> -Original Message-
> From: ffmpeg-devel [mailto:ffmpeg-devel-boun...@ffmpeg.org] On Behalf Of
> Guo, Yejun
> Sent: Friday, November 22, 2019 4:02 PM
> To: ffmpeg-devel@ffmpeg.org
> Subject: Re: [FFmpeg-devel] [PATCH 4/4] fate/filter-video: add two tests for
> dnn_p
> -Original Message-
> From: Guo, Yejun
> Sent: Friday, November 22, 2019 3:50 PM
> To: ffmpeg-devel@ffmpeg.org
> Cc: Guo, Yejun
> Subject: [PATCH 4/4] fate/filter-video: add two tests for dnn_processing with
> format rgb24 and grayf32
>
> Signed-off-by: Guo,
Signed-off-by: Guo, Yejun
---
tests/fate/filter-video.mak| 4 ++
...filter-dnn_processing-halve_first_channel_float | 55 ++
.../fate/filter-dnn_processing-halve_gray_float| 55 ++
3 files changed, 114 insertions(+)
create
Signed-off-by: Guo, Yejun
---
doc/filters.texi| 8 ++-
libavfilter/vf_dnn_processing.c | 147 ++--
2 files changed, 118 insertions(+), 37 deletions(-)
diff --git a/doc/filters.texi b/doc/filters.texi
index 1f86ae1..c3f7997 100644
--- a/doc
Signed-off-by: Guo, Yejun
---
libavfilter/vf_dnn_processing.c | 90 -
1 file changed, 45 insertions(+), 45 deletions(-)
diff --git a/libavfilter/vf_dnn_processing.c b/libavfilter/vf_dnn_processing.c
index f59cfb0..ce976ec 100644
--- a/libavfilter
Signed-off-by: Guo, Yejun
---
tools/python/convert_from_tensorflow.py | 5 -
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/tools/python/convert_from_tensorflow.py
b/tools/python/convert_from_tensorflow.py
index 605158a..5e87e22 100644
--- a/tools/python
> -Original Message-
> From: Guo, Yejun
> Sent: Wednesday, November 13, 2019 1:41 PM
> To: ffmpeg-devel@ffmpeg.org
> Cc: Guo, Yejun
> Subject: [PATCH] fate/filter-video: add test for dnn_processing with model
> halve_first_channel
>
> Signed-off-by: Guo,
> -Original Message-
> From: Guo, Yejun
> Sent: Wednesday, November 13, 2019 1:41 PM
> To: ffmpeg-devel@ffmpeg.org
> Cc: Guo, Yejun
> Subject: [PATCH] fate/filter-video: add test for dnn_processing with model
> halve_first_channel
>
> Signed-off-by: Guo,
Signed-off-by: Guo, Yejun
---
tests/fate/filter-video.mak| 3 ++
.../fate/filter-dnn_processing-halve_first_channel | 55 ++
2 files changed, 58 insertions(+)
create mode 100644 tests/ref/fate/filter-dnn_processing-halve_first_channel
diff --git a
Signed-off-by: Guo, Yejun
---
libavfilter/vf_dnn_processing.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libavfilter/vf_dnn_processing.c b/libavfilter/vf_dnn_processing.c
index 87ad354..a916d38 100644
--- a/libavfilter/vf_dnn_processing.c
+++ b/libavfilter
> -Original Message-
> From: ffmpeg-devel [mailto:ffmpeg-devel-boun...@ffmpeg.org] On Behalf Of
> leozhang
> Sent: Friday, November 08, 2019 11:55 AM
> To: ffmpeg-devel@ffmpeg.org
> Subject: [FFmpeg-devel] [PATCH] avfilter/vf_dnn_processing: correct duplicate
> statement
>
> Signed-off-b
> > From: Pedro Arthur [mailto:bygran...@gmail.com]
> > Sent: Thursday, November 07, 2019 1:18 AM
> > To: FFmpeg development discussions and patches
>
> > Cc: Guo, Yejun
> > Subject: Re: [FFmpeg-devel] [PATCH V3] avfilter/vf_dnn_processing: add a
> > ge
> -Original Message-
> From: Guo, Yejun
> Sent: Thursday, October 31, 2019 4:33 PM
> To: ffmpeg-devel@ffmpeg.org
> Cc: Guo, Yejun
> Subject: [PATCH V3] avfilter/vf_dnn_processing: add a generic filter for image
> proccessing with dnn networks
>
> This
nn_processing=model=halve_first_channel.pb:input=dnn_in:output=dnn_out:fmt=rgb24:dnn_backend=tensorflow
-y out.tf.png
Signed-off-by: Guo, Yejun
---
configure | 1 +
doc/filters.texi| 44 ++
libavfilter/Makefile| 1 +
If filter changes frame w/h, AVFILTER_FLAG_SUPPORT_TIMELINE_GENERIC
cannot be supported.
