大神们，能通过sample_venc只编码一帧YUV成jpeg么，conf文件应该怎么配？

@lijinlin001 uboot的dts当时没太留意，功能开启是在uboot/common/main.c里配置的，大概是这样：

#include <pwm.h>

void pwm_init(void)
{
	int pwm1;
	pwm1 = pwm_request(1, "pwm1");
    	pwm_config(pwm1, 500000, 500000);
    	pwm_set_polarity(pwm1, PWM_POLARITY_INVERSED);
    	pwm_enable(pwm1);
}

具体的接口函数用法看uboot的pwm.h

有没有只编一帧YUV成jpeg的demo啊

为什么SDK make menuconfig的mpp中，没有enable ai/ao component的选项？如下图

正常应该是这样的，如下图：

被隐藏了？怎么打开

@lingzhi12321 你要用跑仿真推理？

@whycanservice ,了解，谢啦~

@WhycanService 用acuity-toolkit-binary-6.12.0和VivanteIDE5.7.2工具，不管怎么量化，最终运行都是下面的打印

@whycanservice 请问怎么用Verisilicon IDE查看模型呢？在其菜单栏上没找到相关的，硬打开也是当二进制文件处理。我想看模型的输入输出信息。因为跑应用的时候，报了以下错误：

用旧版的工具量化的模型是正常的，同样的指令用新工具量化的模型，运行时就报这个问题，难道改了api？

xxxx@server:~/workspace/v853_quantification/kl_yolo_v3m_int8/yun_output_nbg_unify_test$ nbinfo -out network_binary.nb

Verisilicon NBInfo version 1.2.4, version=0x00010204

nbg file name network_binary.nb
not support NBG format version=0x10016, support version=0x10014
NBG parser is forward compatible with NBG format, please update NBG parser in time
fail to read NBG header
failed to read fix section data
failed to initialize nbg parser

更新了acuity-toolkit和VivanteIDE工具后，用nbinfo工具查看nb文件就报版本不匹配，请问在哪可以下载最新版本的nbinfo工具？

@daizebin 安装后可以编译了。
但报了新的问题

arch/arm/cpu/armv7/built-in.o: In function `cleanup_before_linux_select':
/home/data1/xxxx/workspace/tina-v853/lichee/brandy-2.0/u-boot-2018/arch/arm/cpu/armv7/cpu.c:43: undefined reference to `dcache_disable'
/home/data1/xxxx/workspace/tina-v853/lichee/brandy-2.0/u-boot-2018/arch/arm/cpu/armv7/cpu.c:58: undefined reference to `icache_disable'
/home/data1/xxxx/workspace/tina-v853/lichee/brandy-2.0/u-boot-2018/arch/arm/cpu/armv7/cpu.c:64: undefined reference to `icache_disable'
/home/data1/xxxx/workspace/tina-v853/lichee/brandy-2.0/u-boot-2018/arch/arm/cpu/armv7/cpu.c:69: undefined reference to `icache_enable'
arch/arm/mach-sunxi/built-in.o: In function `enable_caches':
/home/data1/xxxx/workspace/tina-v853/lichee/brandy-2.0/u-boot-2018/arch/arm/mach-sunxi/board.c:170: undefined reference to `dcache_enable'
cmd/built-in.o: In function `do_sprite_test':
/home/data1/xxxx/workspace/tina-v853/lichee/brandy-2.0/u-boot-2018/cmd/sunxi_sprite.c:74: undefined reference to `dcache_disable'
./../tools/toolchain/gcc-linaro-7.2.1-2017.11-x86_64_arm-linux-gnueabi/bin/arm-linux-gnueabi-ld.bfd: BFD (Linaro_Binutils-2017.11) 2.28.2.20170706 assertion fail /home/tcwg-buildslave/workspace/tcwg-make-release/builder_arch/amd64/label/tcwg-x86_64-build/target/arm-linux-gnueabi/snapshots/binutils-gdb.git~users~linaro~binutils-2_28-branch/bfd/elf32-arm.c:9514
./../tools/toolchain/gcc-linaro-7.2.1-2017.11-x86_64_arm-linux-gnueabi/bin/arm-linux-gnueabi-ld.bfd: BFD (Linaro_Binutils-2017.11) 2.28.2.20170706 assertion fail /home/tcwg-buildslave/workspace/tcwg-make-release/builder_arch/amd64/label/tcwg-x86_64-build/target/arm-linux-gnueabi/snapshots/binutils-gdb.git~users~linaro~binutils-2_28-branch/bfd/elf32-arm.c:9514
./../tools/toolchain/gcc-linaro-7.2.1-2017.11-x86_64_arm-linux-gnueabi/bin/arm-linux-gnueabi-ld.bfd: BFD (Linaro_Binutils-2017.11) 2.28.2.20170706 assertion fail /home/tcwg-buildslave/workspace/tcwg-make-release/builder_arch/amd64/label/tcwg-x86_64-build/target/arm-linux-gnueabi/snapshots/binutils-gdb.git~users~linaro~binutils-2_28-branch/bfd/elf32-arm.c:9514
./../tools/toolchain/gcc-linaro-7.2.1-2017.11-x86_64_arm-linux-gnueabi/bin/arm-linux-gnueabi-ld.bfd: BFD (Linaro_Binutils-2017.11) 2.28.2.20170706 assertion fail /home/tcwg-buildslave/workspace/tcwg-make-release/builder_arch/amd64/label/tcwg-x86_64-build/target/arm-linux-gnueabi/snapshots/binutils-gdb.git~users~linaro~binutils-2_28-branch/bfd/elf32-arm.c:9514
Makefile:1496: recipe for target 'u-boot' failed
make: *** [u-boot] Error 1
muboot stop for build error in brandy, Please check!

xxxx@server:~/workspace/tina-v85x_1.1$ mboot
v853_vision v853 v853-vision
build_boot platform:sun8iw21p1 o_option:uboot
grep: /home/data1/xxxx/workspace/tina-v85x_1.1/lichee/brandy-2.0/spl/Makefile: No such file or directory
Prepare toolchain ...
build for sun8iw21p1_defconfig ...
cat: .tmp_config_from_defconfig.o.md5sum: No such file or directory
md5sum: .config: No such file or directory
  CLEAN   scripts/basic
  HOSTCC  scripts/basic/fixdep
  HOSTCC  scripts/kconfig/conf.o
  YACC    scripts/kconfig/zconf.tab.c
/bin/sh: 1: bison: not found
scripts/Makefile.lib:228: recipe for target 'scripts/kconfig/zconf.tab.c' failed
make[1]: *** [scripts/kconfig/zconf.tab.c] Error 127
Makefile:547: recipe for target 'sun8iw21p1_defconfig' failed
make: *** [sun8iw21p1_defconfig] Error 2
muboot stop for build error in brandy, Please check!

这是V1.1版本的SDK，不确定之前V1.0是否就有这个问题

@whycanservice 量化工具好像没有适配cuda，用不了GPU

pegasus quantize --model yolo.json --model-data yuntai.data --MLE --algorithm kl_divergence --batch-size 1 --iterations 1 --divergence-nbins 4096 --device CPU --with-input-meta ./yolo_inputmeta.yml --rebuild --model-quantize yolo.quantize --quantizer asymmetric_affine --qtype uint8 

怎么加了--MLE（最小化每层误差）参数后，量化的时候超级久，batch-size为1时差不多要2个小时， 再设大点好几天都没量化完，哪位大神了解嘛？