Skip to main content
  1. Planet
  2. Global
  3. Page 157

the avatar of Stefan's openSUSE Blog
the avatar of Stefan's openSUSE Blog

How to install SLE-15-SP6 on NVIDIA Jetson platform (Jetson AGX Orin/IGX Orin)

This covers the installation of updated Kernel, out-of-tree nvidia kernel modules package, how to get GNOME desktop running and installation/run of glmark2 benchmark. Also it describes how to get some CUDA and TensorRT samples running.

SP6

Download SLE-15-SP6 (Arm) installation image. This you can put on a regular USB stick or on an SD card using dd command.

Boot from the USB stick/SD card, that you wrote above and install SP6. You can install via serial console or connect a monitor to the display port.

When using a connected monitor for installation

This needs for the installation a special setting in the Firmware of the machine.

--> UEFI Firmware Settings
 --> Device Manager
  --> NVIDIA Configuration
   --> Boot Configuration
    --> SOC Display Hand-Off Mode <Always>

This setting for SOC Display Hand-Off Mode will change automatically to Never later with the installation of the graphics driver.

Installation

Once grub starts you need to edit the grub entry Installation. Press e for doing this and add console=tty0 exec="date -s 2025-01-27" (when using a connected monitor for intallation) or exec="date -s 2025-01-27" (when installing on a serial console) to the linux [...] line. Replace 2025-01-27 with the current date.

### When using a connected monitor for intallation
[...]
linux /boot/aarch64/linux splash=silent console=tty0 exec="date -s 2025-01-27"
[...]
### When installing on a serial console
[...]
linux /boot/aarch64/linux splash=silent exec="date -s 2025-01-27"
[...]

The reason for this is that during installation the driver nvvrs-pseq-rtc for the battery-backed RTC0 (Real Time Clock) is not yet available and therefore the non-battery-backed RTC1 is used, which doesn’t have the correct time set during installation. So this is a workaround to avoid a product registration failure later due to a certificate, which is not valid yet.

Then press F10 to continue to boot.

Make sure you select the following modules during installation:

  • Basesystem (enough for just installing the kernel driver)
  • Containers (needed for podman for CUDA libraries)
  • Desktop Applications (needed for running a desktop)
  • Development Tools (needed for git for CUDA samples)

Select SLES with GNOME for installation.

In Clock and Time Zone dialogue chose Other Settings to open Change Date and Time dialogue. There enable Synchronize with NTP Server.

--> Clock and Time Zone dialogue
 --> Other Settings
  --> Change Date and Time dialogue
   --> (x) Synchronize with NTP Server

Kernel + KMP drivers

After installation update kernel and install our KMP (kernel module package) for all nvidia kernel modules.

We plan to make the KMP available as a driver kit via the SolidDriver Program. For now please install an updated kernel and the KMP after checking the build status (type ‘jetson’ in Search… field; rebuilding can take a few hours!) from our open buildservice:

# flavor either default or 64kb (check with `uname -r` command)
sudo zypper up kernel-<flavor>
sudo zypper ar https://download.opensuse.org/repositories/X11:/XOrg/SLE_15_SP6/  jetson-kmp
sudo zypper ref
sudo zypper in -r jetson-kmp nvidia-jetson-36_4-kmp-<flavor> kernel-firmware-nvidia-jetson-36_4

Userspace/Desktop

Unfortunately installing the userspace is a non-trivial task.

Installation

Download Driver Package (BSP) from this location. Extract Jetson_Linux_R36.4.0_aarch64.tbz2.

tar xf Jetson_Linux_R36.4.0_aarch64.tbz2

Then you need to convert debian packages from this content into tarballs.

pushd Linux_for_Tegra
sed -i -e 's/lbzip2/bzip2/g' -e 's/-I zstd //g' nv_tools/scripts/nv_repackager.sh
./nv_tools/scripts/nv_repackager.sh -o ./nv_tegra/l4t_tar_packages --convert-all
popd

From the generated tarballs you only need these:

nvidia-l4t-3d-core_36.4.0-20240912212859_arm64.tbz2
nvidia-l4t-camera_36.4.0-20240912212859_arm64.tbz2
nvidia-l4t-core_36.4.0-20240912212859_arm64.tbz2
nvidia-l4t-cuda_36.4.0-20240912212859_arm64.tbz2
nvidia-l4t-gbm_36.4.0-20240912212859_arm64.tbz2
nvidia-l4t-multimedia-utils_36.4.0-20240912212859_arm64.tbz2
nvidia-l4t-multimedia_36.4.0-20240912212859_arm64.tbz2
nvidia-l4t-nvfancontrol_36.4.0-20240912212859_arm64.tbz2
nvidia-l4t-nvml_36.4.0-20240912212859_arm64.tbz2
nvidia-l4t-nvpmodel_36.4.0-20240912212859_arm64.tbz2
nvidia-l4t-nvsci_36.4.0-20240912212859_arm64.tbz2
nvidia-l4t-pva_36.4.0-20240912212859_arm64.tbz2
nvidia-l4t-tools_36.4.0-20240912212859_arm64.tbz2
nvidia-l4t-vulkan-sc-sdk_36.4.0-20240912212859_arm64.tbz2
nvidia-l4t-wayland_36.4.0-20240912212859_arm64.tbz2
nvidia-l4t-x11_36.4.0-20240912212859_arm64.tbz2
nvidia-l4t-nvml_36.4.0-20240912212859_arm64.tbz2

