\ No newline at end of file +`--device=/dev/dri:/dev/dri` passes the host GPU through to the container for GPU inference (`device=GPU`). + +### 6. Run inference + +Mount your model directory and a test video, then run the desired model. The examples below use `docker run --rm` for a one-shot test; replace the image tag, model path, and video path as needed. + +**Enhanced EDSR (2× upscale, `rgb24` input, `model_type=2`):** + +```bash +docker run --rm \ + -v /path/to/models:/models:ro \ + -v /path/to/input.mp4:/input.mp4:ro \ + -v /path/to/output:/output \ + ffmpeg_ivsr_sdk_
-The patches included in `patches` folder are specifically for FFmpeg n7.1.
+The patches included in `patches` folder are specifically for FFmpeg n8.1.
Additionally, there are other parameters that you can use. These parameters are listed in the table below:
@@ -18,12 +22,12 @@ Additionally, there are other parameters that you can use. These parameters are |output|output name of the model|NULL|output| |device|device for inference task|CPU|CPU or GPU| |model_type|type for models|0|0 for Enhanced BasicVSR, 1 for SVP models, 2 for Enhanced EDSR, 3 for one CUSTOM VSR, 4 for TSENet| -|normalize_factor|factor for normalization|1.0|255.0 for Enhanced EDSR, 1.0 for other models supported in current version| -|num_streams|number of execution streams for the throughput mode (now valid only for GPU devices).|1|use `benchmark_app` (a tool provided by OpenVINO Toolkit), to get the appropriate value for the best throughput| -|extension|extension lib file full path, required for loading Enhanced BasicVSR model| -|op_xml|custom op xml file full path, required for loading Enhanced BasicVSR model| -|nif|number of input frames in batch sent to the DNN backend|1|3 for Enhanced BasicVSR, 1 for other models supported in current version| -|nireq|number of request|0|use the default setting or set it to match the number of cpu cores| +|normalize_factor|normalizing factor for models that do not require input normalization to [0, 1]|1.0|255.0 for Enhanced EDSR, 1.0 for all other models| +|num_streams|number of execution streams for throughput mode (valid only for GPU devices)|1|use `benchmark_app` to determine the best value| +|extension|extension lib file full path, required for Enhanced BasicVSR|—|—| +|op_xml|custom op xml file full path, required for Enhanced BasicVSR|—|—| +|nif|number of input frames in batch sent to the DNN backend|1|3 for Enhanced BasicVSR, 1 for other models| +|nireq|number of infer requests|0|leave as default or set to match CPU core count| Here are some examples of FFmpeg command lines to run inference with the supported models using the `ivsr` backend.
diff --git a/ivsr_ffmpeg_plugin/dockerfiles/rockylinux9/Dockerfile b/ivsr_ffmpeg_plugin/dockerfiles/rockylinux9/Dockerfile index ee1298b..aada420 100644 --- a/ivsr_ffmpeg_plugin/dockerfiles/rockylinux9/Dockerfile +++ b/ivsr_ffmpeg_plugin/dockerfiles/rockylinux9/Dockerfile @@ -167,7 +167,7 @@ RUN cmake .. \ #build ffmpeg with iVSR SDK backend RUN dnf -y --enablerepo=crb install nasm RUN dnf -y --enablerepo=devel install yasm -RUN dnf -y install diffutils +RUN dnf -y install diffutils patchutils # build libx264 WORKDIR ${WORKSPACE} @@ -195,16 +195,23 @@ ARG FFMPEG_IVSR_SDK_PLUGIN_DIR=${IVSR_DIR}/ivsr_ffmpeg_plugin ARG FFMPEG_DIR=${FFMPEG_IVSR_SDK_PLUGIN_DIR}/ffmpeg ARG FFMPEG_REPO=https://github.com/FFmpeg/FFmpeg.git -ARG FFMPEG_VERSION=n7.1 +ARG FFMPEG_VERSION=n8.1 WORKDIR ${FFMPEG_DIR} RUN git clone ${FFMPEG_REPO} ${FFMPEG_DIR} && \ git checkout ${FFMPEG_VERSION} COPY ./ivsr_ffmpeg_plugin/patches/*.patch ${FFMPEG_DIR}/ -RUN { set -e; \ - for patch_file in $(find -iname "*.patch" | sort -n); do \ - echo "Applying: ${patch_file}"; \ - git am --whitespace=fix ${patch_file}; \ - done; } +# Patch 0001 was written for n7.1 context; exclude the three files covered by +# the n8.1-native patch 0004, then apply 0004 for those files, and apply 0002/0003. +RUN filterdiff \ + -x '*/configure' \ + -x '*/dnn_interface.c' \ + -x '*/swscale_unscaled.c' \ + 0001-*.patch | \ + git apply --3way --ignore-whitespace - && \ + git apply --ignore-whitespace 0004-*.patch && \ + git add -A && \ + git apply --3way --whitespace=fix 0002-*.patch && \ + git apply --3way --whitespace=fix 0003-*.patch RUN ./configure \ --extra-cflags=-fopenmp \ diff --git a/ivsr_ffmpeg_plugin/dockerfiles/rockylinux9/ov2024.5s.dockerfile b/ivsr_ffmpeg_plugin/dockerfiles/rockylinux9/ov2024.5s.dockerfile index ca73eab..aec916a 100644 --- a/ivsr_ffmpeg_plugin/dockerfiles/rockylinux9/ov2024.5s.dockerfile +++ b/ivsr_ffmpeg_plugin/dockerfiles/rockylinux9/ov2024.5s.dockerfile @@ -80,7 +80,7 @@ RUN cmake .. \ #build ffmpeg with iVSR SDK backend RUN dnf -y --enablerepo=crb install nasm RUN dnf -y --enablerepo=devel install yasm -RUN dnf -y install diffutils +RUN dnf -y install diffutils patchutils # build libx264 WORKDIR ${WORKSPACE} @@ -108,16 +108,23 @@ ARG FFMPEG_IVSR_SDK_PLUGIN_DIR=${IVSR_DIR}/ivsr_ffmpeg_plugin ARG FFMPEG_DIR=${FFMPEG_IVSR_SDK_PLUGIN_DIR}/ffmpeg ARG FFMPEG_REPO=https://github.com/FFmpeg/FFmpeg.git -ARG FFMPEG_VERSION=n7.1 +ARG FFMPEG_VERSION=n8.1 WORKDIR ${FFMPEG_DIR} RUN git clone ${FFMPEG_REPO} ${FFMPEG_DIR} && \ git checkout ${FFMPEG_VERSION} COPY ./ivsr_ffmpeg_plugin/patches/*.patch ${FFMPEG_DIR}/ -RUN { set -e; \ - for patch_file in $(find -iname "*.patch" | sort -n); do \ - echo "Applying: ${patch_file}"; \ - git am --whitespace=fix ${patch_file}; \ - done; } +# Patch 0001 was written for n7.1 context; exclude the three files covered by +# the n8.1-native patch 0004, then apply 0004 for those files, and apply 0002/0003. +RUN filterdiff \ + -x '*/configure' \ + -x '*/dnn_interface.c' \ + -x '*/swscale_unscaled.c' \ + 0001-*.patch | \ + git apply --3way --ignore-whitespace - && \ + git apply --ignore-whitespace 0004-*.patch && \ + git add -A && \ + git apply --3way --whitespace=fix 0002-*.patch && \ + git apply --3way --whitespace=fix 0003-*.patch RUN ./configure \ --extra-cflags=-fopenmp \ diff --git a/ivsr_ffmpeg_plugin/dockerfiles/ubuntu22/ov2026.1.dockerfile b/ivsr_ffmpeg_plugin/dockerfiles/ubuntu22/ov2026.1.dockerfile new file mode 100644 index 0000000..a2feb5e --- /dev/null +++ b/ivsr_ffmpeg_plugin/dockerfiles/ubuntu22/ov2026.1.dockerfile @@ -0,0 +1,154 @@ +# SPDX-License-Identifier: BSD 3-Clause License +# +# Copyright (c) 2025, Intel Corporation +# All rights reserved. +# +# Redistribution and use