Unverified50% confidenceSolutionExact time
Xiaomi MiMo's fix involved training a lightweight repetition-specific reinforcement learning teacher—approximately 7,000 samples over 12 steps—and then merging it into the main model via MoPD, at a cost of approximately 4% of a full retraining.
1
Sources
50%
Confidence
Long-term
Relevance
10/8/2026
First Seen
Sources
Related Entities
Related Claims
Unverified大型训练任务中因故障和恢复造成的算力浪费可达总训练时间的5%到20%63% similarUnverifiedTraining uses batch size 256, 400,000 steps, EMA 0.9999, linear learning rate decay over the final quarter, and torch.compile achieves a 2.4× speedup over eager mode.61% 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.60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/988756API
curl https://kongchang.com/api/v1/knowledge/claims/988756MCP
get_claim(id=988756)