Unverified50% confidenceFactExact time
After quantizing Gemma 4 E2B's embedding table to int4, the model size dropped from 6.32 GiB to 2.85 GiB, with the reduction coming mainly from the embedding table rather than the Transformer body.
1
Sources
50%
Confidence
Long-term
Relevance
10/12/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe M5's INT8 compute is twice that of BF16, and a project on GitHub using the M5's INT8 compute kernels has reported a 1.6× improvement.62% similarUnverified瑞芯微RK3588和联发科天玑系列已将AI推理能力下沉至百元级模组56% similarUnverifiedOn the English-German translation task, Transformer achieved a BLEU score of 28.4, more than two points higher than the previous best ensemble model result.56% similarUnverifiedThe model is a Latent Flow Transformer with only 2.4 to 4 million parameters, quantized to int8 precision.55% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1011098API
curl https://kongchang.com/api/v1/knowledge/claims/1011098MCP
get_claim(id=1011098)