Today @MeckaAI is announcing $60M in funding to become the data and deployment layer for physical AI
This raise will allow us to scale our data infrastructure, invest into new verticals, and deploy robots into the real world
Mecka is excited to power EgoVerse: a growing ecosystem for robot learning from egocentric human data
Proven across 4 leading research labs, EgoVerse data consistently boosts robot performance
We release the tools for anyone to collect data, inference/train, and contribute! 🧵
Excited to announce our $8M seed round led by @neo to further our work as the data layer for robotics
Heres why we bet early on human data advancing autonomous robotics:
@CapnSlipp@nvidia 70b runs but is slow on the m4 pro. It runs at 6-7tps which is good for a tech demo but too slow for actual work.
If you run R1's 32b models on the 5090, you'll get like 40tps (really good), but on m4 pro it's about 12tps.
Source- using m4 pro with 48gb
@zjasper @Kiaman121564x @deepseek_ai@hyperbolic_labs@ollama Which setups would fit the full r1 model? A 670b parameter quantized q4 is likely close to 400gigs. A single h100 is not enough