in source and binary forms, with or without +# modification, are permitted provided that the following conditions are met: +# +# * Redistributions of source code must retain the above copyright notice, this +# list of conditions and the following disclaimer. +# +# * Redistributions in binary form must reproduce the above copyright notice, +# this list of conditions and the following disclaimer in the documentation +# and/or other materials provided with the distribution. +# +# * Neither the name of the copyright holder nor the names of its +# contributors may be used to endorse or promote products derived from +# this software without specific prior written permission. +# +# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" +# AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE +# IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE +# DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE +# FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL +# DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR +# SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER +# CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, +# OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE +# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + +# Base image +ARG IMAGE=ubuntu@sha256:ed1544e454989078f5dec1bfdabd8c5cc9c48e0705d07b678ab6ae3fb61952d2 +FROM $IMAGE AS base + +# Set non-interactive frontend for package installation +ARG DEBIAN_FRONTEND=noninteractive + +# Install essential utilities +RUN apt-get update && \ + apt-get install -y --no-install-recommends \ + curl ca-certificates gpg-agent software-properties-common && \ + rm -rf /var/lib/apt/lists/* + +# Configure shell +SHELL ["/bin/bash", "-o", "pipefail", "-c"] + +FROM base AS build +LABEL vendor="Intel Corporation" + +# Create workspace directory +RUN mkdir -p /workspace/ivsr + +ARG WORKSPACE=/workspace + +# Install common dependencies +RUN apt-get update && \ + apt-get install -y --no-install-recommends --fix-missing \ + autoconf \ + automake \ + build-essential \ + apt-utils cmake cython3 flex bison gcc g++ git make patch pkg-config wget \ + libdrm-dev libudev-dev libtool libusb-1.0-0-dev xz-utils ocl-icd-opencl-dev opencl-headers \ + && apt-get clean && rm -rf /var/lib/apt/lists/* + +# Install OpenVINO 2026.1 from apt +# Uses the unified repo (no year in path) introduced for 2026.x releases +RUN wget https://apt.repos.intel.com/intel-gpg-keys/GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB && \ + apt-key add GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB && \ + echo "deb https://apt.repos.intel.com/openvino ubuntu22 main" | tee /etc/apt/sources.list.d/intel-openvino.list && \ + apt-get update && \ + apt-get install -y openvino-2026.1.0 && \ + rm -f GPG-PUB-KEY-INTEL-SW-PRODUCTS.PUB + +# Set the working directory +WORKDIR /workspace + +# Install FFmpeg dependencies (patchutils provides filterdiff, required for n8.1 patch application) +RUN apt-get update && \ + apt-get install -y --no-install-recommends \ + ca-certificates tar g++ wget pkg-config nasm yasm libglib2.0-dev flex bison gobject-introspection libgirepository1.0-dev \ + python3-dev libx11-dev libxv-dev