Robot Uprising
04.03.202522:01
https://huggingface.co/spaces/nanotron/ultrascale-playbook
#Training #AI #Paper #software #tool
#Training #AI #Paper #software #tool
25.02.202500:27
“Terms of Service” (2024)
Two AI-powered characters trained on a 4Chan dataset talk to each other for eternity on livestream. Only viewer reporting or TikTok content moderation AI will shut the artwork down, at which point it will be considered complete.
#AI #LLMs #NSFW
Two AI-powered characters trained on a 4Chan dataset talk to each other for eternity on livestream. Only viewer reporting or TikTok content moderation AI will shut the artwork down, at which point it will be considered complete.
#AI #LLMs #NSFW
24.02.202520:45
TVs at HUD Played an AI-Generated Video of Donald Trump Kissing Elon Musk’s Feet
Federal employees at the Department of Housing and Urban Development (HUD) were greeted this morning by television sets at the agency’s Washington, DC, headquarters playing an AI-generated video of President Donald Trump kissing the feet of Elon Musk.
🔗
#AI #USA #politics #hacktivism #genAI #video
Federal employees at the Department of Housing and Urban Development (HUD) were greeted this morning by television sets at the agency’s Washington, DC, headquarters playing an AI-generated video of President Donald Trump kissing the feet of Elon Musk.
🔗
#AI #USA #politics #hacktivism #genAI #video
24.02.202515:16
#meme #text #lewd
25.02.202500:10
24.02.202516:47
I finally learned the actual controversy behind this and I couldn't be more disappointed. The "AI they used", was a non-generative tool to tweak the accents, and used they used it to create some background prop drawings of buildings.
"Come on guys please guys if we're just dogmatically uncompromising and judgmental forever we're definitely gonna get the genie back in the bottle guys come on."
I cannot wait until people are normal about this in 2-6 years.
"Come on guys please guys if we're just dogmatically uncompromising and judgmental forever we're definitely gonna get the genie back in the bottle guys come on."
I cannot wait until people are normal about this in 2-6 years.
26.02.202500:42
#nsfw
24.02.202515:55
Physical Intelligence (π) is open sourcing π_0, the first general-purpose robotic foundation model.
https://www.pi.website/blog/openpi
#Robotics #FOSS
https://www.pi.website/blog/openpi
#Robotics #FOSS
25.02.202516:12
Human level sample efficiency? LIMO: Less is More for Reasoning
- LIMO achieves unprecedented performance in mathematical reasoning with only 1% of the training data used by previous approaches, showcasing remarkable data efficiency.
- LIMO exhibits exceptional out-of-distribution generalization, outperforming models trained on 100x more data by a significant 40.5% absolute improvement across diverse benchmarks.
LIMO Hypothesis: In foundation models with comprehensively encoded domain knowledge (achieved through extensive pre-training), sophisticated reasoning can emerge through minimal, precisely orchestrated demonstrations of cognitive processes.
- The core of LIMO's success lies in the meticulous curation of a small, high-quality dataset. The resulting dataset of 817 examples was carefully selected from millions of candidates.
- LIMO fundamentally challenges the assumption that massive datasets are necessary for complex reasoning in LLMs. Quality of the examples, rather than just the number, is the key factor.
- LIMO suggests that modern, well-pretrained models like Qwen already possess latent, rich reasoning capabilities. LIMO demonstrates that these capabilities can be unlocked and activated effectively with the right "cognitive templates" provided by curated examples.
- LIMO indicates that sophisticated reasoning, regardless of complexity, could potentially be activated with minimal samples given sufficient pre-trained domain knowledge and optimal cognitive reasoning chains for activation.
Further research is needed to validate the LIMO hypothesis across different model architectures and reasoning domains beyond mathematics.
📄 Paper
#LLM #Research #Math #GenAi #Training #LIMO
- LIMO achieves unprecedented performance in mathematical reasoning with only 1% of the training data used by previous approaches, showcasing remarkable data efficiency.
- LIMO exhibits exceptional out-of-distribution generalization, outperforming models trained on 100x more data by a significant 40.5% absolute improvement across diverse benchmarks.
LIMO Hypothesis: In foundation models with comprehensively encoded domain knowledge (achieved through extensive pre-training), sophisticated reasoning can emerge through minimal, precisely orchestrated demonstrations of cognitive processes.
- The core of LIMO's success lies in the meticulous curation of a small, high-quality dataset. The resulting dataset of 817 examples was carefully selected from millions of candidates.
- LIMO fundamentally challenges the assumption that massive datasets are necessary for complex reasoning in LLMs. Quality of the examples, rather than just the number, is the key factor.
- LIMO suggests that modern, well-pretrained models like Qwen already possess latent, rich reasoning capabilities. LIMO demonstrates that these capabilities can be unlocked and activated effectively with the right "cognitive templates" provided by curated examples.
- LIMO indicates that sophisticated reasoning, regardless of complexity, could potentially be activated with minimal samples given sufficient pre-trained domain knowledge and optimal cognitive reasoning chains for activation.
Further research is needed to validate the LIMO hypothesis across different model architectures and reasoning domains beyond mathematics.
📄 Paper
#LLM #Research #Math #GenAi #Training #LIMO
25.02.202505:17
#AI #meme #comic
24.02.202520:50
24.02.202515:21
#meme #text #lewd
显示 1 - 16 共 16
登录以解锁更多功能。