Expired90% confidenceFactTime unknown
The training of Composer 2 used four GPU clusters distributed around the world, and also requisitioned inference GPUs from the production environment during off-peak hours.
1
Sources
90%
Confidence
Medium-term (~90 days)
Relevance
5/31/2026
First Seen
This claim has expired. It is kept for the record and no longer describes the current state.
Valid until: 8/29/2026(expired)
Sources
Related Entities
Related Claims
UnverifiedThe training of Composer 2 utilized four clusters worldwide.76% similarUnverifiedTechniques such as mixed-precision training and gradient accumulation enable training of larger models under limited GPU memory.70% similarVerifiedTraining a model at the level of GPT-4 requires tens of thousands of A100/H100 GPUs working in concert for several months.67% similarUnverifiedGPTQ于2022年采用逐层二阶优化将已训练模型压缩至INT4，是训练后量化的代表66% similarUnverified85% of the compute resources in Composer 2.5's training were invested in additional continued pre-training and reinforcement learning post-training.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/5803API
curl https://kongchang.com/api/v1/knowledge/claims/5803MCP
get_claim(id=5803)