Machinelearning – Telegram
383K subscribers
4.46K photos
861 videos
17 files
4.9K 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
This media is not supported in your browser
VIEW IN TELEGRAM
🔥 SeqDeepFake: Detecting and Recovering Sequential DeepFake Manipulation

First Seq-DeepFake dataset, where face images are manipulated sequentially with corresponding annotations of sequential facial manipulation vectors.

Github: https://github.com/rshaojimmy/seqdeepfake

Project: https://rshaojimmy.github.io/Projects/SeqDeepFake

Paper: https://arxiv.org/pdf/2207.02204.pdf

Dataset: https://paperswithcode.com/dataset/imagenet

@ai_machinelearning_big_data
🔥10👍4🥰1
🔀 No Language Left Behind

Meta's open-sources models capable of delivering high-quality translations directly between any pair of 200+ languages.

Github: https://github.com/facebookresearch/fairseq/tree/nllb

Paper: https://research.facebook.com/publications/no-language-left-behind/

Website: https://ai.facebook.com/research/no-language-left-behind/

Demo: https://nllb.metademolab.com/

@ai_machinelearning_big_data
👍10
👍8👎3
⬆️ YOLOv7

YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors

Github: https://github.com/wongkinyiu/yolov7

Paper: https://arxiv.org/abs/2207.02696v1

Dataset: https://paperswithcode.com/dataset/coco

@ai_machinelearning_big_data
👍157🎉3
👀 Object Centric Open Vocabulary Detection

Object-centric alignment of the language embeddings from the CLIP model.

Github: https://github.com/hanoonaR/object-centric-ovd

Paper: https://arxiv.org/abs/2207.03482v1

Dataset: https://paperswithcode.com/dataset/imagenet

@ai_machinelearning_big_data
👍14🔥2
👍19🔥5👎1
🔸 An Efficiency Study for SPLADE Models

SParse Lexical AnD Expansion Model for First Stage Ranking.

Github: https://github.com/naver/splade

Paper: https://arxiv.org/abs/2207.03834v1

Dataset: https://paperswithcode.com/dataset/ms-marco

@ai_machinelearning_big_data
👍12
🧍‍♂ PeopleSansPeople: A Synthetic Data Generator for Human-Centric Computer Vision

Human-centric privacy-preserving synthetic data generator with highly parametrized domain randomization.

Github: https://github.com/unity-technologies/peoplesanspeople

Paper: https://arxiv.org/abs/2207.05025v1

Demo Video: https://www.youtube.com/watch?v=mQ_DUdB70dc

@ai_machinelearning_big_data
👍13
🚀 Dynamic Low-Resolution Distillation for Cost-Efficient End-to-End Text Spotting

Github: https://github.com/hikopensource/davar-lab-ocr

Paper: https://arxiv.org/abs/2207.06694v1

Dataset: https://paperswithcode.com/dataset/total-text

@ai_machinelearning_big_data
👍84
Language Modelling with Pixels

PIXEL is a language model that operates on text rendered as images, fully removing the need for a fixed vocabulary.

Github: https://github.com/xplip/pixel

Paper: https://arxiv.org/abs/2207.06991v1

Dataset: https://paperswithcode.com/dataset/glue

Pretrained: https://huggingface.co/Team-PIXEL/pixel-base

@ai_machinelearning_big_data
👍20😢2
⚡️ CLOSE: Curriculum Learning On the Sharing Extent Towards Better One-shot NAS

Github: https://github.com/walkerning/aw_nas

Paper: https://arxiv.org/abs/2207.07868v1

Dataset: https://paperswithcode.com/dataset/nas-bench-201

@ai_machinelearning_big_data
👍15😢1
HiFormer: Hierarchical Multi-scale Representations Using Transformers for Medical Image Segmentation

Github: https://github.com/amirhossein-kz/hiformer

Paper: https://arxiv.org/abs/2207.08518v1

Tasks: https://paperswithcode.com/task/medical-image-segmentation

@ai_machinelearning_big_data
👍11🎉1