Signed-off-by: Guo, Yejun
---
libavfilter/vf_sr.c | 1 -
1 file changed, 1 deletion(-)
diff --git a/libavfilter/vf_sr.c b/libavfilter/vf_sr.c
index 0433246..b90643c 100644
--- a/libavfilter/vf_sr.c
+++ b
> -Original Message-
> From: Paul B Mahol [mailto:one...@gmail.com]
> Sent: Monday, October 28, 2019 8:01 PM
> To: Guo, Yejun
> Cc: FFmpeg development discussions and patches
> Subject: Re: [FFmpeg-devel] [PATCH V2 4/4] avfilter/vf_dnn_processing: add a
> gen
> -Original Message-
> From: Paul B Mahol [mailto:one...@gmail.com]
> Sent: Monday, October 28, 2019 4:00 PM
> To: FFmpeg development discussions and patches
> Cc: Guo, Yejun
> Subject: Re: [FFmpeg-devel] [PATCH V2 4/4] avfilter/vf_dnn_processing: add a
> gen
> -Original Message-
> From: Guo, Yejun
> Sent: Monday, October 21, 2019 8:39 PM
> To: ffmpeg-devel@ffmpeg.org
> Cc: Guo, Yejun
> Subject: [PATCH V2 4/4] avfilter/vf_dnn_processing: add a generic filter for
> image proccessing with dnn networks
>
> This
-i input.jpg -vf
dnn_processing=model=halve_first_channel.model:input=dnn_in:output=dnn_out:fmt=rgb24:dnn_backend=native
-y out.native.png
./ffmpeg -i input.jpg -vf
dnn_processing=model=halve_first_channel.pb:input=dnn_in:output=dnn_out:fmt=rgb24:dnn_backend=tensorflow
-y out.tf.png
Sign
to support dnn networks more general, we need to know the input info
of the dnn model.
background:
The data type of dnn model's input could be float32, uint8 or fp16, etc.
And the w/h of input image could be fixed or variable.
Signed-off-by: Guo, Yejun
---
libavfilter/dnn/dnn_backend_nat
ed, a flag named
has_bias is added, so change the version number.
Signed-off-by: Guo, Yejun
---
libavfilter/dnn/dnn_backend_native.c | 2 +-
libavfilter/dnn/dnn_backend_native_layer_conv2d.c | 37 +++-
libavfilter/dnn/dnn_backend_native_layer_conv2d.h | 1 +
tests/dnn/dnn-la
so, we can make a filter more general to accept different network
models, by adding a data type convertion after getting data from network.
After we add dt field into struct DNNData, it becomes the same as
DNNInputData, so merge them with one struct: DNNData.
Signed-off-by: Guo, Yejun
> -Original Message-
> From: Paul B Mahol [mailto:one...@gmail.com]
> Sent: Wednesday, October 16, 2019 7:30 PM
> To: Guo, Yejun
> Cc: FFmpeg development discussions and patches
> Subject: Re: [FFmpeg-devel] [PATCH 4/4] avfilter: add a generic filter for rgb
>
> -Original Message-
> From: Paul B Mahol [mailto:one...@gmail.com]
> Sent: Wednesday, October 16, 2019 5:17 PM
> To: FFmpeg development discussions and patches
> Cc: Guo, Yejun
> Subject: Re: [FFmpeg-devel] [PATCH 4/4] avfilter: add a generic filter for rgb
>
-Original Message-
From: Paul B Mahol
Sent: Wednesday, October 16, 2019 5:17 PM
To: FFmpeg development discussions and patches
Cc: Guo, Yejun
Subject: Re: [FFmpeg-devel] [PATCH 4/4] avfilter: add a generic filter for rgb
proccessing with dnn networks
There should be only one
so, we can make a filter more general to accept different network
models, by adding a data type convertion after getting data from network.
