Опівночні Балачки
412 posts
@O_Balachky
Подкаст про технології: програмування, машине навчання, драми і кеки. Питання до @ikroosh, @sudodoki, @denysxftr
Опівночні Балачки: №43: Дизайн-інтерв'ю open.spotify.com/episode/0FHVhV…
Python programming is easy to learn, right? Only if you speak English! What if you could code in your native language? We translate Python and create a prototype for exactly this. Paper: rb.gy/wqvvls #NLProc #SWE #ICSME #Python #Universal
Our new text-to-image model, DALL·E 3, can translate nuanced requests into extremely detailed and accurate images. Coming soon to ChatGPT Plus & Enterprise, which can help you craft amazing prompts to bring your ideas to life: openai.com/dall-e-3
I wrote about a 23x improvement (!) in LLM live-inference throughput, measured on OPT-13B on A100. There are 2 new innovations which make this possible: Continuous batching & PagedAttention. Short thread below; see writeup, experiments, and results at anyscale.com/blog/continuou…
Охохо ехехехе! OpenAI змерджив до себе мою Eval Model яку я побудував на основі UA-GEC: Grammatical Error Correction and Fluency Corpus for the Ukrainian Language. Тобто наступна версія ChatGPT буде евалюватись на моделі яку зробив я. А ми усі отримаємо більш якісну підтримку української в ChatGPT
I still don't know the full details (and frankly, that's part of the problem), but I wanted to explain what happened from my perspective, based on my fractured understanding.
MMS: Massively Multilingual Speech. - Can do speech2text and text speech in 1100 languages. - Can recognize 4000 spoken languages. - Code and models available under the CC-BY-NC 4.0 license. - half the word error rate of Whisper. Code+Models: github.com/facebookresear… Paper: scontent-lga3-2.xx.fbcdn.net/v/t39.8562-6/3… Blog: ai.facebook.com/blog/multiling…
