Partially Verified75% confidenceFactExact time
NVFP4 occupies 4 bits per parameter, i.e., 0.5 bytes, theoretically halving weight memory usage and load bandwidth compared with FP8.
3
Sources
75%
Confidence
Long-term
Relevance
10/15/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedIn FP16 format each parameter occupies 2 bytes, while in INT4 format it occupies only 0.5 bytes, theoretically achieving a 4x compression ratio.78% similarUnverifiedNVFP4 is actually approximately 3.75 bits, and is not strictly a 4-bit quantization in the rigorous sense.75% similarUnverifiedTaking the Q4_0 format as an example, every 32 FP32 weights are packed into one quantization block containing one scale factor and 32 4-bit integers, averaging approximately 4.5 bits per parameter, reducing memory usage to approximately 1/8 of the original.74% similarUnverifiedCompared to FP16, FP8 reduces memory footprint and bandwidth consumption by half, and the AIC's compute throughput for FP8 can reach twice that of FP1672% similarUnverifiedBitNet 1.58b quantization uses ternary weights {-1, 0, +1} in PCEI's 24-layer network, reducing parameter memory by about 4 times compared with FP16 precision, and at inference time integer addition can replace floating-point multiplication.70% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1032580API
curl https://kongchang.com/api/v1/knowledge/claims/1032580MCP
get_claim(id=1032580)