After we add dt field into struct DNNData, it becomes the same as
DNNInputData, so merge them with one struct: DNNData.
Signed-off-by: Guo, Yejun
y
- try with following commands
./ffmpeg -i input.jpg -vf
dnn_rgb_processing=model=halve_first_channel.model:input=dnn_in:output=dnn_out:fmt=rgb24:dnn_backend=native
-y out.native.png
./ffmpeg -i input.jpg -vf
dnn_rgb_processing=model=halve_first_channel.pb:input=dnn_in:output=dnn_out:fmt=r
ed, a flag named
has_bias is added, so change the version number.
Signed-off-by: Guo, Yejun
---
libavfilter/dnn/dnn_backend_native.c | 2 +-
libavfilter/dnn/dnn_backend_native_layer_conv2d.c | 37 +++-
libavfilter/dnn/dnn_backend_native_layer_conv2d.h | 1 +
tests/dnn/dnn-la
to support dnn networks more general, we need to know the input info
of the dnn model.
background:
The data type of dnn model's input could be float32, uint8 or fp16, etc.
And the w/h of input image could be fixed or variable.
Signed-off-by: Guo, Yejun
---
libavfilter/dnn/dnn_backend_nat
> -Original Message-
> From: ffmpeg-devel [mailto:ffmpeg-devel-boun...@ffmpeg.org] On Behalf Of
> Zhao Zhili
> Sent: Wednesday, October 16, 2019 12:21 AM
> To: ffmpeg-devel@ffmpeg.org
> Cc: Zhao Zhili
> Subject: [FFmpeg-devel] [PATCH 2/2] FATE/dnn: add .gitignore
>
> ---
> tests/dnn/.g
ify the layer load
> function in native mode
>
> Em ter, 15 de out de 2019 às 02:29, Guo, Yejun
> escreveu:
>
> >
> >
> > > -Original Message-
> > > From: Guo, Yejun
> > > Sent: Wednesday, October 09, 2019 10:08 PM
> > > To: ffm
> -Original Message-
> From: Guo, Yejun
> Sent: Wednesday, October 09, 2019 10:08 PM
> To: ffmpeg-devel@ffmpeg.org
> Cc: Guo, Yejun
> Subject: [PATCH V3 3/3] avfilter/dnn: unify the layer load function in native
> mode
>
> Signed-off-by: Guo, Yeju
Signed-off-by: Guo, Yejun
---
libavfilter/dnn/dnn_backend_native.c | 114 +++--
libavfilter/dnn/dnn_backend_native.h | 2 +-
libavfilter/dnn/dnn_backend_native_layer_conv2d.c | 46 +
libavfilter/dnn/dnn_backend_native_layer_conv2d.h | 1
Signed-off-by: Guo, Yejun
---
libavfilter/dnn/Makefile | 1 +
libavfilter/dnn/dnn_backend_native.c | 34 --
libavfilter/dnn/dnn_backend_native.h | 4 ++-
libavfilter/dnn/dnn_backend_native_layer_conv2d.c | 4
and also change CONV to DLT_CONV2D for better description
Signed-off-by: Guo, Yejun
---
libavfilter/dnn/dnn_backend_native.c | 25 +++--
libavfilter/dnn/dnn_backend_native.h | 12 +++-
libavfilter/dnn/dnn_backend_tf.c | 10 +-
3 files changed, 27
n: unify the layer
> execution function in native mode
>
> Hi,
>
> Em sex, 4 de out de 2019 às 09:59, Guo, Yejun
> escreveu:
>
> >
> >
> > > -Original Message-
> > > From: Guo, Yejun
> > > Sent: Tuesday, September 24, 2019 1:34 PM
> &
> -Original Message-
> From: Guo, Yejun
> Sent: Tuesday, September 24, 2019 1:34 PM
> To: ffmpeg-devel@ffmpeg.org
> Cc: Guo, Yejun
> Subject: [PATCH V2 2/2] avfilter/dnn: unify the layer execution function in
> native
> mode
>
> with this change, the
> -Original Message-
> From: ffmpeg-devel [mailto:ffmpeg-devel-boun...@ffmpeg.org] On Behalf Of
> quinkbl...@foxmail.com
> Sent: Tuesday, October 01, 2019 2:37 PM
> To: ffmpeg-devel@ffmpeg.org
> Cc: Zhao Zhili
> Subject: [FFmpeg-devel] [PATCH] FATE/dnn: fix stack buffer overflow
>
> Fro
with this change, the code will be simpler when more layers supported.
