ML Research Hub – Telegram
ML Research Hub
32.7K subscribers
4K photos
228 videos
23 files
4.31K links
Advancing research in Machine Learning – practical insights, tools, and techniques for researchers.

Admin: @HusseinSheikho || @Hussein_Sheikho
Download Telegram
🚀 NAUTILUS: boosting Bayesian importance nested sampling with deep learning

A novel approach to boost the efficiency of the importance nested sampling (INS) technique for Bayesian posterior and evidence estimation using deep learning.

Install:
pip install nautilus-sampler

import corner
import numpy as np
from nautilus import Prior, Sampler
from scipy.stats import multivariate_normal

prior = Prior()
for key in 'abc':
prior.add_parameter(key)

def likelihood(param_dict):
x = [param_dict[key] for key in 'abc']
return multivariate_normal.logpdf(x, mean=[0.4, 0.5, 0.6], cov=0.01)

sampler = Sampler(prior, likelihood)
sampler.run(verbose=True)
points, log_w, log_l = sampler.posterior()
corner.corner(points, weights=np.exp(log_w), labels='abc')

🖥 Github: https://github.com/johannesulf/nautilus

⭐️ Docs: https://nautilus-sampler.readthedocs.io/

📕 Paper: https://arxiv.org/abs/2306.16923v1

https://news.1rj.ru/str/DataScienceT
6
🏌️ GlOttal-flow LPC Filter (GOLF)

A DDSP-based neural vocoder.

🖥 Github: https://github.com/yoyololicon/golf

📕 Paper: https://arxiv.org/abs/2306.17252v1

🔗Demo: https://yoyololicon.github.io/golf-demo/

https://news.1rj.ru/str/DataScienceT
❤‍🔥31👍1
This media is not supported in your browser
VIEW IN TELEGRAM
🔮 SAM-PT: Segment Anything + Tracking 🔮

⭐️ SAM-PT is the first method to utilize sparse point propagation for Video Object Segmentation (VOS).

🌐 Review https://t.ly/QLMG

🌐 Paper arxiv.org/pdf/2307.01197.pdf

🌐 Project www.vis.xyz/pub/sam-pt/

🌐 Code github.com/SysCV/sam-pt

https://news.1rj.ru/str/DataScienceT
❤‍🔥11👍1
❤‍🔥2
This media is not supported in your browser
VIEW IN TELEGRAM
🪩 DISCO: Human Dance Generation

⭐️ NTU (+ #Microsoft) unveils DISCO: a big step towards the Human Dance Generation.

🌐 Review https://t.ly/cNGX

🌐 Paper arxiv.org/pdf/2307.00040.pdf

🌐Project: disco-dance.github.io/

🌐 Code github.com/Wangt-CN/DisCo

https://news.1rj.ru/str/DataScienceT
👍31
Building an Image Recognition API using Flask.

Step 1: Set up the project environment

1. Create a new directory for your project and navigate to it.
2. Create a virtual environment (optional but recommended):
(Image 1.)
3. Install the necessary libraries (image 2.)

Step 2: Create a Flask Web Application
Create a new file called app.py in the project directory (image 3.)

Step 3: Launch the Flask Application
Save the changes and run the Flask application (image 4.)

Step 4: Test the API
Your API is now up and running and you can send images to /predict via HTTP POST requests.
You can use tools such as curl or Postman to test the API.
• An example of using curl (image 5.)
• An example using Python queries (image 6.)

https://news.1rj.ru/str/DataScienceT
❤‍🔥2😍2👍1
Deep Learning
NLP
AI
Python
ML
Data Mining
Tensorflow
Keras

👇👇👇👇👇
@Machine_learn
❤‍🔥32😍1
This media is not supported in your browser
VIEW IN TELEGRAM
🚀 Hierarchical Open-vocabulary Universal Image Segmentation

Decoupled text-image fusion mechanism and representation learning modules for both "things" and "stuff".

🖥 Github: https://github.com/berkeley-hipie/hipie

📕 Paper: https://arxiv.org/abs/2307.00764v1

🔗Project: http://people.eecs.berkeley.edu/~xdwang/projects/HIPIE/

🔗 Dataset: https://paperswithcode.com/dataset/pascal-panoptic-parts

https://news.1rj.ru/str/DataScienceT
❤‍🔥1
❤‍🔥3
We launched a special bot some time ago to download all scientific, software and mathematics books The bot contains more than thirty million books, and new books are downloaded first, In addition to the possibility of downloading all articles and scientific papers for free

To request a subnoscription: talk to @Hussein_Sheikho
3👍3
This media is not supported in your browser
VIEW IN TELEGRAM
🎨 Making ML-powered web games with Transformers.js

The goal of this tutorial is to show you how easy it is to make your own ML-powered web game.

🖥 Github: https://github.com/xenova/doodle-dash

🤗 Hugging face: https://huggingface.co/blog/ml-web-games

⭐️ Code: https://github.com/xenova/doodle-dash

🔗Demo: https://huggingface.co/spaces/Xenova/doodle-dash

🔗 Dataset: https://huggingface.co/datasets/Xenova/quickdraw-small

https://news.1rj.ru/str/DataScienceT
1👍1
🦙 Focused Transformer: Contrastive Training for Context Scaling

LongLLaMA, a large language model capable of handling long contexts of 256k tokens or even more.

🖥 Github: https://github.com/cstankonrad/long_llama

📕 Paper: https://arxiv.org/abs/2307.03170v1

🖥 Colab: https://colab.research.google.com/github/CStanKonrad/long_llama/blob/main/long_llama_colab.ipynb

🔗 Dataset: https://paperswithcode.com/dataset/pg-19

https://news.1rj.ru/str/DataScienceT
👍31❤‍🔥1
🖥 Chat Downloader

A simple tool used to retrieve chat messages from livestreams, videos, clips and past broadcasts.

- YouTube.com
- Zoom.us
- Facebook.com
- Twitch.tv

$ pip install chat-downloader

Using:
# termimal
$ chat_downloader https://www.youtube.com/watch?v=video_link --output chat.json


# Python noscript
from chat_downloader import ChatDownloader

url = 'https://www.youtube.com/watch?v=video_link'
chat = ChatDownloader().get_chat(url)

for message in chat:
chat.print_formatted(message)


🖥 Github
📝 Docs

https://news.1rj.ru/str/DataScienceT
5👍2