OB1K is a new RPC container. it belongs to a new breed of frameworks that tries to improve on the classic JEE model by embedding the server and reducing redundant bloatware.
OB1K supports two modes of operations: sync and async, the async mode aims for maximum performance by adopting reactive principals like using non-blocking code and functional composition using futures.
Ob1k also aims to be ops/devops friendly by being self contained and easily configured.
Massimiliano Dess狸 is a speaker who gave a presentation in Rome in April 2014 about Vert.x, a lightweight polyglot application framework for building highly responsive and reactive applications on the JVM. Vert.x uses an event-driven architecture based on the reactor pattern and event loops to enable asynchronous and non-blocking development. It utilizes verticle components that react to event messages to allow for scalable and resilient applications.
This chapter discusses complex events and event hierarchies in complex event processing (CEP). It defines a complex event as an aggregation of other events. It describes how event pattern rules can be used to create complex events signifying activities consisting of multiple events. It also discusses how event abstraction hierarchies organize events into different levels of abstraction and define rules for aggregating lower-level events into higher-level complex events. Personalized views for different stakeholders can be built by defining event types and abstraction levels relevant to their roles and interests.
The document discusses the development of a Netty 4-based RPC system. It describes using Netty 4 as a non-blocking I/O framework for building asynchronous network applications. It then outlines the design of the RPC system, including components like stubs, skeletons, command handling, and deployment. Finally, it discusses performance testing done on the RPC system to evaluate aspects like throughput, response times, resource usage and scalability under different loads.
Auto Scalable Deep Learning Production AI Serving Infra 蟲 覦 AI DevOps...hoondong kim
油
[Tensorflow-KR Offline 碁碁 覦襭]
Auto Scalable Deep Learning Production AI Serving Infra 蟲 覦 AI DevOps Cycle 蟲 覦覯襦. (Azure Docker PaaS 1襷 TPS Tensorflow Inference Serving 覦覯襦 螻旧)
The document discusses various machine learning clustering algorithms like K-means clustering, DBSCAN, and EM clustering. It also discusses neural network architectures like LSTM, bi-LSTM, and convolutional neural networks. Finally, it presents results from evaluating different chatbot models on various metrics like validation score.
The document discusses challenges with using reinforcement learning for robotics. While simulations allow fast training of agents, there is often a "reality gap" when transferring learning to real robots. Other approaches like imitation learning and self-supervised learning can be safer alternatives that don't require trial-and-error. To better apply reinforcement learning, robots may need model-based approaches that learn forward models of the world, as well as techniques like active localization that allow robots to gather targeted information through interactive perception. Closing the reality gap will require finding ways to better match simulations to reality or allow robots to learn from real-world experiences.
[243] Deep Learning to help students Deep LearningNAVER D2
油
This document describes research on using deep learning to predict student performance in massive open online courses (MOOCs). It introduces GritNet, a model that takes raw student activity data as input and predicts outcomes like course graduation without feature engineering. GritNet outperforms baselines by more than 5% in predicting graduation. The document also describes how GritNet can be adapted in an unsupervised way to new courses using pseudo-labels, improving predictions in the first few weeks. Overall, GritNet is presented as the state-of-the-art for student prediction and can be transferred across courses without labels.
[234]Fast & Accurate Data Annotation Pipeline for AI applicationsNAVER D2
油
This document provides a summary of new datasets and papers related to computer vision tasks including object detection, image matting, person pose estimation, pedestrian detection, and person instance segmentation. A total of 8 papers and their associated datasets are listed with brief descriptions of the core contributions or techniques developed in each.
[226]NAVER 蟯螻 deep click prediction: 覈碁覿 觜蟾讌NAVER D2
油
This document presents a formula for calculating the loss function J(慮) in machine learning models. The formula averages the negative log likelihood of the predicted probabilities being correct over all samples S, and includes a regularization term 了 that penalizes predicted embeddings being dissimilar from actual embeddings. It also defines the cosine similarity term used in the regularization.