Signed-off-by: Guo, Yejun
---
libavfilter/dnn/dnn_backend_native.c | 38 --
libavfilter/dnn/dnn_backend_native_layer_conv2d.c | 4 ++-
libavfilter/dnn/dnn_backend_native_layer_conv2d.h
and also change CONV to DLT_CONV2D for better description
Signed-off-by: Guo, Yejun
---
libavfilter/dnn/dnn_backend_native.c | 25 +++--
libavfilter/dnn/dnn_backend_native.h | 12 +++-
libavfilter/dnn/dnn_backend_tf.c | 10 +-
3 files changed, 27
please ignore this patch set, there is an issue found in this patch, I'll fix
and send v2 patch.
> -Original Message-
> From: Guo, Yejun
> Sent: Tuesday, September 24, 2019 8:56 AM
> To: ffmpeg-devel@ffmpeg.org
> Cc: Guo, Yejun
> Subject: [PATCH 2/2] avfilt
and also change CONV to DLT_CONV2D for better description
Signed-off-by: Guo, Yejun
---
libavfilter/dnn/dnn_backend_native.c | 25 +++--
libavfilter/dnn/dnn_backend_native.h | 12 +++-
libavfilter/dnn/dnn_backend_tf.c | 10 +-
3 files changed, 27
with this change, the code will be simpler when more layers supported.
Signed-off-by: Guo, Yejun
---
libavfilter/dnn/dnn_backend_native.c | 38 --
libavfilter/dnn/dnn_backend_native_layer_conv2d.c | 4 ++-
libavfilter/dnn/dnn_backend_native_layer_conv2d.h
Signed-off-by: Guo, Yejun
---
libavcodec/libx264.c | 16 ++--
libavcodec/libx265.c | 11 ++-
2 files changed, 24 insertions(+), 3 deletions(-)
diff --git a/libavcodec/libx264.c b/libavcodec/libx264.c
index 86e3530..8788286 100644
--- a/libavcodec/libx264.c
+++ b/libavcodec
layer maximum for
> native mode.
>
> Em sex, 20 de set de 2019 às 11:50, Guo, Yejun
> escreveu:
> >
> >
> >
> > > -Original Message-
> > > From: ffmpeg-devel [mailto:ffmpeg-devel-boun...@ffmpeg.org] On Behalf
> Of
> > > Pedro Arthu
layer maximum for
> native mode.
>
> Hi,
>
> Em sex, 20 de set de 2019 às 01:00, Guo, Yejun
> escreveu:
> >
> > The reason to add this layer is that it is used by srcnn in vf_sr.
> > This layer is currently ignored in native mode. After this patch,
> > we c
Signed-off-by: Guo, Yejun
---
libavfilter/dnn/dnn_backend_native.c | 43 +++-
libavfilter/dnn/dnn_backend_native.h | 2 ++
2 files changed, 34 insertions(+), 11 deletions(-)
diff --git a/libavfilter/dnn/dnn_backend_native.c
b/libavfilter/dnn
Signed-off-by: Guo, Yejun
---
libavfilter/dnn/dnn_backend_native.c | 39 +---
1 file changed, 23 insertions(+), 16 deletions(-)
diff --git a/libavfilter/dnn/dnn_backend_native.c
b/libavfilter/dnn/dnn_backend_native.c
index 22a9a33..1b0aea2 100644
--- a
Signed-off-by: Guo, Yejun
---
tests/dnn/Makefile | 1 +
tests/dnn/dnn-layer-maximum-test.c | 71 ++
tests/fate/dnn.mak | 5 +++
3 files changed, 77 insertions(+)
create mode 100644 tests/dnn/dnn-layer-maximum-test.c
diff
The reason to add this layer is that it is used by srcnn in vf_sr.
This layer is currently ignored in native mode. After this patch,
we can add multiple outputs support for native mode.
