Unverified50% confidenceBenchmarkExact time
The per-text cosine similarity between the int8 quantized repack and the original float32 version ranges from 0.9888 to 0.9978; the cosine similarity between bf16 L4 and the original float32 reaches 0.9999 to 1.0000.
1
Sources
50%
Confidence
Long-term
Relevance
10/13/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedFP32 uses 32 bits to represent floating-point numbers, consisting of 1 sign bit, 8 exponent bits, and 23 mantissa bits.64% similarUnverified传统Stable Diffusion使用CLIP-L/14作为文本编码器，嵌入维度768，最大token长度77，参数量约1.2亿60% similarUnverifiedINT4 has only 16 discrete values available, while INT8 has 256 discrete values.59% similarUnverified445 字节是经过 base64 编码后的最终大小，实际压缩后的二进制数据约 333 字节57% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1014141API
curl https://kongchang.com/api/v1/knowledge/claims/1014141MCP
get_claim(id=1014141)