@DanilBaibak Profile picture

Danylo Baibak

@DanilBaibak

Machine learning engineer, open-source enthusiast

Joined November 2012
Similar User
YerevaNN photo

@YerevaNN

Ananya (ಅನನ್ಯ) photo

@punarpuli

mftang photo

@mftang

Carrick Capital photo

@carrickcapital

Sakinat Folorunso photo

@SakinatTijani

Sourodip Kundu photo

@KunduSourodip

Nezihe Merve Gürel (nmervegurel.bsky.social) photo

@nmervegurel

Katrina Drozdov (Evtimova) photo

@stochasticdoggo

Justin Shenk photo

@AskJustinShenk

Gigi Datta photo

@okay2085

Kevin Du photo

@Kevin_Duu

Alain Rakotomamonjy photo

@rakotal1

Shadi Rezapour photo

@shadi_rezapour

Grigoris Chrysos photo

@Grigoris_c

Code Llama, a state-of-the-art large language model for coding - ai.meta.com/blog/code-llam…. State-of-the-art LLM capable of generating code, and natural language about code. Will software engineers survive? 🦕 #LLM #AI #Llama2 #Meta


Does the @UN support war criminals and murderers? Any statements about an ecological catastrophe at the Kakhovka HPP? Why hasn't @UnitedNationsRU been fired from the internal organization yet? Is the UN leadership involved in the terrorist activities of russia?

Tuesday is Russian Language Day. Follow @UnitedNationsRU for updates on the UN's work in Russian.



This document is my running literature review for people trying to catch up on AI. It covers 22 models, 11 architectural changes, 7 post-pre-training techniques and 3 training techniques (and 5 things that are none of the above) #DeepLearning #LLM #AI kipp.ly/blog/transform…


An example machine learning project on how to convert an audio recording of a meeting into a list of the actions for a team #MachineLearning #OpenAI GitHub: github.com/DanilBaibak/au… Article: danylo-baibak.medium.com/machine-learni…

Tweet Image 1

3D-aware Conditional Image Synthesis (pix2pix3D). Pix2pix3D synthesizes 3D objects (neural fields) given a 2D label map, such as a segmentation or edge map #ComputerVision #GAN #3D Github: github.com/dunbar12138/pi… Paper: arxiv.org/abs/2302.08509 Project: cs.cmu.edu/~pix2pix3D/


3D Object Tracking - Multi-modality tracker that fuses information from visual appearance and geometry to estimate object poses #DeepLearning #ComputerVision Paper: arxiv.org/abs/2302.11458… GitHub: github.com/dlr-rm/3dobjec… Dataset : paperswithcode.com/dataset/ycb-vi…


Token Merging: Your ViT but faster - @MetaAI is sharing new research to reduce the latency of existing Vision Transformer models without the need for additional training #ComputerVision #DeepLearning #AI GitHub: github.com/facebookresear… Paper: arxiv.org/abs/2210.09461


Box-supervised Instance Segmentation via Level-set Evolution - a toolbox that aims to provide #SOTA box-supervised instance #segmentation algorithms. It supports #InstanceSegmentation with only box annotations Github: github.com/LiWentomng/Box… Paper: arxiv.org/pdf/2212.01579…


#MetaResearch presented Controllable Agent: RL agent whose reward function can be set in real time, without any further learning or fine-tuning, based on a reward-free pretraining phase #ReinforcementLearning #AI #ML Demo: controllable-agent.metademolab.com GitHub: github.com/facebookresear…


New Reinforcement Learning Environment! If you don't have plans for the Christmas holidays yet 😉 #ReinforcementLearning #AI #MachineLearning

Sneak peak of WIP of an upcoming FPS environment for my @godotengine Reinforcement Learning library. Agents trained using async PPO and population-based training with sample-factory. 👉 github.com/edbeeching/god… It will soon be available on the @huggingface hub!



Self-supervised Learning with data2vec 2.0 - efficient self-supervised algorithm built by Meta AI for speech, vision, and text. Up to 16x faster than the most popular existing algorithm #DeepLearning #PyTorch #AI Paper: scontent-sjc3-1.xx.fbcdn.net/v/t39.2365-6/3… GitHub: github.com/facebookresear…

Tweet Image 1

LaViLa - video-language representations by leveraging Large Language Models. The authors repurpose pre-trained LLMs to be conditioned on visual input, and finetune them to create automatic video narrators #ComputerVision #NLP arxiv.org/abs/2212.04501 github.com/facebookresear…


Gymnasium is a new open source Python library for developing and comparing reinforcement learning algorithms by providing a standard API to communicate between learning algorithms and environments #ReinforcementLearning #Python farama.org/Announcing-The…


Real-time Online Video Detection with Temporal Smoothing Transformers using PyTorch #DeepLearning #ComputerVision #PyTorch Github: github.com/zhaoyue-zephyr… Paper: arxiv.org/abs/2209.09236… Dataset: paperswithcode.com/dataset/activi…

Tweet Image 1

New state of the art for methods without lookahead search, and even surpasses MuZero #ReinforcementLearning #DeepLearning #PyTorch Paper: arxiv.org/abs/2209.00588… GitHub: github.com/eloialonso/iris Dataset: paperswithcode.com/dataset/atari-…


2022 Instagram Workshop on Recommendation Systems At Scale - a great event for researchers and practitioners interested in large scale recommender and ranking systems #Personalization #Ranking #Meta #Instagram research.facebook.com/blog/2022/8/re…


Line by line explanation implementations of neural networks and related algorithms using PyTorch. An excellent opportunity for those who want to understand the details #PyTorch #neuralnetworks #DeepLearning nn.labml.ai


TorchGeo is a #PyTorch domain library, similar to #torchvision, that provides datasets, samplers, transforms, and pre-trained models specific to #geospatial data Github: github.com/microsoft/torc… Article: pytorch.org/blog/geospatia…


Loading...

Something went wrong.


Something went wrong.