doing something – Telegram

doing something

@doing_something

712 subscribers

910 photos

36 videos

36 files

1.94K links

@smlkw doing something, just my notes to keep updates

Download Telegram

About

Blog

Apps

Platform

doing something

712 subscribers

doing something

https://github.com/bips-hb/innsight

#ai

GitHub - bips-hb/innsight: Interpretability methods to analyze the behavior and individual predictions of modern neural networks…

Interpretability methods to analyze the behavior and individual predictions of modern neural networks in R. - bips-hb/innsight

👍1

192 views07:34

doing something

https://arxiv.org/abs/2408.16589

#ai #asr

CrisperWhisper: Accurate Timestamps on Verbatim Speech Transcriptions

We demonstrate that carefully adjusting the tokenizer of the Whisper speech recognition model significantly improves the precision of word-level timestamps when applying dynamic time warping to...

181 views14:28

doing something

Forwarded from Hacker News

TeX and Typst: Layout Models (2024)
Article, Comments

TeX and Typst: Layout Models | Laurenz's Blog

An exploration of the layout models of TeX and Typst.

159 views14:19

doing something

doing something

https://github.com/rust-dd/iTransformer #ai #rust

https://github.com/lucidrains/iTransformer

Python version also exists

#python #ai

GitHub - lucidrains/iTransformer: Unofficial implementation of iTransformer - SOTA Time Series Forecasting using Attention networks…

Unofficial implementation of iTransformer - SOTA Time Series Forecasting using Attention networks, out of Tsinghua / Ant group - lucidrains/iTransformer

203 views13:21

doing something

https://github.com/tensorchord/VectorChord-bm25/releases/tag/0.1.1

There is pg_search extension for PostgreSQL from ParadeDB but TensorChord (authors of pgvecto.rs) also made the same thing.

#pgsql

Release 0.1.1 · tensorchord/VectorChord-bm25

Fix stopword and long tokens
What's Changed

docs: update README to enhance getting started section and clarify tokenizer support by @VoVAllen in #37
docs: update wrong guc default value in rea...

474 views17:01

doing something

https://github.com/ml-explore/mlx/releases/tag/v0.23.0

#ai #mlx

Release v0.23.0 · ml-explore/mlx

Highlights

4-bit Mistral 7B generates at 131 toks/sec out of the box on an M2 Ultra
More performance improvements across the board:

Faster small batch quantized matmuls. Speeds up speculative dec...

❤1

474 views17:49

doing something

Forwarded from Viacheslav

всім привіт. виклав свої тулзи для speech generation evaluation в паблік: https://github.com/balacoon/speech_gen_eval
і зробив leaderboard на основі: https://huggingface.co/spaces/balacoon/TTSLeaderboard.
більш детально тут: https://balacoon.com/blog/tts_leaderboard/

пишіть якщо хочете додати якусь модельку.

цікаво б ще було зробити testset для українських TTS систем, і додати в leaderboard.
це звісно не TTSArena, але повчально. в Leaderboard можна послухати семпли side-by-side тому доволі практично.

GitHub - balacoon/speech_gen_eval: Collection of tools to run objective evaluation of speech generation

Collection of tools to run objective evaluation of speech generation - balacoon/speech_gen_eval

🔥2

286 views12:38

doing something

https://github.com/huggingface/transformers/releases/tag/v4.49.0

#ai

Release v4.49.0: Helium, Qwen2.5-VL, SuperGlue, Granite Vision, Zamba2, GOT-OCR 2.0, DAB-DETR, Depth Pro, RT-DETRv2, GPTQModel…

New models
Helium
Helium-1 preview is a lightweight language model with 2B parameters, targeting edge and mobile devices. It supports the following languages: English, French, German, Italian, Port...

487 views17:53

doing something

https://github.com/roboflow/maestro

#ai #cv

GitHub - roboflow/maestro: streamline the fine-tuning process for multimodal models: PaliGemma 2, Florence-2, and Qwen2.5-VL

streamline the fine-tuning process for multimodal models: PaliGemma 2, Florence-2, and Qwen2.5-VL - roboflow/maestro

425 views20:39

doing something

https://er.ucu.edu.ua/bitstreams/1faace86-5855-4737-af2c-d8559cbec52a/download

The author used my model to make evaluations

#asr #ai

👍3

156 views21:51

doing something

https://huggingface.co/knowledgator

#ai #nlp

knowledgator (Knowledgator Engineering)

Text classification, relations extraction, NER, computational biology

134 views07:25

doing something

qwen architecture is so lil 🥺

#ai #nlp

Please open Telegram to view this post

VIEW IN TELEGRAM

127 views07:52

doing something

A quick check of https://huggingface.co/Ihor/Text2Graph-R1-Qwen2.5-0.5b model

Colab: https://colab.research.google.com/drive/1Cgv6DcrNj-j6BsP3GiCKPWbMmOP8tkeM?usp=sharing

#ai #nlp

❤1

122 views07:57

doing something

doing something

https://huggingface.co/knowledgator #ai #nlp

https://huggingface.co/collections/knowledgator/moderngliner-676ac733bec0aa50e5923c00

btw, ModernBERT is cool

#nlp

❤2

124 views07:59

doing something

why it is so painful...

https://huggingface.co/datasets/Helsinki-NLP/opus-100

#nlp

😢1

144 views09:11

doing something

doing something

why it is so painful... https://huggingface.co/datasets/Helsinki-NLP/opus-100 #nlp

https://huggingface.co/collections/jbochi/madlad-400-65491e6a78726cac9a4b84b7

#nlp

MADLAD-400 - a jbochi Collection

Models and spaces for MADLAD-400: A Multilingual And Document-Level Large Audited Dataset

273 views09:25

doing something

doing something

why it is so painful... https://huggingface.co/datasets/Helsinki-NLP/opus-100 #nlp

https://github.com/MicrosoftTranslator/NTREX

No ы in Ukrainian split

#nlp

GitHub - MicrosoftTranslator/NTREX: NTREX -- News Test References for MT Evaluation

NTREX -- News Test References for MT Evaluation. Contribute to MicrosoftTranslator/NTREX development by creating an account on GitHub.

126 views09:33

doing something

https://github.com/benbrandt/text-splitter

#rust #ai #nlp

GitHub - benbrandt/text-splitter: Split text into semantic chunks, up to a desired chunk size. Supports calculating length by characters…

Split text into semantic chunks, up to a desired chunk size. Supports calculating length by characters and tokens, and is callable from Rust and Python. - benbrandt/text-splitter

464 views09:39

doing something

https://github.com/segment-any-text/wtpsplit

#ai #nlp

GitHub - segment-any-text/wtpsplit: Toolkit to segment text into sentences or other semantic units in a robust, efficient and adaptable…

Toolkit to segment text into sentences or other semantic units in a robust, efficient and adaptable way. - segment-any-text/wtpsplit

136 views09:43

doing something

https://huggingface.co/cardiffnlp

#nlp #ai

cardiffnlp (Cardiff NLP)

Natural Language Processing

137 viewsedited 13:01

doing something

https://predibase.com/blog/teaching-ai-to-write-gpu-code-a-deep-dive-into-reinforcement-fine-tuning

#ai #nlp

Train AI to Write GPU Code via Reinforcement Fine-Tuning

In this post, we’ll discuss how we taught an AI model to convert PyTorch code into efficient Triton kernels using a reinforcement learning algorithm inspired by PPO called Group Relative Preference Optimization (GRPO).

143 views13:56