Signed-off-by: Guo, Yejun
---
libavfilter/dnn/Makefile | 1 +
libavfilter/dnn
> -Original Message-
> From: Guo, Yejun
> Sent: Thursday, September 05, 2019 2:01 PM
> To: ffmpeg-devel@ffmpeg.org
> Cc: Guo, Yejun
> Subject: [PATCH 4/4] FATE/dnn: add unit test for dnn depth_to_space layer
>
> 'make fate-dnn-layer-depth2space' to ru
> -Original Message-
> From: Guo, Yejun
> Sent: Thursday, September 12, 2019 6:25 PM
> To: ffmpeg-devel@ffmpeg.org
> Cc: Guo, Yejun
> Subject: [PATCH V9] avcodec/libvpxenc: add ROI-based encoding support for
> VP8/VP9
>
> example command line to verify it:
> -Original Message-
> From: ffmpeg-devel [mailto:ffmpeg-devel-boun...@ffmpeg.org] On Behalf Of
> Guo, Yejun
> Sent: Wednesday, September 11, 2019 8:55 AM
> To: ffmpeg-devel@ffmpeg.org
> Subject: Re: [FFmpeg-devel] [PATCH] libavcodec/libx264 and libx265: add a fl
> -Original Message-
> From: Aman Gupta [mailto:a...@tmm1.net] On Behalf Of Aman Gupta
> Sent: Saturday, September 14, 2019 7:31 AM
> To: ffmpeg-devel@ffmpeg.org
> Cc: Guo, Yejun ; Aman Gupta
> Subject: [PATCH] avcodec/v4l2_m2m_enc: check for
> V4L2_CID_MPEG_
hi
> -Original Message-
> From: ffmpeg-devel [mailto:ffmpeg-devel-boun...@ffmpeg.org] On Behalf Of
> Aman Gupta
> Sent: Tuesday, September 03, 2019 9:02 AM
> To: ffmpeg-devel@ffmpeg.org
> Cc: loru...@gmail.com; Aman Gupta ;
> jorge.ramirez.or...@gmail.com; Aman Gupta
> Subject: [FFmpeg-de
example command line to verify it:
./ffmpeg -i input.stream -vf addroi=0:0:iw/3:ih/3:-0.8 -c:v libvpx -b:v 2M
tmp.webm
Signed-off-by: Guo, Yejun
---
libavcodec/libvpxenc.c | 193 +
libavcodec/version.h | 2 +-
2 files changed, 194 insertions
> -Original Message-
> From: James Zern [mailto:jz...@google.com]
> Sent: Thursday, September 12, 2019 2:18 PM
> To: FFmpeg development discussions and patches
> Cc: Guo, Yejun
> Subject: Re: [FFmpeg-devel] [PATCH V8] avcodec/libvpxenc: add ROI-based
> encodin
example command line to verify it:
./ffmpeg -i input.stream -vf addroi=0:0:iw/3:ih/3:-0.8 -c:v libvpx -b:v 2M
tmp.webm
Signed-off-by: Guo, Yejun
---
libavcodec/libvpxenc.c | 188 +
libavcodec/version.h | 2 +-
2 files changed, 189 insertions
please ignore this patch since the version number in libavcodec/version.h has
been changed,
and so conflicts, I'll rebase it and send out a new V8 patch.
> -Original Message-
> From: Guo, Yejun
> Sent: Thursday, August 29, 2019 10:35 AM
> To: ffmpeg-devel@ffmpeg.org
> -Original Message-
> From: Guo, Yejun
> Sent: Monday, September 02, 2019 12:38 PM
> To: ffmpeg-devel@ffmpeg.org
> Cc: Guo, Yejun
> Subject: [PATCH] libavcodec/libx264 and libx265: add a flag to output ROI
> warnings only once.
>
> Signed-off-by: Guo,
'make fate-dnn-layer-conv2d' to run the test
Signed-off-by: Guo, Yejun
---
tests/dnn/Makefile| 1 +
tests/dnn/dnn-layer-conv2d-test.c | 238 ++
tests/fate/dnn.mak| 5 +
3 files changed, 244 insertions(+)
create m
the logic is that one layer in one separated source file to make
the source files simple for maintaining.
