ML4 reportedly trained on 3,800 NVIDIA Grace Blackwell GPUs
A repost places the claimed training in European datacenters and mentions a cluster in Bruyères-le-Châtel.
TLDR
A repost of Guillaume Lample’s post claims ML4 was trained on 3,800 NVIDIA Grace Blackwell GPUs in European datacenters and mentions a cluster in Bruyères-le-Châtel. Another post repeated the hardware count without naming ML4. The attached posts do not describe ML4’s capabilities, training duration or release status.
Combined views
203.7K
14 Sources, first seen ago

Comments