Mon Jul 29
09:56:40 2019 +0800| [3805aae47966b691f825abab6843f55676437a02] | committer:
Pedro Arthur
fate: add unit test for dnn-layer-pad
'make fate-dnn-layer-pad' to run the test
Signed-off-by: Guo, Yejun
Signed-off-by: Pedro Arthu
ffmpeg | branch: master | Guo, Yejun | Mon Jul 29
09:56:54 2019 +0800| [ccbab41039af424237eaac5c302c293ab97540f8] | committer:
Pedro Arthur
dnn: convert tf.pad to native model in python script, and load/execute it in
the c code.
since tf.pad is enabled, the conv2d(valid) changes back to its
ffmpeg | branch: master | Guo, Yejun | Tue Jul 30
09:25:56 2019 +0800| [2c01434d60dbc607076a34d129e93a3585cbb277] | committer:
Pedro Arthur
convert_from_tensorflow.py: add option to dump graph for visualization in
tensorboard
Signed-off-by: Guo, Yejun
Signed-off-by: Pedro Arthur
> h
ffmpeg | branch: master | Guo, Yejun | Tue Jul 30
09:26:18 2019 +0800| [ddd92ba2c6c6d0b5f3d5b4496ab07fbcf73b58a2] | committer:
Pedro Arthur
convert_from_tensorflow.py: support conv2d with dilation
conv2d with dilation > 1 generates tens of nodes in graph, it is not
easy to parse each node
ffmpeg | branch: master | Guo, Yejun | Wed Aug 7
10:44:27 2019 +0800| [d0fa1a58da71492f665759e73c47daf04a9b658c] | committer:
Pedro Arthur
FATE/dnn: let fate/dnn tests depend on ffmpeg static libraries
background:
DNN (deep neural network) is a sub module of libavfilter, and FATE/dnn
is unit
ffmpeg | branch: master | Guo, Yejun | Fri Aug 16
10:36:56 2019 +0800| [29aeeb3e3e7b21a56ff4b1684fb65fa242a66c8d] | committer:
Pedro Arthur
libavfilter/dnn/dnn_backend_tf: fix typo that variable uninitialized.
if it is initialized randomly, the tensorflow lib will report
error message such as
ffmpeg | branch: master | Guo, Yejun | Fri Aug 16
10:37:09 2019 +0800| [67889d4715d4a65864fedba1b8e5e2918bd8152a] | committer:
Pedro Arthur
libavfilter/dnn/dnn_backend_tf: add tf.pad support for tensorflow backend with
native model.
Signed-off-by: Guo, Yejun
Signed-off-by: Pedro Arthur
ffmpeg | branch: master | Guo, Yejun | Tue Aug 20
16:50:34 2019 +0800| [83e0b71f66f2f0cc293305e2bd138d87660a8c5b] | committer:
Pedro Arthur
dnn: export operand info in python script and load in c code
Signed-off-by: Guo, Yejun
Signed-off-by: Pedro Arthur
> http://git.videolan.
ffmpeg | branch: master | Guo, Yejun | Thu Aug 29
13:53:33 2019 +0800| [09a455a24649af36e8eea81029be7a410201be4c] | committer:
Pedro Arthur
dnn: introduce dnn operand (in c code) to hold operand infos within network
the info can be saved in dnn operand object without regenerating again and
ffmpeg | branch: master | Guo, Yejun | Thu Aug 29
13:53:41 2019 +0800| [2d5e39c13e500f96a18729ac16472ff69f8d2cfe] | committer:
Pedro Arthur
dnn: change .model file format to put layer number at the end of file
currently, the layer number is at the beginning of the .model file,
so we have to
ffmpeg | branch: master | Guo, Yejun | Mon Sep 2
12:35:58 2019 +0800| [022f50d3febe5232e67d38b0292d86f7aa8a360c] | committer:
Pedro Arthur
libavfilter/dnn: add header into native model file
Signed-off-by: Guo, Yejun
Signed-off-by: Pedro Arthur
> http://git.videolan.org/gitweb.
ffmpeg | branch: master | Guo, Yejun | Tue Nov 17
20:55:13 2020 +0800| [2b177033bb0fe7637448b08ec6a84325b5b25945] | committer:
Guo, Yejun
dnn_backend_openvino.c: separate function execute_model_ov
function fill_model_input_ov and infer_completion_callback are
extracted, it will help the async
ffmpeg | branch: master | Guo, Yejun | Wed Nov 18
14:16:24 2020 +0800| [38089925fa41d0b767e8c47fb236dec390ec38ac] | committer:
Guo, Yejun
dnn_backend_openvino.c: refine code for error handle
Signed-off-by: Xie, Lin
Signed-off-by: Wu Zhiwen
Signed-off-by: Guo, Yejun
> h
ffmpeg | branch: master | Guo, Yejun | Wed Oct 7
10:47:31 2020 +0800| [1b64954e42b4dc685a82d5576527a2166d818773] | committer:
Guo, Yejun
vf_dnn_processing.c: replace filter_frame with activate func
with this change, dnn_processing can use DNN async interface later.
Signed-off-by: Xie, Lin
ffmpeg | branch: master | Guo, Yejun | Wed Nov 18
14:28:06 2020 +0800| [39f5cb4bd1bbb12632baac24dd7b9bb5a68bef8d] | committer:
Guo, Yejun
dnn_interface: add interface to support async execution
Signed-off-by: Xie, Lin
Signed-off-by: Wu Zhiwen
Signed-off-by: Guo, Yejun
> h
ffmpeg | branch: master | Guo, Yejun | Wed Nov 18
14:30:11 2020 +0800| [e67b5d0a247cc7c896c6c1042cf140c429866738] | committer:
Guo, Yejun
dnn: add async execution support for openvino backend
Signed-off-by: Xie, Lin
Signed-off-by: Wu Zhiwen
Signed-off-by: Guo, Yejun
> h
ffmpeg | branch: master | Guo, Yejun | Wed Nov 18
14:54:10 2020 +0800| [5024286465acc35ae005715e0152a135df30ad0a] | committer:
Guo, Yejun
dnn_interface: change from 'void *userdata' to 'AVFilterContext *filter_ctx'
'void *' is too flexible, since we can derive
ffmpeg | branch: master | Guo, Yejun | Wed Dec 16
14:07:00 2020 +0800| [c720286ee3bf59420e6f6ddd1f999f2f327c8967] | committer:
Guo, Yejun
vf_dnn_processing.c: add async support
Signed-off-by: Xie, Lin
Signed-off-by: Wu Zhiwen
Signed-off-by: Guo, Yejun
> http://git.videolan.org/gitweb.
ffmpeg | branch: master | Guo, Yejun | Wed Dec 30
09:37:14 2020 +0800| [8e78d5d394a3cdeb1da195aa901ff2f7b5b1b3ea] | committer:
Guo, Yejun
dnn: fix redefining typedefs and also refine naming with correct prefix
The prefix for symbols not exported from the library and not
local to one
ffmpeg | branch: master | Guo, Yejun | Wed Dec 30
10:47:29 2020 +0800| [97f520b700b9c250883bf6e3a06fbb6ebffc3c91] | committer:
Guo, Yejun
dnn: fix issue when pthread is not supported
Signed-off-by: Guo, Yejun
> http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commi
1 - 100 of 149 matches
Mail list logo