Signed-off-by: Guo, Yejun
---
libavfilter/dnn/Makefile | 1 +
libavfilter/dnn/dnn_backend_native.c | 80 +
libavfilter/dnn
'make fate-dnn-layer-depth2space' to run the test
Signed-off-by: Guo, Yejun
---
tests/dnn/Makefile | 1 +
tests/dnn/dnn-layer-depth2space-test.c | 100 +
tests/fate/dnn.mak | 5 ++
3 files changed, 106
the logic is that one layer in one separated source file to make
the source files simple for maintaining.
Signed-off-by: Guo, Yejun
---
libavfilter/dnn/Makefile | 1 +
libavfilter/dnn/dnn_backend_native.c | 44 +-
libavfilter/dnn
Signed-off-by: Guo, Yejun
---
libavcodec/libx264.c | 14 --
libavcodec/libx265.c | 9 -
2 files changed, 20 insertions(+), 3 deletions(-)
diff --git a/libavcodec/libx264.c b/libavcodec/libx264.c
index 86e3530..62316e6 100644
--- a/libavcodec/libx264.c
+++ b/libavcodec
Signed-off-by: Guo, Yejun
---
libavfilter/dnn/dnn_backend_native.c| 43 +++--
tools/python/convert_from_tensorflow.py | 3 +++
tools/python/convert_header.py | 26
3 files changed, 70 insertions(+), 2 deletions(-)
create mode 100644
> -Original Message-
> From: James Zern [mailto:jz...@google.com]
> Sent: Thursday, August 29, 2019 2:14 PM
> To: Guo, Yejun
> Cc: FFmpeg development discussions and patches
> Subject: Re: [FFmpeg-devel] [PATCH V5] avcodec/libvpxenc: add ROI-based
> encoding suppor
Signed-off-by: Guo, Yejun
---
libavfilter/dnn/dnn_backend_native.c| 49 +++---
libavfilter/dnn/dnn_backend_native.h| 2 +-
libavfilter/dnn_interface.h | 2 +-
tools/python/convert_from_tensorflow.py | 111 +---
4 files changed, 142
the info can be saved in dnn operand object without regenerating again and
again,
and it is also needed for layer split/merge, and for memory reuse.
to make things step by step, this patch just focuses on c code,
the change within python script will be added later.
Signed-off-by: Guo, Yejun
currently, the layer number is at the beginning of the .model file,
so we have to scan twice in python script, the first scan to get the
layer number. Only one scan needed after put the layer number at the
end of .model file.
Signed-off-by: Guo, Yejun
---
libavfilter/dnn/dnn_backend_native.c
in python
> script and load in c code
>
> hi,
>
> Em ter, 20 de ago de 2019 às 05:54, Guo, Yejun
> escreveu:
> >
> > Signed-off-by: Guo, Yejun
> > ---
> > libavfilter/dnn/dnn_backend_native.c| 49 +++---
> > libavfilter/dnn
> -Original Message-
> From: James Zern [mailto:jz...@google.com]
> Sent: Thursday, August 29, 2019 12:39 PM
> To: Guo, Yejun
> Cc: FFmpeg development discussions and patches
> Subject: Re: [FFmpeg-devel] [PATCH V5] avcodec/libvpxenc: add ROI-based
> encodin
d (in c code)
> to hold operand infos within network
>
> Hi,
>
>
> Em ter, 20 de ago de 2019 às 05:54, Guo, Yejun
> escreveu:
> >
> > the info can be saved in dnn operand object without regenerating again and
> again,
> > and it is also needed for lay
example command line to verify it:
./ffmpeg -i input.stream -vf addroi=0:0:iw/3:ih/3:-0.8 -c:v libvpx -b:v 2M
tmp.webm
Signed-off-by: Guo, Yejun
---
libavcodec/libvpxenc.c | 189 +
libavcodec/version.h | 2 +-
2 files changed, 190 insertions
please ignore this patch since there is a conflict in version.h, will send out
a new version, thanks.
