Subiecte populare
#
Bonk Eco continues to show strength amid $USELESS rally
#
Pump.fun to raise $1B token sale, traders speculating on airdrop
#
Boop.Fun leading the way with a new launchpad on Solana.
Rahul
Membru fondator și conducător al echipei de modele a fundației @sarvamai.
Acum că echipa noastră LLM de 15 membri este faimoasă, e timpul să ne extindem data viitoare!
Dacă ai făcut una sau mai multe dintre următoarele variante, te rugăm să ne contactezi.
- pre-antrenat un model de orice dimensiune, de la zero
- Post-antrenat orice model de bază, de la un capăt la altul (Data Curation, SFT, RL)
- sunt vrăjitori cu torță
- sunt un maestru de kernel CUDA
- ai alte abilități relevante și muncă care să le susțină
<at>Prenumele: Sarvam<dot>Ai
14
Dezvăluire a lore-ului: asta este adevărat. Introduceam slide-uri benchmark 105B în deck ÎN TIMP @pratykumar vorbeam despre cifrele 30B pe scenă!
19 feb. 2026
Ok, so for anyone who's upset about the @SarvamAI models not beating GPT5/other 1T parameter models, my response: are you f kidding me? A year ago you were crying about India not having our own LLMs trained from scratch. The team has launched 14 amazing products in 14 days and it took years of work, each. And they span text, OCR, voice, LLMs, STT, hardware! I mean hardware! Edge models! Celebrate. They are just getting started. Sure, I am biased as their first investor, but we invested because we thought so highly of them. Of course we are biased. And the odds were enormous. But the opportunity is enormous-er :-) and these guys are absolute killers, and the team hunts! I haven't chatted with @pratykumar today but I heard someone say he saw that PPT for today's LLM launch session literally ON THE STAGE for the first time and laid it all out without a prompter. You can only do it when you know your material cold, and your heart is 100% living and breathing this stuff. Just incredible, congrats to everyone at @SarvamAI. Start of an era.
1
Limită superioară
Clasament
Favorite