And from this tarball nvidia-l4t-init_36.4.0-20240912212859_arm64.tbz2 you only need these files:

etc/asound.conf.tegra-ape
etc/asound.conf.tegra-hda-jetson-agx
etc/asound.conf.tegra-hda-jetson-xnx
etc/nvidia-container-runtime/host-files-for-container.d/devices.csv
etc/nvidia-container-runtime/host-files-for-container.d/drivers.csv
etc/nvsciipc.cfg
etc/sysctl.d/60-nvsciipc.conf
etc/systemd/nv_nvsciipc_init.sh
etc/systemd/nvpower.sh
etc/systemd/nv.sh
etc/systemd/system.conf.d/watchdog.conf
etc/systemd/system/multi-user.target.wants/nv_nvsciipc_init.service
etc/systemd/system/multi-user.target.wants/nvpower.service
etc/systemd/system/multi-user.target.wants/nv.service
etc/systemd/system/nv_nvsciipc_init.service
etc/systemd/system/nvpower.service
etc/systemd/system/nv.service
etc/udev/rules.d/99-tegra-devices.rules
usr/share/alsa/cards/tegra-ape.conf
usr/share/alsa/cards/tegra-hda.conf
usr/share/alsa/init/postinit/00-tegra.conf
usr/share/alsa/init/postinit/01-tegra-rt565x.conf
usr/share/alsa/init/postinit/02-tegra-rt5640.conf

So first let’s repackage nvidia-l4t-init_36.4.0-20240912212859_arm64.tbz2:

pushd Linux_for_Tegra/nv_tegra/l4t_tar_packages/
cat > nvidia-l4t-init.txt << EOF
etc/asound.conf.tegra-ape
etc/asound.conf.tegra-hda-jetson-agx
etc/asound.conf.tegra-hda-jetson-xnx
etc/nvidia-container-runtime/host-files-for-container.d/devices.csv
etc/nvidia-container-runtime/host-files-for-container.d/drivers.csv
etc/nvsciipc.cfg
etc/sysctl.d/60-nvsciipc.conf
etc/systemd/nv_nvsciipc_init.sh
etc/systemd/nvpower.sh
etc/systemd/nv.sh
etc/systemd/system.conf.d/watchdog.conf
etc/systemd/system/multi-user.target.wants/nv_nvsciipc_init.service
etc/systemd/system/multi-user.target.wants/nvpower.service
etc/systemd/system/multi-user.target.wants/nv.service
etc/systemd/system/nv_nvsciipc_init.service
etc/systemd/system/nvpower.service
etc/systemd/system/nv.service
etc/udev/rules.d/99-tegra-devices.rules
usr/share/alsa/cards/tegra-ape.conf
usr/share/alsa/cards/tegra-hda.conf
usr/share/alsa/init/postinit/00-tegra.conf
usr/share/alsa/init/postinit/01-tegra-rt565x.conf
usr/share/alsa/init/postinit/02-tegra-rt5640.conf
EOF
tar xf nvidia-l4t-init_36.4.0-20240912212859_arm64.tbz2
rm nvidia-l4t-init_36.4.0-20240912212859_arm64.tbz2
tar cjf nvidia-l4t-init_36.4.0-20240912212859_arm64.tbz2 $(cat nvidia-l4t-init.txt)
popd

On IGX Orin platform with dedicated graphics card (dGPU systems) you need to get rid of some files due to conflicts with dGPU userspace drivers.