> -Original Message-
> From: Guo, Yejun
> Sent: Thursday, August 29, 2019 9:29 AM
> To: ffmpeg-devel@ffmpeg.org
> Cc: Guo, Yejun
> Subject: [PATCH V6] avcodec/libvpx
example command line to verify it:
./ffmpeg -i input.stream -vf addroi=0:0:iw/3:ih/3:-0.8 -c:v libvpx -b:v 2M
tmp.webm
Signed-off-by: Guo, Yejun
---
libavcodec/libvpxenc.c | 189 +
libavcodec/version.h | 2 +-
2 files changed, 190 insertions
> -Original Message-
> From: James Zern [mailto:jz...@google.com]
> Sent: Tuesday, August 27, 2019 12:03 PM
> To: FFmpeg development discussions and patches
> Cc: Guo, Yejun
> Subject: Re: [FFmpeg-devel] [PATCH V5] avcodec/libvpxenc: add ROI-based
> encoding suppor
example command line to verify it:
./ffmpeg -i input.stream -vf addroi=0:0:iw/3:ih/3:-0.8 -c:v libvpx -b:v 2M
tmp.webm
Signed-off-by: Guo, Yejun
---
libavcodec/libvpxenc.c | 194 +
1 file changed, 194 insertions(+)
diff --git a/libavcodec
enc: add ROI-based
> encoding support for VP8/VP9 support
>
> On Thu, Aug 22, 2019 at 08:56:28 +0800, Guo, Yejun wrote:
> > +/* This list must be iterated in reverse, so for the case that
> > + * two regions overlapping, the more important area takes effect.
>
> Nit:
example command line to verify it:
./ffmpeg -i input.stream -vf addroi=0:0:iw/3:ih/3:-0.8 -c:v libvpx -b:v 2M
tmp.webm
Signed-off-by: Guo, Yejun
---
libavcodec/libvpxenc.c | 193 +
1 file changed, 193 insertions(+)
diff --git a/libavcodec
I-based
> encoding support for VP8/VP9 support
>
> Hi,
>
> On Tue, Aug 13, 2019 at 8:29 PM Guo, Yejun wrote:
> > example command line to verify it:
> > ./ffmpeg -i input.stream -vf addroi=0:0:iw/3:ih/3:-0.8 -c:v libvpx -b:v 2M
> tmp.webm
> >
> > Signed-o
Signed-off-by: Guo, Yejun
---
libavfilter/dnn/dnn_backend_native.c| 49 +++---
libavfilter/dnn/dnn_backend_native.h| 2 +-
libavfilter/dnn_interface.h | 2 +-
tools/python/convert_from_tensorflow.py | 111 +---
4 files changed, 142
currently, the layer number is at the beginning of the .model file,
so we have to scan twice in python script, the first scan to get the
layer number. Only one scan needed after put the layer number at the
end of .model file.
Signed-off-by: Guo, Yejun
---
libavfilter/dnn/dnn_backend_native.c
the info can be saved in dnn operand object without regenerating again and
again,
and it is also needed for layer split/merge, and for memory reuse.
to make things step by step, this patch just focuses on c code,
the change within python script will be added later.
Signed-off-by: Guo, Yejun
> -Original Message-
> From: ffmpeg-devel [mailto:ffmpeg-devel-boun...@ffmpeg.org] On Behalf Of
> Jun Zhao
> Sent: Monday, August 19, 2019 9:01 PM
> To: ffmpeg-devel@ffmpeg.org
> Cc: Jun Zhao
> Subject: [FFmpeg-devel] [PATCH V1 3/5] lavfi/dnn/dnn_backend_native: Refine
> the coding style
> -Original Message-
> From: ffmpeg-devel [mailto:ffmpeg-devel-boun...@ffmpeg.org] On Behalf Of
> Jun Zhao
> Sent: Monday, August 19, 2019 9:01 PM
> To: ffmpeg-devel@ffmpeg.org
> Cc: Jun Zhao
> Subject: [FFmpeg-devel] [PATCH V1 4/5] lavfi/dnn/dnn_backend_native: fix
> memory leak in erro
if it is initialized randomly, the tensorflow lib will report
error message such as:
Attempt to add output -7920 of depth_to_space4 not in range [0, 1) to node with
type Identity
Signed-off-by: Guo, Yejun
---
libavfilter/dnn/dnn_backend_tf.c | 1 +
1 file changed, 1 insertion(+)
diff --git a
Signed-off-by: Guo, Yejun
---
libavfilter/dnn/dnn_backend_tf.c | 48
1 file changed, 19 insertions(+), 29 deletions(-)
diff --git a/libavfilter/dnn/dnn_backend_tf.c b/libavfilter/dnn/dnn_backend_tf.c
index ca7434a..626fba9 100644
--- a/libavfilter/dnn
> -Original Message-
> From: Guo, Yejun
> Sent: Wednesday, August 14, 2019 3:06 PM
> To: ffmpeg-devel@ffmpeg.org
> Cc: Guo, Yejun
> Subject: [PATCH 2/2] libavfilter/dnn/dnn_backend_tf: add tf.pad support for
> tensorflow backend with native model.
