少し遅くなりましたが、OpenVINO 2021.4.1 LTSがリリースされました
https://software.intel.com/content/www/us/en/develop/articles/openvino-2021-4-lts-relnotes.html
Intelからのリリースを転記します
Major Features and Improvements
This 2021.4.1 LTS release provides functional bug fixes, and minor capability changes for the previous 2021.4 Long-Term Support (LTS) release, enabling developers to deploy applications powered by Intel® Distribution of OpenVINO™ toolkit with confidence. To learn more about long-term support and maintenance, go to the Long-Term Support Policy.
NOTE: A new LTS version is released every year and is supported for 2 years (1 year of bug fixes, and 2 years for security patches). The LTS version is intended for developers taking OpenVINO™ toolkit to production. For developers that prefer the very latest features and leading performance, standard releases are recommended. Standard releases will continue to be made available three to four times a year.
- Learn more about what components are included in the LTS release in the Included in this release section. Specific fixes to the known issues:
- Model Optimizer, Inference Engine (Inference Engine Python API, CPU, GPU, MYRAID, HDDL, and GNA plugin), Deep Learning Streamer, and Post-Training Optimization Tool (POT)
- Minor capability changes and bug fixes to the Open Model Zoo
- New Jupyter Notebook tutorials that simplify getting started with OpenVINO™ toolkit:
- Optical Character Recognition (OCR) tutorial – detects text and then performs scene text recognition
- HuggingFace BERT Quantization tutorial – demonstrates post-training INT8 quantization pipeline for NLP models
- Hello Detection – demonstrates basic object detection with OpenVINO using a text detection model
- Hello Segmentation – demonstrates basic semantic segmentation with a road segmentation model
- New PaddlePaddle-to-OpenVINO tutorials: Image Classification and PaddleGAN Super Resolution
- Download the 2021.4.1 LTS release of the Intel® Distribution of OpenVINO™ toolkit to upgrade to the latest LTS release.
Long Term Support版なので、バグフィクスなどがメインですが、Post-Training Optimization Toolが入っています。
これは、モデルを再トレーニング無しで、最適化するツールと説明されています。
必要なのは、FP16/32のモデルとキャリブレーションモデルとのことです。
これがあれば、再トレーニングしなくても良さそうなので、openvino.jpでも試してみたいと思います。
産業用画像処理装置開発、
ゲームコンソール開発、半導体エンジニアなどを経て、
Webエンジニア&マーケティングをやっています
好きな分野はハードウェアとソフトウェアの境界くらい