# repackage nvidia-l4t-x11_ package
tar tf nvidia-l4t-x11_36.4.0-20240912212859_arm64.tbz2 | grep -v /usr/bin/nvidia-xconfig \
  > nvidia-l4t-x11_36.4.0-20240912212859.txt
tar xf  nvidia-l4t-x11_36.4.0-20240912212859_arm64.tbz2
rm      nvidia-l4t-x11_36.4.0-20240912212859_arm64.tbz2
tar cjf nvidia-l4t-x11_36.4.0-20240912212859_arm64.tbz2 $(cat nvidia-l4t-x11_36.4.0-20240912212859.txt)

# repackage nvidia-l4t-3d-core_ package
tar tf nvidia-l4t-3d-core_36.4.0-20240912212859_arm64.tbz2 | \
  grep -v \
       -e /etc/vulkan/icd.d/nvidia_icd.json \
       -e /usr/lib/xorg/modules/drivers/nvidia_drv.so \
       -e /usr/lib/xorg/modules/extensions/libglxserver_nvidia.so \
       -e /usr/share/glvnd/egl_vendor.d/10_nvidia.json \
       > nvidia-l4t-3d-core_36.4.0-20240912212859.txt
tar xf  nvidia-l4t-3d-core_36.4.0-20240912212859_arm64.tbz2
rm      nvidia-l4t-3d-core_36.4.0-20240912212859_arm64.tbz2
tar cjf nvidia-l4t-3d-core_36.4.0-20240912212859_arm64.tbz2 $(cat nvidia-l4t-3d-core_36.4.0-20240912212859.txt)

Then extract the generated tarballs to your system.

pushd Linux_for_Tegra/nv_tegra/l4t_tar_packages
for i in \
nvidia-l4t-core_36.4.0-20240912212859_arm64.tbz2 \
nvidia-l4t-3d-core_36.4.0-20240912212859_arm64.tbz2 \
nvidia-l4t-cuda_36.4.0-20240912212859_arm64.tbz2 \
nvidia-l4t-gbm_36.4.0-20240912212859_arm64.tbz2 \
nvidia-l4t-multimedia-utils_36.4.0-20240912212859_arm64.tbz2 \
nvidia-l4t-multimedia_36.4.0-20240912212859_arm64.tbz2 \
nvidia-l4t-nvfancontrol_36.4.0-20240912212859_arm64.tbz2 \
nvidia-l4t-nvpmodel_36.4.0-20240912212859_arm64.tbz2 \
nvidia-l4t-tools_36.4.0-20240912212859_arm64.tbz2 \
nvidia-l4t-x11_36.4.0-20240912212859_arm64.tbz2 \
nvidia-l4t-nvsci_36.4.0-20240912212859_arm64.tbz2 \
nvidia-l4t-pva_36.4.0-20240912212859_arm64.tbz2 \
nvidia-l4t-wayland_36.4.0-20240912212859_arm64.tbz2 \
nvidia-l4t-camera_36.4.0-20240912212859_arm64.tbz2 \
nvidia-l4t-vulkan-sc-sdk_36.4.0-20240912212859_arm64.tbz2 \
nvidia-l4t-nvml_36.4.0-20240912212859_arm64.tbz2 \
nvidia-l4t-init_36.4.0-20240912212859_arm64.tbz2; do
  sudo tar xjf $i -C /
done
popd

On systems without dedicated graphics (internal GPU systems) card you still need to move

/usr/lib/xorg/modules/drivers/nvidia_drv.so
/usr/lib/xorg/modules/extensions/libglxserver_nvidia.so

to

/usr/lib64/xorg/modules/drivers/nvidia_drv.so
/usr/lib64/xorg/modules/extensions/libglxserver_nvidia.so

So let’s do this.

sudo mv /usr/lib/xorg/modules/drivers/nvidia_drv.so \
          /usr/lib64/xorg/modules/drivers/
sudo mv /usr/lib/xorg/modules/extensions/libglxserver_nvidia.so \
          /usr/lib64/xorg/modules/extensions/
sudo rm -rf /usr/lib/xorg

Then add /usr/lib/aarch64-linux-gnu and /usr/lib/aarch64-linux-gnu/tegra-egl to /etc/ld.so.conf.d/nvidia-tegra.conf.

echo /usr/lib/aarch64-linux-gnu | sudo tee -a /etc/ld.so.conf.d/nvidia-tegra.conf
echo /usr/lib/aarch64-linux-gnu/tegra-egl | sudo tee -a /etc/ld.so.conf.d/nvidia-tegra.conf

Run ldconfig 

sudo ldconfig

Video group for regular users

A regular user needs to be added to the group video to be able to log in to the GNOME desktop as regular user. This can be achieved by using YaST, usermod or editing /etc/group manually.

Reboot the machine with the previously updated kernel

sudo reboot

In Mokmanager (Perform MOK management) select Continue boot. Although Secureboot is enabled by default in BIOS it seems it hasn’t been implemented yet (BIOS from 04/04/2024). Select first entry SLES 15-SP6 for booting.

