BigDL release 2.4.0

Highlights

Note: BigDL v2.4.0 has been updated to include functional and security updates. Users should update to the latest version.

BigDL release 2.3.0

Highlights

Note: BigDL v2.3.0 has been updated to include functional and security updates. Users should update to the latest version.

Nano

• Enhanced trace and quantization process (for PyTorch and TensorFlow model optimizations)
• New inference optimization methods (including Intel ARC series GPU support, CPU fp16, JIT int8, etc.)
• New inference/training features (including TorchCCL support, async inference pipeline, compressed model saving, automatic channels_last_3d, multi-instance training for customized TF train loop, etc.)
• Performance enhancement and overhead reduction for inference optimized model
• More user-friendly document and API design

Orca:

• Step-by-step distributed TensorFlow and PyTorch tutorials for different data inputs.
• Improvement and examples for distributed MMCV pipelines.
• Further enhancement for Orca Estimator (more flexible PyTorch train loops via Hook, improved multi-output prediction, memory optimization for OpenVINO, etc.)

Chronos

• 70% latency reduction for Forecasters
• New bigdl.chronos.aiops module for AIOps use case on top of Chronos algorithms.
• Enhanced TF-based TCNForecaster to better accuracy

Friesian:

• Automatic deployment of RecSys serving pipeline on Kubernetes with Helm Chart

PPML

• TDX (both VM and CoCo) support for Big Data, DL Training & Serving (including TDX-VM orchestration & k8s deployment, TDXCC installation & deployment, attestation and key management support, etc.)
• New Trusted Machine Learning toolkit (with secure and distributed SparkML & LightGBM support)
• Trusted Big Data toolkit upgrade (>2x EPC usage reduction, Apache Flink support, Azure MAA support, multi-KMS support, etc.)
• Trusted Deep Learning toolkit upgrade (with improved performance using BigDL Nano, tcmalloc, etc.)
• Trusted DL Serving toolkit upgrade (with Torch Serve, TF-Serving, and improved throughput and latency)

BigDL release 2.2.0

Highlights

Note: BigDL v2.2.0 has been updated to include functional and security updates. Users should update to the latest version.

• Nano
  • Extend BigDL Nano inference to support iGPU and more data types (INT8/BF16/FP16 quantization)
  • More performance features (e.g., InferenceOptimizer for Keras, Nano decorator for PyTorch training loop, Nano Context Manager for thread number control and autocast, etc.)
  • Support installation with more PyTorch/TensorFlow versions and conditional dependencies on different platforms
• PPML
  • Upgrade BigDL PPML solution to support new LibOS (e.g., Gramine1.3.1, Occlum0.29.2) with better security, higher performance, more stability and easier deployment.
  • Support more Big Data frameworks (Spark 3.1.3, Flink, Hive etc.), more Python and Data Science tools (Numpy, Pandas, sklearn, Torch Serv, Triton, Flask etc.), and distributed DL training using Orca
  • Improve the Attestation (e.g., MREnclave Attestation), Key Management (e.g., multi-KMS) & Encryption (e.g., transparent encryption) features for better end-to-end secure pipeline.
  • Initial support of BigDL PPML on SPR TDX (Virtual Machine and TDX Confidential Container)
• Chronos
  • Extend BigDL Chronos to support Windows and Mac, and new Python versions (3.8/3.9)
  • Provide a benchmark tool for Chronos users to evaluate Chronos performance on their platform
  • More performance features (e.g., accuracy and performance improvement for TCNForecaster, lower memory usage, auto optimization search, faster and portable TSDataset, etc.)
• Friesian
  • LightGBM training support
  • Performance improvements for online serving pipeline
• Orca
  • Improve Orca Estimator APIs for better user experience
  • Memory optimization for distributed training with Spark DataFrame,
  • Better support for image inputs and visualization with Xshards
  • Distributed MMCV applications using Orca
• Documentation
  • Tutorials for running BigDL Orca on YARN/K8s/Databricks
  • BigDL PPML solutions on Azure
  • How-to guides and examples for Chronos forecasting and deployment process

BigDL release 2.1.0

Highlights

Note: BigDL v2.1.0 has been updated to include functional and security updates. Users should update to the latest version.

