Unverified50% confidenceBenchmarkExact time
Meta's Llama 2 (70-billion-parameter version) used 1.8 million GPU hours for training.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/9/2026
First Seen
Valid until: 12/8/2026
Sources
Related Entities
Related Claims
UnverifiedLarge-scale pretraining often requires thousands of high-end GPUs (such as NVIDIA H100) running continuously for months, with the training cluster's power consumption reaching tens of megawatts.72% similarUnverifiedMeta在OPT-175B训练日志中披露，175天的训练过程中经历了超过100次硬件相关的中断事件，平均每1-2天出现一次需要人工干预的故障71% similarVerifiedFine-tuning requires hundreds to thousands of high-quality training samples, hours of GPU compute time, and ongoing maintenance costs, and the entire process must be repeated whenever knowledge is updated.70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/884582API
curl https://kongchang.com/api/v1/knowledge/claims/884582MCP
get_claim(id=884582)