Suchergebnisse fรผr "๐งต@cass03821โatpํ๋งค์ฒโโญ"
Kurzvideos
Videos

Besties really pulled up UNINVITED ๐ญ๐ Part 2/4 ๐งต๐
Asmeeyโฝ๏ธ๐11,643,890 Aufrufe โข vor 2 Tagen

She got everyone talking bout it Thread ๐งต in comment ๐
Gowies PR4,179,768 Aufrufe โข vor 4 Tagen

Introducing Contrastive Language Model (CLM): an ultra-fast System One Model trained with a contrastive learning objective that connects states and actions. CLM-8B is pre-trained on internet-scale data and delivers up to 9ร faster inference than Jev โก while achieving comparable performance across computer-use, gaming, and tool-calling tasks. With lightweight fine-tuning, CLM-8B sets a new SOTA on challenging agentic coding benchmarks, such as DeepSWE (81.6%) and Terminal-Bench 2.1 (87.6%). In contrast, Jev fails to serve as an effective verifier for these long-horizon tasks. We also build an efficient training and serving infra for CLMs by disaggregating states and actions, allowing their embeddings to be cached and reused independently. This substantially reduces inference latency in settings where the state evolves continuously while the action set remains fixed. Finally, we establish scaling laws for CLMs and show that the test contrastive loss decreases predictably as a power law in training compute, model size, and dataset size. ๐ Blog: ๐ป Code: ๐ฃ๏ธ Discord: ๐ค Data & Models: More details on CLMโs architecture, data recipe, and scaling laws in the thread below ๐งต
Jacky Kwok928,813 Aufrufe โข vor 1 Tag

Introducing ScienceBuddy โ a free workspace for scientific agents that improve through researcher collaboration. Use GPT-6 in ScienceBuddy at no cost. GPU-accelerated, and fused with the JEV framework. ๐งต Two loops: ๐น Inner loop โ refines the agent harness ๐น Outer loop โ trains the model with rubric-guided RL Together: Recursive-in-Recursive Self-Improvement. ScienceBuddy explores how scientific agents can improve through sustained collaboration with researchers. ๐ฌTry it free: #ScienceBuddy #PhAILabs #AI4Science
PhAI Labs1,166,272 Aufrufe โข vor 2 Tagen

ํธ๋ฆฌํ์์ค '๋ณด๋ํ๋' ๊น์ฑ์ฐ์ด ๋ค๋ฆ๊ฒ ์ธ์คํ๋ก ์ง์ ๋ฐํ ํญํ ๋ ผ๋์ ์ ๋ง 1/ ๋ณด๋ํ๋ ๋น๋น๋งจ(์ต์์)-๊น์ฑ์ฐ ํญํ ๋ ผ๋, 1๋ 4๊ฐ์ ๋ง์ ๊น์ฑ์ฐ์ด ์ง์ ์ธ์คํ๋ก ์ ํ ํด๋ช ์์ฝ ํ๋ ๐งต 2/ 2019๋ ๋ผ์ด๋ธ ๋ฐฉ์ก ์ค '๋ฏธ์ฑ๋ ์ ํญํ' ์ํน์ด ํฐ์ง๋ฉฐ ์ต์์๋ ๋ฐฉ์ก์์ ์ฆ๊ฐ ํ์ฐจํ๊ณ ์ดํ ๊ฒ์ฐฐ์์ ๋ฌดํ์ ์ฒ๋ถ์ ๋ฐ์์. ํ์ง๋ง ๋น์ฌ์์๋ ํ๋(๊น์ฑ์ฐ)์ ์ง์ ์ ์ธ ์ค๋ช ์ ์ข ์ ์์ ๊น์ง ๋์ค์ง ์์๋ ์ํฉ. 3/ 2021๋ 4์ ๊น์ฑ์ฐ์ ์ธ์คํ๋ฅผ ํตํด "๋น์ ์ผ๋ก ์์ฒ๋ฅผ ๋ฐ์ ์ ์ ์ ํ ์์๋ค"๋ฉฐ ๋จ์ ํดํ๋์ด์์์ ๋ช ํํ ๋ฐํ. ์ต์์์๋ ์ฌ์ ํ ์ฐ๋ฝ์ ์ฃผ๊ณ ๋ฐ์ผ๋ฉฐ ์ข์ ์ ํ๋ฐฐ๋ก ์ ์ง๋ด๊ณ ์๋ค๊ณ ์ ํจ. 4/ ์ด์ด ๋น์ ์ ์์์ฌ ๋ช ์์ ๊ณต์ง๋ง ๋ด๊ณ ๋ณธ์ธ์ด ์ง์ ๋น ๋ฅด๊ฒ ํด๋ช ํ์ง ๋ชปํ๋ ์ ์ ํํํ๋ฉฐ, ์ค๋ ์๊ฐ์ด ํ๋ ์ง๋ง ์ด์ ๋ผ๋ ์คํด๊ฐ ์์ ํ ํ๋ฆฌ๊ธธ ๋ฐ๋๋ค๋ฉฐ ์ฌ๊ณผ๋ฅผ ๋ง๋ถ์. ์คํด๊ฐ ํ๋ ค์ ๋คํ์ด๊ณ ๋ช์ด์ง๋ฆฌ ์์์ ํ์ฅ์ด ์ ๋ง ์ปธ๋ค์.
๋๋ฒ ๋ฅด๋ง657,799 Aufrufe โข vor 1 Tag

This is happening right now in New York. Thousands of people transferring their sins into live chickens. Then the chickens are slaughtered and sent to the same stores as any other chicken. Someone is eating a bird a stranger used as a vessel for their guilt. ๐งต (film link in thread)
Kameron Waters5,138,643 Aufrufe โข vor 7 Tagen

๐งตThe White House repeatedly begged Volodymyr Zelenskyy to stop bombing Russian refineries. Forty-eight hours later, Ukrainian drones incinerated Rosneft's Kuibyshev plant, wiping out 1.4 million tons of annual diesel capacity before dawn. 1/๐งต
Prof. Bonk ๐บ๐ฆ Perverse & Unintended Consequences541,352 Aufrufe โข vor 1 Tag

Seguro has visto este video como el "peor fail del deporte" o un simple meme de risa, pero la historia real es brutal. รl es John Elmerson Fabriga y sacรณ un 0.00 perfecto tras caer de espalda. Tras las burlas, la verdad avergonzรณ a todo su paรญs. La cruda realidad abajo. ๐งต๐
PE_layo7,089,364 Aufrufe โข vor 10 Tagen

Introducing Diarization Bench, Voice Arena's benchmark for who spoke when. The metric we measure is diarization error rate, or DER. It measures how much of a conversation a model attaches to the wrong person. Every second it misses, invents, or gives to another speaker counts against it. Here's what we did. 139 conversations recorded for this board, in rooms and on calls, with people interrupting each other throughout. 12 systems scored under a single goal: testing models in the real-world conditions they are actually deployed in. NVIDIA's Nemotron 3 leads at 14.72%. ๐งต
Shobhit Banga629,621 Aufrufe โข vor 2 Tagen

Homer had a lyre. You have Grok Imagine. Create a compelling scene from Homerโs The Odyssey that shows what Grok @Imagineโs video and voice capabilities can do. Weโre awarding $100K, $50K, and $25K to the top three videos submitted by quoting this post. ๐งต
Grok35,127,419 Aufrufe โข vor 1 Monat