[214] Ai Serving Platform: 襭 蟇伎 誤朱一るゼ 豌襴蠍 螻蟲磯蠍NAVER D2
油
The document discusses running a TensorFlow Serving (TFS) container using Docker. It shows commands to:
1. Pull the TFS Docker image from a repository
2. Define a script to configure and run the TFS container, specifying the model path, name, and port mapping
3. Run the script to start the TFS container exposing port 13377
The document discusses linear algebra concepts including:
- Representing a system of linear equations as a matrix equation Ax = b where A is a coefficient matrix, x is a vector of unknowns, and b is a vector of constants.
- Solving for the vector x that satisfies the matrix equation using linear algebra techniques such as row reduction.
- Examples of matrix equations and their component vectors are shown.
This document describes the steps to convert a TensorFlow model to a TensorRT engine for inference. It includes steps to parse the model, optimize it, generate a runtime engine, serialize and deserialize the engine, as well as perform inference using the engine. It also provides code snippets for a PReLU plugin implementation in C++.
The document discusses machine reading comprehension (MRC) techniques for question answering (QA) systems, comparing search-based and natural language processing (NLP)-based approaches. It covers key milestones in the development of extractive QA models using NLP, from early sentence-level models to current state-of-the-art techniques like cross-attention, self-attention, and transfer learning. It notes the speed and scalability benefits of combining search and reading methods for QA.
4. ろ磯 ?
Java network application framework
http://netty.io/
@netty_project
Asynchronous & event-driven
High-throughput & highly concurrent connection
.. with less resources threads, memory & CPU cycles
API as an I/O abstraction layer
Works with NIO, OIO, AIO & more without many changes
Well-defined event model & thread model
Flexible event handling with bi-directional chain of responsibility pattern
Promotes 'separation of concerns'
I/O layer (Netty)
Protocol codec (Netty or user)
Business logic (user)
5. 蟲煙
Business Logic
Custom Event Handlers & Codecs
User Code
HTTP
TCP UDP In-VM Codec Framework
Event Handlers
Transports Handlers
I/O Abstraction Channels, Event Loops, and Pipelines
Buffers
Core
9. 蠍一ヾ 企欧 覈
企欧 = 覦 螳豌 Is GC cheap?
No at extreme scale
ChannelOpen (instanceof ChannelStateEvent)
Can buffer pool beat JVM's memory allocator?
ChannelBound (instanceof ChannelStateEvent) Think memory bandwidth took by memset
ChannelConnected (instanceof ChannelStateEvent) Modern concurrency constructs enable close-to-native
efficiency in object pooling
MessageEvent (inbound)
Too many state events!
MessageEvent (outbound)
Open bound connected is an overkill in most TCP conns
ChannelDisconnected (instanceof ChannelStateEvent)
What's interesting to user is 'connected'.
ChannelUnbound (instanceof ChannelStateEvent)
ChannelClosed (instanceof ChannelStateEvent)
New buffer is created on every MessageEvent.
GC pressure
User has no control over the buffer
Heap or direct
Bounded or unbounded
No buffered flush
Every write is an immediate flush
Sometimes not desired
16. ろ 豌伎 ろ
螳 碁ろ語 誤蠍 覈 碁ろ 譟壱螻 譯殊 襦貊 譟壱 蟲谿 ろ
TCP: NIO x AIO x OIO = 9 1 (OIO x OIO)
UDP: NIO x OIO = 4
襦貊 : Echo, SSL, SPDY = 3 * TCP (8) = 24
豌 螻 碁る 蟆曙磯 EmbeddedChannel 伎
蠍壱
Jenkins
Animal Sniffer: Build fails on violation before packaging
Checkstyle: Build fails on violation before compilation
17. 螳 蟇磯Μ
Dynamic event routing
Statistical Distributed Performance Analysis on Mesos Cluster
Management and Monitoring
Native Transport
Scalable Community