7 8 74

Abror Shopulatov PRO

murodbek

https://murodbek.substack.com/

AI & ML interests

Machine Learning, NLP, NMT

Recent Activity

upvoted a paper about 1 hour ago

Multilingual Machine Translation with Open Large Language Models at Practical Scale: An Empirical Study

liked a Space about 4 hours ago

HuggingFaceFW/blogpost-fineweb-v1

liked a dataset 4 days ago

tahrirchi/uzlib

View all activity

Organizations

murodbek's activity

upvoted a paper about 1 hour ago

Multilingual Machine Translation with Open Large Language Models at Practical Scale: An Empirical Study

Paper • 2502.02481 • Published Feb 4 • 13

liked a Space about 4 hours ago

905

FineWeb: decanting the web for the finest text data at scale

🍷

Generate high-quality web text data for LLM training

liked a dataset 4 days ago

tahrirchi/uzlib

Viewer • Updated 9 days ago • 1.86k • 273 • 3

updated a dataset 9 days ago

tahrirchi/uzlib

Viewer • Updated 9 days ago • 1.86k • 273 • 3

liked a model 25 days ago

nvidia/stt_uz_fastconformer_hybrid_large_pc

Automatic Speech Recognition • Updated Feb 18 • 101 • 6

upvoted a collection 29 days ago

UzLLM

Collection

A collection of Uzbek-adapted LLMs. • 4 items • Updated Dec 4, 2024 • 5

New activity in tahrirchi/uz-crawl about 1 month ago

O'zbek tili manbalarning `ranking` xaqida

#2 opened about 2 months ago by

Sardor-Odil

liked a model about 2 months ago

issai/tilmash

Translation • Updated Nov 29, 2023 • 425 • 3

reacted to hexgrad's post with 🔥 about 2 months ago

Post

7037

Wanted: Peak Data. I'm collecting audio data to train another TTS model:
+ AVM data: ChatGPT Advanced Voice Mode audio & text from source
+ Professional audio: Permissive (CC0, Apache, MIT, CC-BY)

This audio should *impress* most native speakers, not just barely pass their audio Turing tests. Professional-caliber means S or A-tier, not your average bloke off the street. Traditional TTS may not make the cut. Absolutely no low-fi microphone recordings like Common Voice.

The bar is much higher than last time, so there are no timelines yet and I expect it may take longer to collect such mythical data. Raising the bar means evicting quite a bit of old data, and voice/language availability may decrease. The theme is *quality* over quantity. I would rather have 1 hour of A/S-tier than 100 hours of mid data.

I have nothing to offer but the north star of a future Apache 2.0 TTS model, so prefer data that you *already have* and costs you *nothing extra* to send. Additionally, *all* the new data may be used to construct public, Apache 2.0 voicepacks, and if that arrangement doesn't work for you, no need to send any audio.

Last time I asked for horses; now I'm asking for unicorns. As of writing this post, I've currently got a few English & Chinese unicorns, but there is plenty of room in the stable. Find me over on Discord at rzvzn: https://discord.gg/QuGxSWBfQy