An open source machine learning framework that accelerates the path from research prototyping to production deployment. PyTorch is an open source project at the Linux Foundation.

PyTorch

Upcoming webinar: Dinosaur Bone Hunting with Intel AI! 🦖🔎 Join us for a live, hands on workshop on February 6.

Details + registration: https://hubs.la/Q02jwmrk0

Check out what's new in PyTorch Documentation 👀 Featuring 3 new tutorials and PyTorch 2.2 documentation + tutorials!

Details 👉  https://hubs.la/Q02jwrbr0

PyTorch 2.2 is here 🎉

Featuring: 
- SDPA support of FlashAttention-2
- New ahead-of-time extension of TorchInductor
- device_mesh, a new abstraction for initializing and representing ProcessGroups 
- A standardized, configurable logging mechanism called TORCH_LOGS

More announcement details in our latest blog: https://hubs.la/Q02jc6kB0

With this release, we've also added new libraries. 📚 These updates demonstrate our focus on developing common and extensible APIs across all domains to make it easier for our community to build ecosystem projects on PyTorch. 🔥

Check out the library updates here: https://hubs.la/Q02jc4gH0

Colossal-AI has achieved remarkable results on 7B LLaMA-2 model by utilizing only 0.0085 trillion tokens of data, investing 15 hours, and incurring a few hundred training costs! 💯🔥

Read more: https://hubs.la/Q02j0Chq0

Check it out on GitHub: https://hubs.la/Q02j0wkY0 

Happy to share an early version of float8_experimental, a library to accelerate training with float8 + torch.compile written in native PyTorch: https://lnkd.in/ghsfYSJr 🔥 

NVIDIA H100s support float8, with theoretical advantages of 2x in throughput and memory usage over 16-bit formats. We aim to make it easy to use float8 in PyTorch.

Per-tensor scaling is commonly used to maintain accuracy, introducing significant overhead in the absence of fused kernels.

We use torch.compile to generate fast kernels for per-tensor scaling, achieving speedups of up to 1.20x on pre-training LLaMa 7B benchmarks on one H100-SXM GPU.

What about distributed? FSDP with 16-bit all-gather is supported today. 

Support for tensor parallel, sequence parallel and FSDP with 8-bit all-gather is planned. 

Learn more here 👉  https://lnkd.in/gp_McA_S

A powerful and mature deep learning library with strong visualization capabilities and several options for high-level model development. It has production-ready deployment options and support for mobile platforms.

TensorFlow

A high-level neural networks API, written in Python and capable of running on top of TensorFlow, designed to enable fast experimentation with deep neural networks.

Keras

A lightweight, modular, and scalable deep learning framework developed by Facebook. It is designed with expression, speed, and modularity in mind.

Caffe2

A flexible and efficient library for deep learning, offering both imperative and symbolic programming interfaces, making it easy to use and efficient to scale.

MXNet

An open-source machine learning library, a predecessor to PyTorch, providing a wide range of algorithms for deep learning and machine learning tasks.

Torch

PyTorch is a machine learning library based on the Torch library, used for applications such as computer vision and natural language processing. It is recognized as one of the two most popular machine learning libraries alongside TensorFlow, offering free and open-source software released under the modified BSD license.

industry_trends_dossier

news_impact_assignment_basic

key_people

company_dossier

network_dossier

public_peers

news_daily_refresh

news_processing_daily_refresh

news_impact_assignment_daily

news_processing_basic

legal

legal_dossier

industry trends dossier	dataroom
company dossier	dataroom
network	dataroom
network mapping	dataroom
transactions dossier	dataroom
network onboarding	dataroom	Mar 8, 2024 12:13 AM	e14f200e
news impact assignment basic	general	Mar 8, 2024 12:13 AM	b6cc41d5
public peers	general
news daily refresh	general
news processing daily refresh	general
news impact assignment daily	general
news processing basic	general	Mar 8, 2024 12:13 AM	5ebd1b03
company summary	general
signals	general	Mar 8, 2024 12:13 AM	ad7bcf59
news advanced	general
news impact assignment advanced	general
news processing advanced	general
peers	general	Mar 8, 2024 12:13 AM	29a7c6de
news basic	general	Mar 8, 2024 12:13 AM	db74e4f8
events	markdown
bundesanzeiger	filings
financial statements	filings
regulatory analysis	company_analysis
industry analysis	company_analysis
swot analysis	company_analysis
competition analysis	company_analysis
financial legal advisories	company_analysis
merger acquisitions	company_analysis
customer profile	customers

ORG:PYTORCH

PyTorch