Unverified50% confidenceOpinionExact time
The 6.3-million-parameter network is only about the size of an early convolutional classification model.
1
Sources
50%
Confidence
Long-term
Relevance
10/18/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedIn the 120B-A12B model, the total model weights are 120 billion parameters, but only about 12 billion parameters actually participate in computation during each token's forward pass.66% similarUnverifiedIn the experiment, ResNet-56 has only 2,752 more parameters than plain-56 (approximately one-third of a percent of the total), and these extra parameters come solely from 1×1 projection layers at locations where the number of channels changes.65% similarUnverifiedA model with 100 billion parameters requires approximately 80 GB of VRAM to process a single inference request, while a model with tens of billions of parameters requires only about 20 GB.64% similarUnverifiedA model with 11,000 parameters occupies less than 50KB of memory.64% similarUnverifiedGraphite counted 12,877 distinct tics across nine models, of which 65% appear in only a single model series.63% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1058682API
curl https://kongchang.com/api/v1/knowledge/claims/1058682MCP
get_claim(id=1058682)