fcan

running llms on 3090 /
not ai generated ✨

[fcan @fcan_traces]

Haters say it is fake

[fcan @fcan_traces]

@messimerize10 i know messi is a vibe-killer but it tastes surprisingly good

[fcan @fcan_traces]

@twid @0xSero about the 2nd paragraph, i mean, it is already impractical at a level on its own. that setup gives me head aches.

[fcan @fcan_traces]

@xalturicx @0x0SojalSec same for me, not a single one of them passed the vibe check.

[fcan @fcan_traces]

@sudoingX did laguna s 2.1 write this tweet

[fcan @fcan_traces]

@ucuncMahmut aglattın kardeş

[fcan @fcan_traces]

@tarih101YT abi haha, kıyamet kilisesindeki şu ahşap merdiven de aynısını hissettiriyor

[fcan @fcan_traces]

full duplex voice model moshi from @kyutai_labs trained on 100x H100 for a month 🤯

[fcan @fcan_traces]

@jesafarin atmanız gereken emojileri biliyorsunuz 🥲

[fcan @fcan_traces]

@WescheNex1q this is a great question. i mostly use llama.cpp as inference engine with -np1 flag.
Let me check with vllm.

[fcan @fcan_traces]

@ben_burtenshaw this is so good

[fcan @fcan_traces]

@lagrimacanta @vertigomsu kardes linkedine cevirdiniz burayı

[fcan @fcan_traces]

@mandylu plot twist, they are talking about dario himself

[fcan @fcan_traces]

@seanhn what kind of work?
if its some kind of data manipulation, i would export the file as csv and use llms to write deterministic python scripts.

[fcan @fcan_traces]

@nikita_builds there are number of companies doing this, maybe you can check hugging face inference providers

[fcan @fcan_traces]

@anathema_2001 katılıyorum.. patlıcan, bayrak emojiisi falan cok komik ama kanal amacının dışına çıktı twittera düştüğünden beri

[fcan @fcan_traces]

@fkadev bizim sokaktaki simitçi de bu tasarıma geçti 😭

[fcan @fcan_traces]

@jun_song @huggingface doesnt it make hugging face single point of failure?

[fcan @fcan_traces]

@mckinley_1776 ona bu kadar şuna bu kadar vergiden çok shanghai communique'ye dayanıyor kalkınma, anlamaya başlayınca gözünüzde normalleşecektir