A video has been published on nnabla's Youtube Channel!
We explain DeFace, an approach to model 3D facial textures under challenging illumination affected by occlusions.
[AI Paper] Perform 3D face texture modeling with DeFace!
https://t.co/MH9GjkbuEU
#nnabla_youtube
A video has been published on NNabla's Youtube Channel!
[AI Paper] Perform efficient portrait animation with LivePortrait!
https://t.co/AGBLpCsBml
We explain the LivePortrait, portrait animation framework on better generalization, controllability and efficiency.
#nnabla_youtube
NNC Desktop has been newly added!
In addition to the Cloud and Windows versions, we are excited to announce the release of the open-source Desktop version. This supports both Windows/Linux and enables training with GPUs.
https://t.co/0D6hemkNWV
#nnc#sony#NeuralNetworkConsole
A video has been published on YouTube nnabla channel!
Sony's Research Minds:
[AI paper] CTM: Advanced Single-Step DIffusion Model for Fast and High-Quality Sampling
https://t.co/ja3XoJ3tQ8
#nnabla_youtube#SonysResearchMinds
A video has been published on YouTube nnabla channel!
[AI Paper] SAN: Improving the performance of any GANs by reconstructing the discriminator! Explaining an ICLR accepted paper | Sony's Research Minds
https://t.co/psSbBmhAX7
#nnabla_youtube#SonysResearchMinds
#ICLR2024 Feature Month!
We will be releasing explanatory videos starting next week for the notable ICLR papers from Sony!
SAN: https://t.co/G4aaxDVrlx
MPGD: https://t.co/LSpSUqYtFk
CTM: https://t.co/Ll8nrnldgm
Don't miss it!
#nnabla_youtube#SonysResearchMinds
We have released Neural Network Libraries v1.39.0 and nnabla-rl v0.15.0!
nnabla started supporting MLFlow in monitor functions.
nnabla-rl has added support for Gymnasium and HyAR, SRSAC, and AMP algorithms!
https://t.co/gcrMDc6A0s
#sony#neuralnetworklibraries
A video has been published on YouTube nnabla channel!
[AI Paper] Perform one-shot face reenactment with HyperReenact!
https://t.co/ponGPoK9SM
HyperReenact was published in ICCV 2023 and proposes a neural face reenactment method robust to large head pose shifts.
#nnabla_youtube
A new video has been published on YouTube nnabla channel!
[AI Paper] Using Generative AI for Image Classification!
https://t.co/WhHlHxLzO0
We introduce ICLR2024 paper by Google DeepMind and explain how generative AI models can be used for image classification.
#nnabla_youtube
We have released Neural Network Libraries v1.37.0, nnabla-rl v0.14.0 and nnabla-nas v0.15.0!
nnabla started supporting CUDA 12.0 extension, and nnabla-rl prepared a template to start reinforcement learning easily.
https://t.co/CIe30d9EJI
#sony#neuralnetworklibraries
A video has been published on YouTube nnabla channel!
This video explains a paper that proposes a real-time high-fidelity portrait avatar reconstruction.
Perform high fidelity portrait avatar reconstruction in real-time with StyleAvatar!
https://t.co/kQEyp0kPor
#nnabla_youtube
We have released Neural Network Libraries v1.35.0!
In addition to adding Lion Optimizer, improve tflite support in nnabla-converter significantly!
We also released nnabla-rl v0.13.0!
https://t.co/GnFTAB3Ooh
#sony#neuralnetworklibraries
We have released Neural Network Libraries v1.34.0!
We started supporting CUDA 11.6 extension.
Also, we fixed a bug in channel last layout of SyncBatchNormalization.
https://t.co/GUyope9VOC
#sony#neuralnetworklibraries
We have released Neural Network Libraries v1.33.0!
We add an example of CLIPort which is a multi-task manipulation policy learning algorithm, and zero-short image classification demo!
Also, this release has many bugfixes.
https://t.co/9m5gzOLgyE
#sony#neuralnetworklibraries
We have released Neural Network Libraries v1.32.0!
In addition to improvements to auto-forward mode that make it more than double memory efficiency and faster with overhead reduction, we also started supporting python3.10.
https://t.co/icKfYyAaDW
#sony#neuralnetworklibraries
The new video is available on YouTube nnabla channel!
This video explains NAS papers presented in CVPR 2022 and ECCV 2022, including trend analysis.
【Conference report】Neural architecture search in CVPR 2022 and ECCV 2022
https://t.co/Lp3CvkNHd9
#nnabla_youtube
We have released Neural Network Libraries v1.31.0!
In addition to adding attractive demos, we also released nnabla-rl, nnabla-browser and nnabla-nas!
Now nnabla-rl can work with nnabla-browser, so give it a try!
https://t.co/jKt6snXmWL
#sony#neuralnetworklibraries
Neural Network Libraries v1.30.0 and nnabla-nas v0.13.0 has been released!
CUDA11.4 is supported now along with new powerful features including all-in-one wheels, prejudice remover regularizer, CLIP training!
https://t.co/woQKlGO9lb
#neuralnetworklibraries#sony
We have released Neural Network Libraries v1.29.0!
We provide QATScheduler to support QAT, and nnabla-ext-cuda* supports distributed training. Also, more reinforcement learning examples have been added!
https://t.co/OnLMz7vOWT
#sony#neuralnetworklibraries
We have released Neural Network Libraries v1.28.0!
We have added a new graph converter PruningModifier to support weight pruning!
https://t.co/2k7gm2HPY5
#sony#neuralnetworklibraries