libxt-dev libasound2-dev libpango1.0-dev libtheora-dev libvisual-0.4-dev libgl1-mesa-dev \ + libcurl4-gnutls-dev librtmp-dev libx264-dev libx265-dev libde265-dev libva-dev libtbb-dev patchutils && \ + rm -rf /var/lib/apt/lists/* + +# Build iVSR SDK +COPY ./ivsr_sdk ${WORKSPACE}/ivsr/ivsr_sdk +WORKDIR ${WORKSPACE}/ivsr/ivsr_sdk +RUN rm -rf build && mkdir -p build && cd build && \ + cmake .. -DENABLE_LOG=OFF -DENABLE_PERF=OFF -DENABLE_THREADPROCESS=ON -DCMAKE_BUILD_TYPE=Release && \ + make -j16 && \ + make install && \ + echo "Building iVSR SDK finished." + +# Build and install FFmpeg n8.1 with libivsr support +ARG FFMPEG_REPO=https://github.com/FFmpeg/FFmpeg.git +ARG FFMPEG_VERSION=n8.1 +ARG FFMPEG_IVSR_SDK_PLUGIN_DIR=${WORKSPACE}/ivsr/ivsr_ffmpeg_plugin +WORKDIR ${FFMPEG_IVSR_SDK_PLUGIN_DIR}/ffmpeg +RUN git config --global user.email "noname@example.com" && \ + git config --global user.name "no name" && \ + git clone ${FFMPEG_REPO} . && \ + git checkout ${FFMPEG_VERSION} + +COPY ./ivsr_ffmpeg_plugin/patches/*.patch ./ +# Patch 0001 was written for n7.1 context; exclude the three files covered by the +# n8.1-native patch 0004, then apply 0004 for those files, and finally apply the +# fix-up patches 0002/0003. +RUN filterdiff \ + -x '*/configure' \ + -x '*/dnn_interface.c' \ + -x '*/swscale_unscaled.c' \ + 0001-*.patch | \ + git apply --3way --ignore-whitespace - && \ + git apply --ignore-whitespace 0004-*.patch && \ + git add -A && \ + git apply --3way --whitespace=fix 0002-*.patch && \ + git apply --3way --whitespace=fix 0003-*.patch + +RUN ./configure \ + --enable-gpl \ + --enable-nonfree \ + --disable-static \ + --disable-doc \ + --enable-shared \ + --enable-version3 \ + --enable-libivsr \ + --enable-libx264 \ + --enable-libx265 && \ + make -j16 && \ + make install + +# for GPU +RUN apt-get update && \ + apt-get install -y --no-install-recommends ocl-icd-libopencl1 && \ + apt-get clean ; \ + rm -rf /var/lib/apt/lists/* && rm -rf /tmp/* +# hadolint ignore=DL3003 +# 24.31.30508.7 +RUN mkdir /tmp/gpu_deps && cd /tmp/gpu_deps && \ + curl -L -O https://github.com/intel/intel-graphics-compiler/releases/download/igc-1.0.17384.11/intel-igc-core_1.0.17384.11_amd64.deb && \ + curl -L -O https://github.com/intel/intel-graphics-compiler/releases/download/igc-1.0.17384.11/intel-igc-opencl_1.0.17384.11_amd64.deb && \ + curl -L -O https://github.com/intel/compute-runtime/releases/download/24.31.30508.7/intel-level-zero-gpu-dbgsym_1.3.30508.7_amd64.ddeb && \ + curl -L -O https://github.com/intel/compute-runtime/releases/download/24.31.30508.7/intel-level-zero-gpu_1.3.30508.7_amd64.deb && \ + curl -L -O https://github.com/intel/compute-runtime/releases/download/24.31.30508.7/intel-opencl-icd-dbgsym_24.31.30508.7_amd64.ddeb && \ + curl -L -O https://github.com/intel/compute-runtime/releases/download/24.31.30508.7/intel-opencl-icd_24.31.30508.7_amd64.deb && \ + curl -L -O https://github.com/intel/compute-runtime/releases/download/24.31.30508.7/libigdgmm12_22.4.1_amd64.deb && \ + dpkg -i ./*.deb && rm -Rf /tmp/gpu_deps + +ENV LIBVA_DRIVER_NAME=iHD +ENV LIBVA_DRIVERS_PATH=/usr/lib/x86_64-linux-gnu/dri + +# Set the working directory back to /workspace +WORKDIR /workspace +CMD ["/bin/bash"] diff --git a/ivsr_ffmpeg_plugin/patches/0004-n8.1-compat-dnn_interface-swscale.patch b/ivsr_ffmpeg_plugin/patches/0004-n8.1-compat-dnn_interface-swscale.patch new file mode 100644 index 0000000..ec9509a --- /dev/null +++ b/ivsr_ffmpeg_plugin/patches/0004-n8.1-compat-dnn_interface-swscale.patch @@ -0,0 +1,213 @@ +diff --git a/configure b/configure +index 1759694..4b15289 100755 +--- a/configure ++++ b/configure +@@ -262,6 +262,8 @@ External library support: + --enable-libopencolorio enable color management via OpenColorIO [no] + --enable-libopenvino enable OpenVINO as a DNN module backend + for DNN based filters like dnn_processing [no] ++ --enable-libivsr enable iVSR SDK as a DNN module backend ++ for DNN based Super Resolution filters [no] + --enable-libopus enable Opus de/encoding via libopus [no] + --enable-libplacebo enable libplacebo library [no] + --enable-libpulse enable Pulseaudio input via libpulse [no] +@@ -2070,6 +2072,7 @@ EXTERNAL_LIBRARY_LIST=" + libopenjpeg + libopenmpt + libopenvino ++ libivsr + libopus + libplacebo + libpulse +@@ -3026,7 +3029,7 @@ dirac_parse_select="golomb" + dovi_rpudec_select="golomb" + dovi_rpuenc_select="golomb" + dnn_deps="avformat swscale" +-dnn_deps_any="libtensorflow libopenvino libtorch" ++dnn_deps_any="libtensorflow libopenvino libtorch libivsr" + error_resilience_select="me_cmp" + evcparse_select="golomb" + faandct_deps="faan" +@@ -7347,6 +7350,7 @@ enabled libopenmpt && require_pkg_config libopenmpt "libopenmpt >= 0.2.65 + enabled libopenvino && { { check_pkg_config libopenvino openvino openvino/c/openvino.h ov_core_create && enable openvino2; } || + { check_pkg_config libopenvino openvino c_api/ie_c_api.h ie_c_api_version || + require libopenvino c_api/ie_c_api.h ie_c_api_version -linference_engine_c_api; } } ++enabled libivsr && require libivsr ivsr.h ivsr_init -livsr + enabled libopus && { + enabled libopus_decoder && { + require_pkg_config libopus opus opus_multistream.h opus_multistream_decoder_create +diff --git a/libavfilter/dnn/dnn_interface.c b/libavfilter/dnn/dnn_interface.c +index 7080ab1..6d47920 100644 +--- a/libavfilter/dnn/dnn_interface.c ++++ b/libavfilter/dnn/dnn_interface.c +@@ -33,6 +33,7 @@ + extern const DNNModule ff_dnn_backend_openvino; + extern const DNNModule ff_dnn_backend_tf; + extern const DNNModule ff_dnn_backend_torch; ++extern const DNNModule ff_dnn_backend_ivsr; + + #define OFFSET(x) offsetof(DnnContext, x) + #define FLAGS AV_OPT_FLAG_FILTERING_PARAM +@@ -78,6 +79,9 @@ static const DnnBackendInfo dnn_backend_info_list[] = { + #if CONFIG_LIBTORCH + {offsetof(DnnContext, torch_option), .module = &ff_dnn_backend_torch}, + #endif ++#if CONFIG_LIBIVSR ++ {offsetof(DnnContext, ivsr_option), .module = &ff_dnn_backend_ivsr}, ++#endif + }; + + const DNNModule *ff_get_dnn_module(DNNBackendType backend_type, void *log_ctx) +@@ -103,6 +107,14 @@ void ff_dnn_init_child_class(DnnContext *ctx) + } + } + ++void ff_dnn_uninit_child_class(DnnContext *ctx) ++{ ++ for (int i = 0; i < FF_ARRAY_ELEMS(dnn_backend_info_list); i++) { ++ const AVClass **ptr = (const AVClass **) ((char *) ctx + dnn_backend_info_list[i].offset); ++ av_opt_free(ptr); ++ } ++} ++ + void *ff_dnn_child_next(DnnContext *obj, void *prev) { + size_t pre_offset; + +diff --git a/libswscale/swscale_unscaled.c b/libswscale/swscale_unscaled.c +index f612f88..d61f7c8 100644 +--- a/libswscale/swscale_unscaled.c ++++ b/libswscale/swscale_unscaled.c +@@ -2121,6 +2121,106 @@ static int float_y_to_uint_y_wrapper(SwsInternal *c, + return srcSliceH; + } + ++static int uint16_y_to_float_y_wrapper(SwsInternal *c, ++ const uint8_t *const src[], ++ const int srcStride[], int srcSliceY, ++ int srcSliceH, uint8_t *const dst[], ++ const int dstStride[]) ++{ ++ int y, x; ++ const ptrdiff_t srcStrideUint16 = srcStride[0] >> 1; ++ const ptrdiff_t dstStrideFloat = dstStride[0] >> 2; ++ const uint16_t *srcPtr = (const uint16_t *)(src[0] + srcStride[0] * srcSliceY); ++ float *dstPtr = (float *)(dst[0] + dstStride[0] * srcSliceY); ++ const float float_norm_factor = 1.0f / 65535.0f; ++ ++ for (y = 0; y < srcSliceH; ++y) { ++ for (x = 0; x < c->opts.src_w; ++x) { ++ dstPtr[x] = (float)srcPtr[x] * float_norm_factor; ++ } ++ srcPtr += srcStrideUint16; ++ dstPtr += dstStrideFloat; ++ } ++ ++ return srcSliceH; ++} ++ ++static int float_y_to_uint16_y_wrapper(SwsInternal *c, ++ const uint8_t *const src[], ++ const int srcStride[], int srcSliceY, ++ int srcSliceH, uint8_t *const dst[], ++ const int dstStride[]) ++{ ++ int y, x; ++ const ptrdiff_t srcStrideFloat = srcStride[0] >> 2; ++ const ptrdiff_t dstStrideUint16 = dstStride[0] >> 1; ++ const float *srcPtr = (const float *)(src[0] + srcStride[0] * srcSliceY); ++ uint16_t *dstPtr = (uint16_t *)(dst[0] + dstStride[0] * srcSliceY); ++ ++ for (y = 0; y < srcSliceH; ++y) { ++ for (x = 0; x < c->opts.src_w; ++x) { ++ dstPtr[x] = av_clip_uint16(lrintf(65535.0f * srcPtr[x])); ++ } ++ srcPtr += srcStrideFloat; ++ dstPtr += dstStrideUint16; ++ } ++ ++ return srcSliceH; ++} ++ ++static int uint10_y_to_float_y_wrapper(SwsInternal *c, ++ const uint8_t *const src[], ++ const int srcStride[], int srcSliceY, ++ int srcSliceH, uint8_t *const dst[], ++ const int dstStride[]) ++{ ++ int y, x; ++ const ptrdiff_t srcStrideUint16 = srcStride[0] >> 1; ++ const ptrdiff_t dstStrideFloat = dstStride[0] >> 2; ++ const uint16_t *srcPtr = (const uint16_t *)(src[0] + srcStride[0] * srcSliceY); ++ float *dstPtr = (float *)(dst[0] + dstStride[0] * srcSliceY); ++ const float float_norm_factor = 1.0f / 1023.0f; ++ ++ for (y = 0; y < srcSliceH; ++y) { ++ for (x = 0; x < c->opts.src_w; ++x) { ++ dstPtr[x] = (float)srcPtr[x] * float_norm_factor; ++ } ++ srcPtr += srcStrideUint16; ++ dstPtr += dstStrideFloat; ++ } ++ ++ return srcSliceH; ++} ++ ++static int float_y_to_uint10_y_wrapper(SwsInternal *c, ++ const uint8_t *const src[], ++ const int srcStride[], int srcSliceY, ++ int srcSliceH, uint8_t *const dst[], ++ const int dstStride[]) ++{ ++ int y, x; ++ const ptrdiff_t srcStrideFloat = srcStride[0] >> 2; ++ const ptrdiff_t dstStrideUint16 = dstStride[0] >> 1; ++ const float *srcPtr = (const float *)(src[0] + srcStride[0] * srcSliceY); ++ uint16_t *dstPtr = (uint16_t *)(dst[0] + dstStride[0] * srcSliceY); ++ ++ for (y = 0; y < srcSliceH; ++y) { ++ for (x = 0; x < c->opts.src_w; ++x) { ++ int value = lrintf(1023.0f * srcPtr[x]); ++ if (value < 0) { ++ value = 0; ++ } else if (value > 1023) { ++ value = 1023; ++ } ++ dstPtr[x] = (uint16_t)value; ++ } ++ srcPtr += srcStrideFloat; ++ dstPtr += dstStrideUint16; ++ } ++ ++ return srcSliceH; ++} ++ + /* unscaled copy like stuff (assumes nearly identical formats) */ + static int packedCopyWrapper(SwsInternal *c, const uint8_t *const src[], + const int srcStride[], int srcSliceY, int srcSliceH, +@@ -2631,6 +2731,26 @@ void ff_get_unscaled_swscale(SwsInternal *c) + c->convert_unscaled = float_y_to_uint_y_wrapper; + } + ++ /* 16bit Y to float Y */ ++ if (srcFormat == AV_PIX_FMT_GRAY16 && dstFormat == AV_PIX_FMT_GRAYF32){ ++ c->convert_unscaled = uint16_y_to_float_y_wrapper; ++ } ++ ++ /* float Y to 16bit Y */ ++ if (srcFormat == AV_PIX_FMT_GRAYF32 && dstFormat == AV_PIX_FMT_GRAY16){ ++ c->convert_unscaled = float_y_to_uint16_y_wrapper; ++ } ++ ++ /* 10bit Y to float Y */ ++ if (srcFormat == AV_PIX_FMT_GRAY10 && dstFormat == AV_PIX_FMT_GRAYF32){ ++ c->convert_unscaled = uint10_y_to_float_y_wrapper; ++ } ++ ++ /* float Y to 10bit Y */ ++ if (srcFormat == AV_PIX_FMT_GRAYF32 && dstFormat == AV_PIX_FMT_GRAY10){ ++ c->convert_unscaled = float_y_to_uint10_y_wrapper; ++ } ++ + /* LQ converters if -sws 0 or -sws 4*/ + if (c->opts.flags&(SWS_FAST_BILINEAR|SWS_POINT)) { + /* yv12_to_yuy2 */ diff --git a/ivsr_sdk/CMakeLists.txt b/ivsr_sdk/CMakeLists.txt index 2deecef..efb9688 100644 --- a/ivsr_sdk/CMakeLists.txt +++ b/ivsr_sdk/CMakeLists.txt @@ -15,6 +15,8 @@ SET(CMAKE_CXX_FLAGS_RELEASE "$ENV{CXXFLAGS} -O3 -Wall") set(SDK_PRIVATE_HEADERS ${CMAKE_CURRENT_SOURCE_DIR}/privates/include) +option(ENABLE_IRGUARD "Enable irguard model protection (requires prebuilt libirguard.a)" OFF) + add_subdirectory(src) add_subdirectory(privates) diff --git a/ivsr_sdk/privates/CMakeLists.txt b/ivsr_sdk/privates/CMakeLists.txt index c84e37a..b55dfdf 100644 --- a/ivsr_sdk/privates/CMakeLists.txt +++ b/ivsr_sdk/privates/CMakeLists.txt @@ -6,8 +6,10 @@ cmake_minimum_required(VERSION 3.10) option(REBUILD_IRGUARD "Enables cpu id checking" OFF) -if(REBUILD_IRGUARD) - add_subdirectory(model_guard) -else() - add_subdirectory(model_guard.bin) +if(ENABLE_IRGUARD) + if(REBUILD_IRGUARD) + add_subdirectory(model_guard) + else() + add_subdirectory(model_guard.bin) + endif() endif() diff --git a/ivsr_sdk/src/CMakeLists.txt b/ivsr_sdk/src/CMakeLists.txt index 432a13f..4ac7093 100644 --- a/ivsr_sdk/src/CMakeLists.txt +++ b/ivsr_sdk/src/CMakeLists.txt @@ -38,9 +38,11 @@ target_link_libraries(${TARGET_NAME} PRIVATE openvino::runtime) find_package(OpenMP REQUIRED) target_link_libraries(${TARGET_NAME} PRIVATE OpenMP::OpenMP_CXX) -if (REBUILD_IRGUARD) - target_link_libraries(${TARGET_NAME} PRIVATE irguard z) -else() - target_link_libraries(${TARGET_NAME} PRIVATE ${CMAKE_LIBRARY_OUTPUT_DIRECTORY}/libirguard.a z) +if (ENABLE_IRGUARD) + if (REBUILD_IRGUARD) + target_link_libraries(${TARGET_NAME} PRIVATE irguard z) + else() + target_link_libraries(${TARGET_NAME} PRIVATE ${CMAKE_LIBRARY_OUTPUT_DIRECTORY}/libirguard.a z) + endif() endif() diff --git a/ivsr_sdk/src/ov_engine.cpp b/ivsr_sdk/src/ov_engine.cpp index 95f36d5..7f9fa76 100644 --- a/ivsr_sdk/src/ov_engine.cpp +++ b/ivsr_sdk/src/ov_engine.cpp @@ -13,7 +13,9 @@ #include