Basic testing

First basic testing will be running nvidia-smi. 

sudo nvidia-smi

Graphical desktop (GNOME) should work as well. Unfortunately Linux console is not available. Use either a serial console or a ssh connection if you don’t want to use the graphical desktop or need remote access to the system.

glmark2

Install phoronix-test-suite

sudo zypper ar https://cdn.opensuse.org/distribution/leap/15.6/repo/oss/ repo-oss
sudo zypper ref
sudo zypper in phoronix-test-suite

Run phoronix-test-suite

sudo zypper in gcc gcc-c++
phoronix-test-suite benchmark glmark2

CUDA/Tensorflow

Containers

NVIDIA provides containers available for Jetson that include SDKs such as CUDA. More details here. These containers are Ubuntu based, but can be used from SLE as well. You need to install the NVIDIA container runtime for this. Detailed information here.

1. Install podman and nvidia-container-runtime
sudo zypper install podman
sudo zypper ar https://nvidia.github.io/libnvidia-container/stable/rpm/nvidia-container-toolkit.repo
sudo zypper modifyrepo --enable nvidia-container-toolkit-experimental
sudo zypper --gpg-auto-import-keys install -y nvidia-container-toolkit
sudo nvidia-ctk cdi generate --mode=csv --output=/var/run/cdi/nvidia.yaml
sudo nvidia-ctk cdi list
2. Download the CUDA samples
sudo zypper install git
cd
git clone https://github.com/NVIDIA/cuda-samples.git
cd cuda-samples
git checkout v12.4
3. Start X
sudo rcxdm stop
sudo Xorg -retro &> /tmp/log &
export DISPLAY=:0
xterm &

Monitor should now show a Moiree pattern with an unframed xterm on it. Otherwise check /tmp/log.

4. Download and run the JetPack6 container
sudo podman run --rm -it -e DISPLAY --net=host --device nvidia.com/gpu=all --group-add keep-groups --security-opt label=disable -v $HOME/cuda-samples:/cuda-samples nvcr.io/nvidia/l4t-jetpack:r36.2.0 /bin/bash

CUDA

5. Build and run the samples in the container
cd /cuda-samples
make -j$(nproc)
pushd ./Samples/5_Domain_Specific/nbody
make
popd
./bin/aarch64/linux/release/deviceQuery
./bin/aarch64/linux/release/nbody

Tensorrt

6. Build and run Tensorrt in the container

This is both with the GPU and DLA (deep-learning accelerator).

cd /usr/src/tensorrt/samples/
make -j$(nproc)
cd ..
./bin/sample_algorithm_selector
./bin/sample_onnx_mnist
./bin/sample_onnx_mnist --useDLACore=0
./bin/sample_onnx_mnist --useDLACore=1

Misc

Performance

You can improve the performance by giving the clock a boost. For best performance you can run jetson_clocks to set the device to max clock settings

sudo jetson_clocks --show
sudo jetson_clocks
sudo jetson_clocks --show

The 1st and 3rd command just prints the clock settings.

MaxN Power

For maximum performance you also need to set MaxN Power. This can be done by running

sudo nvpmodel -m 0

Afterwards you need to reboot the system though.

sudo reboot

In order to check for the current value run

sudo nvpmodel -q

a silhouette of a person's head and shoulders, used as a default avatar
a silhouette of a person's head and shoulders, used as a default avatar

Iconos Nature para tu PC, simples y elegantes

Hago una descanso en la promoción de Akademy-es 2024 de València y os traigo una entrada ligera. Os presento el tema de iconos Nature para tu PC, una colección de avatares para aplicaciones, acciones y otros elementos gráficos que destacan por su sencillez y elegancia.

Iconos Nature para tu PC, simples y preciosos

Como he dicho muchas veces, me fascina la variedad que tenemos a nuestra disposición tanto de forma, estilo o colores. Tenemos iconos clásicos, minimalistas, lineales, 3D, que simulan otros sistemas operativos, imaginativos, que recuerdan a objetos cotidianeos, etc.

Hoy os presento un pack de iconos creados por Thalic que ha nombrado como Nature. Se trata de unos iconos minimalistas, simples y de un bonoto color rojo pastel, que quedan muy elegantes en un entorno oscuro.

Iconos Nature para tu PC, simples y elegantes

En palabras de su creador:

Este es mi primer intento de hacer un tema de iconos para KDE ya que normalmente lo hago para DE’s basados en GTK. Pero me he pasado a KDE en todos mis ordenadores, así que necesito algunos temas de iconos que me gusten y la forma de hacerlo es construirlos uno mismo.

El tema de iconos de la naturaleza está todavía en una etapa temprana, así que no dudes en reportar errores o hacer sugerencias y peticiones, para que pueda evolucionar el tema de iconos Nature al máximo. Voy a añadir más colores para elegir, pero empiezo con el color rojo.

