Husky
Features
Docs
Quick Start
Contact Us
Solutions
Research
About Us
User-friendly APIs:
C/C++ API
Functioning
For programmers/software developers
Simple and even more intuitive to understand and reason than Python/Scala APIs
Allow fine control for better utilization of system sources to boost performance
Python API
Functioning
Improving
For data scientists and users who are good at Python
User-friendly and low development cost
Backed-up by highly efficient C++ library for most core operations
Scala API
Functioning
Improving
Offer something in-between C/C++ API (high performance) and Python API (low development cost)
Expressive computing model:
Powerful object interaction model
Naturally model many existing distributed computing frameworks, e.g.:
Parameter server
MapReduce pipeline
Pregel's vertex-centric model
GraphLab's GAS model
Storm's streaming model
Spark's mini-batch streaming model
Interactive data analytics
Support efficient fine-grained operations (e.g., machine learning, graph analytics)
Support efficient coarse-grained transformations (e.g., map, reduce, join)
Support both synchronous and asynchronous execution
Compatibility with other platforms:
Only aim to excel in what is natural with Husky's computing model
Combine with other platforms for workloads that are not best supported by Husky
Connection with the following external components
Notebooks:
Jupyter Notebook
Functioning
Hadoop ecosystems:
HBase
Functioning
,
Hive
In Progress
NoSQL:
MongoDB
Functioning
,
Cassandra
Pending
,
CouchDB
Pending
Distributed data store:
Redis
Pending
Messaging systems:
Kafka
Functioning
Enterprise search engine:
ElasticSearch
Pending
Algorithm Packages Development Status:
Data cube and OLAP
In Progress
Recommendation
Collaborative filtering (SGD, ALS)
Functioning
Tensor factorization
Functioning
Classification
Random forests
Functioning
SVM
Functioning
CNN
Pending
Logistic regression
Functioning
Naive bayes
Pending
kNN
Functioning
Regression
Linear Regression
Functioning
Random forests
Functioning
Clustering
K-means
Functioning
DBSCAN
Functioning
LDA
Pending
Text mining
BM25
Functioning
TF-IDF
Functioning
Feature Selection
PCA
Functioning
Graph analytics
Functioning
LSH for various metrics (e.g., Euclidean, Angular, Jaccard, etc.)
Functioning