Machinelearning – Telegram
385K subscribers
4.47K photos
864 videos
17 files
4.91K links
Погружаемся в машинное обучение и Data Science

Показываем как запускать любые LLm на пальцах.

По всем вопросам - @haarrp

@itchannels_telegram -🔥best channels

Реестр РКН: clck.ru/3Fmqri
Download Telegram
Yandex Research выложил в опенсорс RuLeanALBERT — самую большую BERT-подобную модель на русском языке, которая поместится на ваш компьютер

Нейросеть обучали децентрализованно с помощью вычислительной платформе Яндекса. На бенчмарках по пониманию языка она показывает результаты, сравнимые с другими открытыми моделями и где-то даже близкие к state-of-the-art.

Модель хотя и имеет миллиарды параметров, но вполне способна уместиться на одну домашнюю GPU: вы можете использовать её открытый код в своих проектах для классификации предложений, представления текстов и других языковых задач, не требующих генерации.

Yandex Research — это исследовательская группа в Яндексе, которая занимается фундаментальными проблемами в важнейших областях computer science.

Подробности – в статье на Хабре
👍33👎32
🛠 Generative Visual Prompt: Unifying Distributional Control of Pre-Trained Generative Models

⚙️ Github
🗒 Paper
🦾 Dataset

@ai_machinelearning_big_data
👍15🔥41
🛠 Omni-Dimensional Dynamic Convolution

A novel multi-dimensional attention mechanism with a parallel strategy to learn complementary attentions for convolutional kernels.

⚙️ Github
🗒 Paper
🦾 Dataset

@ai_machinelearning_big_data
👍9🔥51
👁 Real-time Online Video Detection with Temporal Smoothing Transformers

git clone --recursive git@github.com:zhaoyue-zephyrus/TeSTra.git

⚙️ Github
🗒 Paper
🦾 Dataset

@ai_machinelearning_big_data
👍112🔥2
This media is not supported in your browser
VIEW IN TELEGRAM
🗒 Text2Light: Zero-Shot Text-Driven HDR Panorama Generation

Text2Light can generate HDR panoramas in 4K+ resolution using free-form texts solely.

conda env create -f environment.yml
conda activate text2light


⚙️ Github
💡 Project
💻 Model
🗒 Paper
🦾 Tutorial

@ai_machinelearning_big_data
👍11🔥71
🗣 Robust Speech Recognition via Large-Scale Weak Supervision

Whisper is a general-purpose speech recognition model by Open AI.

pip install git+https://github.com/openai/whisper.git

⚙️ Github
💡 Colab
💻 Model
🗒 Paper
🦾 Dataset
✴️ HABR

@ai_machinelearning_big_data
👍22🔥21
This media is not supported in your browser
VIEW IN TELEGRAM
🔄 VToonify: Controllable High-Resolution Portrait Video Style Transfer


git clone https://github.com/williamyang1991/VToonify.git
cd VToonify


⚙️ Github
💡 Colab
💻 Project
🗒 Paper
🦾 Dataset
🎞 Video

@ai_machinelearning_big_data
👍33🔥165
🤗 SetFit - Efficient Few-shot Learning with Sentence Transformers

An efficient and prompt-free framework for few-shot fine-tuning of Sentence Transformers (ST).

python -m pip install setfit

⚙️ Github
🗒 Paper
📌 Blog
🦾 Model and Dataset

@ai_machinelearning_big_data
👍11🔥3👏1
This media is not supported in your browser
VIEW IN TELEGRAM
🔲 TensorStore

Novel open-source C++ / #Python library for storage/manipulation of high-dim data

⚙️ Github
🗒 Tutorial
📌 Google AI
🦾 Docs

@ai_machinelearning_big_data
👍2653🔥2
📎 Dilated Neighborhood Attention Transformer

natural, flexible and efficient extension to NA that can capture more global context and expand receptive fields exponentially at no additional cost.

⚙️ Github
🗒 Model
📋 Paper
📌 Dataset

@ai_machinelearning_big_data
👍15🔥31👏1
🎓 YATO: Yet Another deep learning based Text analysis Open toolkit

pip install ylab-yato

⚙️ Github
📋 Paper
📌 Dataset

@ai_machinelearning_big_data
👍173🔥1
📎 Diffusion Posterior Sampling for General Noisy Inverse Problems

git clone https://github.com/DPS2022/diffusion-posterior-sampling

cd diffusion-posterior-sampling

⚙️ Github
📋 Paper
📌 Dataset

@ai_machinelearning_big_data
👍24🔥32
🖥 One Transformer Can Understand Both 2D & 3D Molecular Data

git clone https://github.com/lsj2408/Transformer-M.git

⚙️ Github
📋 Paper
📌 Dataset

@ai_machinelearning_big_data
Please open Telegram to view this post
VIEW IN TELEGRAM
🔥17👍42
🖥 GLM-130B: An Open Bilingual Pre-trained Model

bidirectional dense model with 130 billion parameters, pre-trained using the algorithm of General Language Model (GLM).

⚙️ Github
📋 Paper
➡️ Model
➡️ Demo
📌 Dataset

@ai_machinelearning_big_data
Please open Telegram to view this post
VIEW IN TELEGRAM
👍122🔥1
🖥 Leveraging Instance Features for Label Aggregation in Programmatic Weak Supervision

git clone https://github.com/JieyuZ2/wrench.git

⚙️ Github
📋 Paper
➡️ Blog
📌 Dataset

@ai_machinelearning_big_data
Please open Telegram to view this post
VIEW IN TELEGRAM
👍101🔥1
🖥 SpeechUT: Bridging Speech and Text with Hidden-Unit for Encoder-Decoder Based Speech-Text Pre-training

SpeechT5 framework that explores the encoder-decoder pre-training for self-supervised speech/text representation learning

⚙️ Github
📋 Paper
➡️ Open SLR
📌 Dataset

@ai_machinelearning_big_data
Please open Telegram to view this post
VIEW IN TELEGRAM
👍95🔥21
🔩 ExtEnD: Extractive Entity Disambiguation

⚙️ Github
➡️ Demo
📋 Paper
📌 Dataset

@ai_machinelearning_big_data
Please open Telegram to view this post
VIEW IN TELEGRAM
11👍5🔥2
Please open Telegram to view this post
VIEW IN TELEGRAM
👍122🔥1
🖥 Bayesian Optimisation & Reinforcement Learning Research

pip install HEBO

⚙️Github: https://github.com/huawei-noah/HEBO

🗒 Docs: https://hebo.readthedocs.io/en/latest/

🖥 T-LBO: https://github.com/huawei-noah/HEBO/blob/master/T-LBO

↪️ Reinforcement Learning Research : https://github.com/huawei-noah/HEBO/tree/master/SAUTE

@ai_machinelearning_big_data
Please open Telegram to view this post
VIEW IN TELEGRAM
👍203🔥3
Please open Telegram to view this post
VIEW IN TELEGRAM
12👍7🔥2
🟡 Benchmarking ML for MD simulation

Molecular dynamics simulation with machine learning force fields.

⚙️ Github: https://github.com/kyonofx/mdsim

🗒 Paper: https://arxiv.org/abs/2210.07237v1

🖥 Guide: https://github.com/kyonofx/mdsim#install-other-dependencies

➡️ Dataset: https://paperswithcode.com/dataset/md17

@ai_machinelearning_big_data
Please open Telegram to view this post
VIEW IN TELEGRAM
👍14🔥32