Iconos Nature para tu PC, simples y elegantes

Y como siempre digo, si os gusta el pack de iconos podéis pagarlo de muchas formas en la página en continua evolución de KDE Store, que estoy seguro que el desarrollador lo agradecer: puntúale positivamente, hazle un comentario en la página o realiza una donación. Ayudar al desarrollo del Software Libre también se hace simplemente dando las gracias, ayuda mucho más de lo que os podéis imaginar, recordad la campaña I love Free Software Day 2017 de la Free Software Foundation donde se nos recordaba esta forma tan sencilla de colaborar con el gran proyecto del Software Libre y que en el blog dedicamos un artículo.

Más información: KDE Store

La entrada Iconos Nature para tu PC, simples y elegantes se publicó primero en KDE Blog.

a silhouette of a person's head and shoulders, used as a default avatar
a silhouette of a person's head and shoulders, used as a default avatar

Lista de reproducción con las emisoras de somaGM

Quiero compartir en el blog una lista de reproducción actualizada que he recopilado con las emisoras disponibles de la radio online SomaFM

Foto: Ron Lach

Por el blog ya he escrito varios artículos relacionados con SomaFM y como escuchar sus diferentes emisoras. Pero hoy quiero simplemente compartir una lista de reproducción que he recopilado con todas las emisoras disponibles, a fecha de publicación del blog, de la emisora online SomaFM.

Se trata de un archivo eb formato .m3u de lista de reproducción con todos los enlaces a las emisoras de los distintos géneros de música disponibles en SomaFM.

Además de las URL a las emisoras de SomaFM, he añadido los metadatos, con la información de qué es cada canal según publican en la propia SomaFM.

El archivo está en un gist en GitHub y lo podéis descargar con wget mediante:

wget https://gist.githubusercontent.com/victorhck/2c9491f8dbd0873ffd3af6bdce07d8e3/raw/d45f4da8c8086497948a0f873334d65195ddc1d8/soma_fm.m3u -O soma_fm.m3u

Después lo podéis abrir con el reproductor de audio que mejor os parezca. VLC, Amarok, etc… Y a disfrutar de música en muchos estilos, sin cortes comerciales.

Enlaces de interés

a silhouette of a person's head and shoulders, used as a default avatar
a silhouette of a person's head and shoulders, used as a default avatar

Lista de reproducción con las emisoras de SomaFM

Quiero compartir en el blog una lista de reproducción actualizada que he recopilado con las emisoras disponibles de la radio online SomaFM

Foto: Ron Lach

Por el blog ya he escrito varios artículos relacionados con SomaFM y como escuchar sus diferentes emisoras. Pero hoy quiero simplemente compartir una lista de reproducción que he recopilado con todas las emisoras disponibles, a fecha de publicación del blog, de la emisora online SomaFM.

Se trata de un archivo eb formato .m3u de lista de reproducción con todos los enlaces a las emisoras de los distintos géneros de música disponibles en SomaFM.

Además de las URL a las emisoras de SomaFM, he añadido los metadatos, con la información de qué es cada canal según publican en la propia SomaFM.

El archivo está en un gist en GitHub y lo podéis descargar con wget mediante:

wget https://gist.githubusercontent.com/victorhck/2c9491f8dbd0873ffd3af6bdce07d8e3/raw/d45f4da8c8086497948a0f873334d65195ddc1d8/soma_fm.m3u -O soma_fm.m3u

Después lo podéis abrir con el reproductor de audio que mejor os parezca. VLC, Amarok, etc… Y a disfrutar de música en muchos estilos, sin cortes comerciales.

Enlaces de interés

the avatar of openSUSE News
the avatar of openSUSE News

openSUSE Asia Summit Set for Tokyo

openSUSE.Asia Summit will come back to Tokyo, Japan

The openSUSE Project is exciting to announce that openSUSE.Asia Summit 2024 is going to be held in Tokyo, Japan. The openSUSE.Asia Summit is an annual conference for users and contributors of openSUSE and FLOSS enthusiasts. During this summit, they will gather in person to share knowledge and experiences about openSUSE including applications running on it.

The venue of the summit will be located in Tokyo, the capital of Japan, blending tradition and cutting-edge technology. Its infrastructure and global connectivity make it a primal location for promoting collaboration among openSUSE users and developers. Moreover, Tokyo is a center of information technology; Many technology companies have their offices in Tokyo, with numerous engineers residing in the surrounding areas.

Tokyo is also a popular place for sightseeing with its unique culture, food, etc. Especially, characters from video games, anime, and comics, which are now common in the world, attract tourists to Japan. In Tokyo, you can easily find character shops and get items related to works you love.

The number of tourists from abroad has recovered last year to the same level as before COVID-19. Due to the currency exchange rate, it will be a great chance to enjoy your trip to Japan while saving your money. Even though you may have attended the last summit in Tokyo, you will discover new facets, developed before the TOKYO 2020 Summer Olympics.