>
>
Signed-off-by: Guo, Yejun
---
libavfilter/dnn/dnn_backend_tf.c | 47
1 file changed, 19 insertions(+), 28 deletions(-)
diff --git a/libavfilter/dnn/dnn_backend_tf.c b/libavfilter/dnn/dnn_backend_tf.c
index ca7434a..75dbe5e 100644
--- a/libavfilter/dnn
if it is not initialized with zero, the tensorflow lib will report
error message such as:
Attempt to add output -7920 of depth_to_space4 not in range [0, 1) to node with
type Identity
Signed-off-by: Guo, Yejun
---
libavfilter/dnn/dnn_backend_tf.c | 1 +
1 file changed, 1 insertion(+)
diff
test it, but let me know if you need them.
yes, I have the .pb files. I missed the patch for such support. Will refine and
send out soon.
>
> Em sex, 9 de ago de 2019 às 12:25, Guo, Yejun
> escreveu:
> >
> >
> >
> > > -Original Message-
> > >
example command line to verify it:
./ffmpeg -i input.stream -vf addroi=0:0:iw/3:ih/3:-0.8 -c:v libvpx -b:v 2M
tmp.webm
Signed-off-by: Guo, Yejun
---
libavcodec/libvpxenc.c | 197 +
1 file changed, 197 insertions(+)
diff --git a/libavcodec
> -Original Message-
> From: James Zern [mailto:jz...@google.com]
> Sent: Wednesday, August 14, 2019 7:32 AM
> To: FFmpeg development discussions and patches
> Cc: Guo, Yejun
> Subject: Re: [FFmpeg-devel] [PATCH V2] avcodec/libvpxenc: add ROI-based
> encodin
example command line to verify it:
./ffmpeg -i input.stream -vf addroi=0:0:iw/3:ih/3:-0.8 -c:v libvpx -b:v 2M
tmp.webm
Signed-off-by: Guo, Yejun
---
configure | 7 ++
libavcodec/libvpxenc.c | 197 +
2 files changed, 204 insertions
example command line to verify it:
./ffmpeg -i input.stream -vf addroi=0:0:iw/3:ih/3:-0.8 -c:v libvpx -b:v 2M
tmp.webm
Signed-off-by: Guo, Yejun
---
libavcodec/libvpxenc.c | 177 +
1 file changed, 177 insertions(+)
diff --git a/libavcodec
> -Original Message-
> From: Guo, Yejun
> Sent: Tuesday, July 30, 2019 9:26 AM
> To: ffmpeg-devel@ffmpeg.org
> Cc: Guo, Yejun
> Subject: [PATCH 2/2] convert_from_tensorflow.py: support conv2d with dilation
>
> conv2d with dilation > 1 generates tens of nodes
n fate/dnn Makefile with ffmpeg static libraries.
This is the same method used in fate/checkasm
Signed-off-by: Guo, Yejun
---
tests/dnn/Makefile | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/tests/dnn/Makefile b/tests/dnn/Makefile
index b2e6680..0e050ea 100644
--- a
Signed-off-by: Guo, Yejun
---
tests/dnn/Makefile | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/tests/dnn/Makefile b/tests/dnn/Makefile
index b2e6680..0e050ea 100644
--- a/tests/dnn/Makefile
+++ b/tests/dnn/Makefile
@@ -4,8 +4,8 @@ DNNTESTOBJS := $(DNNTESTOBJS
it test for
> dnn-layer-pad"
>
> On Thu, Aug 1, 2019 at 4:53 PM Guo, Yejun wrote:
> >
> > This reverts commit 3805aae47966b691f825abab6843f55676437a02.
> >
> > this test calls internal DNN functions within libavfilter, it is
> > not allowed, and so re
501 - 600 of 838 matches
Mail list logo