• Orca
  • Improve user experience and API consistency for Orca Estimators.
  • Support directly save and load TensorFlow model format in Orca TensorFlow2 Estimator.
  • Provide more examples (e.g. PyTorch brain image segmentation, XShards tutorials for distributed Python data processing), etc.
  • Support customized metrics in Orca PyTorch Estimator.
• Nano
  • New inference optimization pipelines, with more optimization methods and a new InferenceOptimizer
  • More training optimization methods (bf16, channel last)
  • Add TorchNano support for PyTorch model customized training loop
  • Auto-scale learning rate for multi-instance training
  • Built-in AutoML support through hyperparameter optimization
  • Support a wide range versions of pytorch (1.9-1.12) and tensorflow (2.7-2.9)
• DLlib
  • Add LightGBM support
  • Improve Keras-style model summary API
  • Add Python support for loading HDFS files
• Chronos
  • Add new Autoformer (https://arxiv.org/abs/2106.13008) Forecaster and pipeline that are optimized on CPU
  • Tensorflow 2 support for LSTM, Seq2Seq, TCN and MTNet Forecasters
  • Add light-weight (does not rely on Spark/Ray Tune) auto tunning
  • Better support on distributed workflow (spark df and distributed pandas processing)
  • Add more installation options is now supported to make the installation lighter
• Friesian:
  • Integration of DeepRec (https://github.com/alibaba/DeepRec) with Friesian.
  • Add more reference examples, e.g. multi-task recommendation, TFRS (https://www.tensorflow.org/recommenders) list-wise ranking, LightGBM training, etc.
  • Add a reference example for offline distributed similarity search (using FAISS)
  • More operations in FeatureTable (e.g. string embeddings with BERT, etc.).
• PPML
  • Upgrade BigDL PPML on Gramine.
  • Improve the attestation and key managing process
  • More Big Data frameworks on BigDL PPML (including spark, flink, hive, hdfs, etc.)
  • Add PPMLContext API for encryption IO and KMS, supports different file formats, encryption algorithms and KMS services
  • Support PSI, Pytorch NN, Keras NN, FGBoost (federated XGBoost) in VFL scenario, linear regression & logistic regression for VFL

BigDL release 2.0.0

Highlights

Note: BigDL v2.0.0 has been updated to include functional and security updates. Users should update to the latest version.

BigDL release 0.13.0

v0.13.0

Update deploy-spark2.sh

BigDL release 0.12.2

v0.12.2

flip version to 0.12.2 (#3119)

BigDL release 0.12.1

v0.12.1

add 0.12 release doc (#3095)

BigDL release 0.11.1

v0.11.1

flip version to 0.11.1 (#3048)

BigDL release 0.10.0

Highlights

• Continue RNN optimization. We support both LSTM and GRU integration with MKL-DNN which acheives ~3x performance

• ONNX support. We support loading third party framework models via ONNX

• Richer data preprocssing support and segmentation inference pipeline support

Details

• [New Feature] Full MaskRCNN model support with data processing
• [New Feature] Support variable-size Resize
• [New Feature] Support batch input for region proposal
• [New Feature] Support samples of different size in one minibatch
• [New Feature] MAP validation method implementation
• [New Feature] ROILabel enhancement to support both object detection and segmentation
• [New Feature] Grey image support for segmentation
• [New Feature] Add TopBlocks support for Feature Pyramid Networks (FPN)
• [New Feature] GRU integration with MKL-DNN support
• [New Feature] MaskHead support for MaskRCNN
• [New Feature] BoxHead support for MaskRCNN
• [New Feature] RegionalProposal support for MaskRCNN
• [New Feature] Shape operation support for ONNX
• [New Feature] Gemm operation support for ONNX
• [New Feature] Gather operation support for ONNX
• [New Feature] AveragePool operation support for ONNX
• [New Feature] BatchNormalization operation support for ONNX
• [New Feature] Concat operation support for ONNX
• [New Feature] Conv operation support for ONNX
• [New Feature] MaxPool operation support for ONNX
• [New Feature] Reshape operation support for ONNX
• [New Feature] Relu operation support for ONNX
• [New Feature] SoftMax operation support for ONNX
• [New Feature] Sum operation support for ONNX
• [New Feature] Squeeze operation support for ONNX
• [New Feature] Const operation support for ONNX
• [New Feature] ONNX model loader implementation
• [New Feature] RioAlign layer support
• [Enhancement] Align batch normalization layer between mklblas and mkl-dnn
• [Enhancement] Python API enhancement to support nested list input
• [Enhancement] Multi-model training/inference support with MKL-DNN
• [Enhancement] BatchNormalization fusion with Scale
• [Enhancement] SoftMax companion object support no argument initialization
• [Enhancement] Python support for training with MKL-DNN
• [Enhancement] Docs enhancement
• [Bug Fix] Fix model version comparison
• [Bug Fix] Fix graph backward bug for ParallelTable
• [Bug Fix] Fix memory leak for training with MKL-DNN
• [Bug Fix] Fix performance caused by denormal values during training
• [Bug Fix] Fix SoftMax segment fault issue under MKL-DNN
• [Bug Fix] Fix TimeDistributedCriterion python API inconsistent with Scala