Please see also:

The expected summit date is Nov. 2 and 3 soon after Open Source Summit Japan. Our call for speakers is going to end around the end of July. For more details including the venue, please stay tuned until the next announcement in a couple of weeks.

the avatar of Alessandro de Oliveira Faria
the avatar of Alessandro de Oliveira Faria

PAG: Orientação de atenção perturbada

Estudos recentes comprovam que modelos de difusão podem gerar amostras de alta qualidade, mas a qualidade dessas amostras muitas vezes depende fortemente de técnicas de orientação durante a amostragem, como a orientação por classificador (CG) e orientação sem classificador (CFG), que não são aplicáveis na geração incondicional ou em diversas tarefas subsequentes, como restauração de imagens. Neste artigo, propomos uma nova técnica de orientação para amostragem em difusão, chamada Orientação por Atenção Perturbada (PAG), que melhora a qualidade das amostras tanto em configurações incondicionais quanto condicionais, sem a necessidade de treinamento adicional ou integração de módulos externos. O PAG é projetado para aprimorar progressivamente a estrutura das amostras sintetizadas durante o processo de desruído, aproveitando a capacidade dos mecanismos de autoatenção de capturar informações estruturais. Ele envolve a geração de amostras intermediárias com estrutura degradada, substituindo mapas de autoatenção selecionados na difusão U-Net por uma matriz identidade e orientando o processo de desruído para se afastar dessas amostras degradadas.

A Orientação por Atenção Perturbada melhora significativamente a qualidade das amostras em modelos de difusão sem necessitar de condições externas, como rótulos de classes ou prompts de texto, nem de treinamento adicional. Isso é particularmente valioso em configurações de geração incondicional, onde a orientação sem classificador (CFG) não é aplicável. Nossa orientação pode ser utilizada para aumentar o desempenho em várias tarefas subsequentes que utilizam modelos de difusão incondicionais, incluindo ControlNet com um prompt vazio e tarefas de restauração de imagem como super-resolução e inpainting.

Comparativos qualitativos entre amostras de difusão guiadas e não guiadas (linha de base). Sem quaisquer condições externas, como rótulos de classes ou prompts de texto, ou treinamento adicional, nosso PAG eleva dramaticamente a qualidade das amostras de difusão mesmo em geração incondicional, onde a orientação sem classificador (CFG) não é aplicável. Nossa orientação também pode melhorar o desempenho base em várias tarefas subsequentes, como ControlNet com prompt vazio e restauração de imagem, incluindo inpainting e desfocagem.

Mais informações:
https://github.com/v0xie/sd-webui-incantations
https://github.com/KU-CVLAB/Perturbed-Attention-Guidance

a silhouette of a person's head and shoulders, used as a default avatar
a silhouette of a person's head and shoulders, used as a default avatar

KDE participa en Google Summer of Code 2024

Como es tradicional, aunque me salté el 2021, hoy quiero comentar que un año más KDE participa en Google Summer of Code (GSoC) en su edición del 2024 . A lo largo de muchos años, esta simbiosis entre la Comunidad KDE y el gigante multicolor ha sido muy provechosa para ambos, como hemos visto en muchas ocasiones en el blog, esperemos que éste también lo sea.

KDE participa en Google Summer of Code 2024

El equipo de KDE es uno de las Comunidades que siempre intentan colaborador con los proyectos sobre Software Libre que suele organizar cualquier compañía, y Google no es ninguna excepción.

KDE participa en Google Summer of Code 2024

Este año tenemos bastantes estudiantes mejorando sus habilidades al tiempo que mejoran las aplicaciones del ecosistema KDE. De esta forma según leemos en el blog de mentorías de KDE, qua acabo de descubrir, tenemos un articulo de Carl Schwan que nos cuenta que los estudiantes van a trabajar en mejorar aplicaciones como KDE Connect, Labplot, Arianna (un visor de ePub), Frameworks, Okular, Snaps, Krita, KDE Games y Kdenlive.

Más información: KDE Mentorship

¿Qué es GSoC?

KDE participará en Google Summer of Code 2016

Vía Somos Libres he encontrado esta magnífica descripción del programa GSoC:

Google Summer of Code (GSoC) es un evento organizado por Google, cuyo objetivo es hacer participar a varios estudiantes en el desarrollo de determinados proyectos Open Source elegidos por Google. Cada grupo debe cumplir con una lista de tareas específicas que deben realizar y elegidas por el representante del proyecto, también conocido como mentor.

Los objetivos del GSoC son:

  • Crear y liberar código Open Source para el beneficio de todos.
  • Inspirar a los jóvenes desarrolladores a participar en el desarrollo de aplicaciones Open Source.
  • Ayudar a los proyectos Open Source a identificar a nuevos y posibles desarrolladores.
  • Dar a los estudiantes la oportunidad de trabajar en algo relacionado a sus estudios. Dar a los estudiantes una mayor exposición a situaciones del mundo real de desarrollo de software.

En definitiva, una excelente iniciativa que beneficia a todo el mundo.

La entrada KDE participa en Google Summer of Code 2024 se publicó primero en KDE Blog.

the avatar of rickspencer3's Blog
the avatar of rickspencer3's Blog

openSUSE Tumbleweed is the Best Distro No One Knows About

openSUSE Tumbleweed is the Best Distro No One Knows About

I've been at SUSE for 4 months now. Of course the company keeps my primary focus on our Enterprise customers, but I have learned a lot more how openSUSE is built and used in the four months, and I have to say, I am impressed. I think Tumbleweed is the best developer distro that nobody knows about.

On my main laptop I opted to install the "stable" verion of openSUSE called "Leap." (you can read about that here). I followed suit on my $65 laptop, but ran into some issues based on the cheapness and newness of the laptop's components. For example, the wifi module was not recognized, and the built in speakers just didn't work. The wifi issue was obvious; the wifi module was too new for Leap 15.5, and I was too lazy to compile and install an up to date kernel driver for it.

As I learned more about openSUSE, I finally understood the difference between Tumbleweed and Leap, and I realized that Tumbleweed would probably work well on my oddball $65 laptop.

How is openSUSE built anyway?

openSUSE is unique, because it is both upstream of Suse Linux Enterprise, and downstream from it. Basically, what happens is:

  1. The openSUSE community is constantly packaging upstream software with the Open Build Server.
  2. Those packages are constantly being built into openSUSE Tumbleweed, which is, therefore, a rolling release. There is a quality assurance process that keeps Tumbleweed stable in the sense of "not crashy."
  3. Periodically, those packages from the Open Build Server, which become highly used and vetted by the community using Tumbleweed, then get moved into SUSE's Internal Build Server. From there, SUSE builds Certified and L3 Supported packages, that go into SUSE Linux Enterprie releases. This is a paid Enterprise product.
  4. Out of those packages, openSUSE Leap is built. Leap, therefore, is essentially the same as SUSE Linux Enterprise, but without the certifications and support.

diagram of SUSE packaging process

I assume I got some details wronng above, but I think that's the gist of it.

Choice happens. You can choose a high quality rolling release, a fully supported Enterprise release with a long lifecycle, or a free (as in speach and beer) release with the same lifecycle and bits as the Enterprise version.

For simplicity, I left out that there are even more options. For example, do you want an immutable OS with transactional updates? The openSUSE community has you covered with Microos.

So How did it Go?

Installing Tumbleweed was actually pretty boring. The main difference from installing Leap was that the wifi driver was recogized by the kernel (as I expected). I was pleasantly surprised to see that I also a built in LTE modem.

Tumbleweed detected and allowed me to configure the wifi

Up to Date

Looks like after install, every single package is up to date with the repositories. I uppose the installer installed all up to date packages from the repositories, which is sweet.

everything up to date

WIFI woes

However, while the built in wifi seemsto work, I noticed that when I am downloading files, they sometimes get "stuck." Either the server times out, or the data trickles in so slowly the files will never download. More on this bellow.

Next Steps

So now I seem to be a happy Tumbleweed user. I have installed my work software (Slack, etc...) so I am planning to take this device as my only laptop on an upcoming work trip to Europe in May. I should be in meetings most of the time, so it's a pretty low risk situation.

Follow up on Issues

So, this wifi issue ... this seems like a good opportunity for me to help out with the community however modestly. I will learn how to log an issue in the right place, and then see if I can help who ever turns out to be the right mainter address the bug.

Connect with the Community

I am motivated to started looking at this issue as openSUSE Conference is coming up at the end of June, and I am looking forward to connecting with community members and generally learning how the openSUSE community works, and seeing how I can collaborate and help.

a silhouette of a person's head and shoulders, used as a default avatar
a silhouette of a person's head and shoulders, used as a default avatar

Décimo audio de Podcast Linux «Raspberry Pi y GNU/Linux» – Podcast Linux #10

Aunque el proyecto Podcast Linux está parado esto no significa que no tenga cabida en el blog y, mientras pueda, seguiré promocionándolo con la esperanza de que reviva, como cierto pájaro mitológico. Y he pensado hacerlo de una forma sencilla para mi y creo que benificiosa para todos, creando poco a poco un índice de todas sus emisiones, de forma que podamos encontrar en este blog una alternativa a su magnífica obra. Así que bienvenidos al décimo audio de Podcast Linux «Raspberry Pi y GNU/Linux» – Podcast Linux #10donde Juan nos habla de la Raspberry Pi y de software que le saca el mayor rendimiento como Kodi.

Décimo audio de Podcast Linux «Raspberry Pi y GNU/Linux» – Podcast Linux #10

Décimo audio de Podcast Linux "Raspberry Pi y GNU/Linux" - Podcast Linux #10

Como los lectores del blog sabrán hace un tiempo Podcast Linux cerró sus emisiones por motivos que solo incumben a su creador. Desde el blog no quiero dejar que su recuerdo se desvanezca así que seguiré publicitando sus audios ya que su calidad no debe caer en el olvido.

Hace un tiempo decidí empezar por el principio, mostrando su primer audio, el cual no promocioné en su día. Lo mismo ocurría con los siguientes, y, si las búsquedas no me engañan no fue hasta los episodios 19, 20, 21 y 22 cuando empecé a hacerlo.

De esta forma continuo con su décimo audio, un programa increíble que presenta al microordenador Raspberry Pi y a una de las aplicaciones con las que mejor combina: Kodi. Pero mejor leer las palabras de Juan:

Muy buenas Linuxero. Bienvenido a otro episodio más de Podcast Linux.
Mi nombre es Juan Febles y cada quincena tengo una cita contigo para hablar de lo que más nos gusta: GNU/Linux.

En el Núcleo Kernel hablaré de la Raspberry Pi y su íntima relación con GNU/Linux.

En el Gestor de Paquetes le damos un repaso a Kodi, un Media Center multiplataforma.

El invitado a Comunidad Linux es Gabriel Viso, podcaster del podcast Pitando.net.
Por último, en Área de Notificaciones, le daré un repaso a los mensajes que he recibido en este último mes.

Hasta dentro de 15 días Linuxeros!!!!
Un abrazo muy fuerte.

Más información: Podcast Linux

Sigue a Podcast Linux

Aprovecho para animaros a seguir Podcast Linux en algunos de los canales de comunicación que tiene:

La entrada Décimo audio de Podcast Linux «Raspberry Pi y GNU/Linux» – Podcast Linux #10 se publicó primero en KDE Blog.

the avatar of Alessandro de Oliveira Faria
the avatar of Alessandro de Oliveira Faria

Experimentar roupas antes de comprar na Internet.

Em breve, graças à Inteligência Artificial, será possível experimentar de maneira virtual as roupas vendidas na internet antes de comprar. Para comprovar a teoria, busquei um projeto no github, e o coloquei em funcionamento na minha maquina. Após, entrei no site Véi Nerd ( https://www.veinerd.com/produto/camiseta-linux ). Escolhi uma camiseta que gostei e salvei na minha maquina. Ao submeter as duas imagens de referencia ao sistema, o resultado foi muito animador e tudo aconteceu de maneira automática.

No teste foi possível gerar imagens com um alto grau de consistência de vestuário em cenários reais. Mesmo com fundos complexos ou poses diversas das pessoas, é possível gerar imagens de alta qualidade.

O projeto IDM-VTON é a experimentação virtual baseada em imagens, que produz uma imagem de uma pessoa usando uma peça de vestuário selecionada, a partir de um par de imagens que retratam, respectivamente, a pessoa e a vestimenta. Tecnologia anteriores adaptaram modelos de difusão baseados em exemplos existentes para a experimentação virtual, a fim de melhorar a naturalidade das imagens geradas em comparação com outros métodos (como os baseados em GAN), mas não conseguiram preservar a identidade das vestimentas.

Para superar este obstáculo, o projeto IDM-VTON propões um modelo de difusão inovador que melhora a fidelidade da vestimenta e gera imagens autênticas de experimentação virtual. Pois utiliza dois módulos diferentes para codificar a semântica da imagem da vestimenta; com base no UNet do modelo de difusão,

1) a semântica de alto nível extraída de um codificador visual é fundida à camada de atenção cruzada, e então 2) os recursos de baixo nível extraídos do UNet paralelo são fundidos à camada de autoatenção. Além disso, é fornecido prompts textuais detalhados para as imagens tanto da vestimenta quanto da pessoa, a fim de realçar a autenticidade das visuais geradas. Então é constatado um método de personalização usando um par de imagens de pessoa-vestimenta, que melhora significativamente a fidelidade e autenticidade

Os meus testes experimentais demonstram que o método supera as abordagens anteriores (tanto baseadas em difusão quanto em GAN) na preservação dos detalhes da vestimenta e na geração de imagens autênticas de experimentação virtual, tanto qualitativa quanto quantitativamente. Além disso, o método de personalização proposto demonstra sua eficácia em um cenário real. Mais visualizações estão disponíveis na página do projeto: https://github.com/yisol/IDM-VTON