Unverified50% confidenceBenchmarkExact time
The Tencent team compressed a 3.3 GB translation model down to 440 MB, a compression ratio of nearly 87%, with almost no visible degradation in BLEU score.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/7/2026
First Seen
Valid until: 12/6/2026
Sources
Related Entities
Related Claims
Unverified动态GGUF的UD-IQ2_XXS档位将原始BF16模型体积从54.7GB压缩到约9GB，体积压缩率83.5%，准确率保留82.5%74% similarUnverifiedA 140 GB FP16 70B model can be compressed to approximately 35–40 GB after 4-bit quantization, reducing communication overhead by about 75%; 2-bit quantization can compress it to approximately 18 GB but introduces accuracy loss.73% similarUnverifiedThe original QW3.6 27B model file size is approximately 54 GB; Prismo compresses it to under 4 GB through mathematical compression techniques, a compression ratio of more than 13×, and claims this does not affect model performance.73% similarUnverified在ComfyUI视频抠图工作流中，将图片尺寸从1024降至512可将APNG文件从160MB压缩至30多MB，压缩比接近80%73% similarUnverifiedBonsai compresses a 27B model from 54GB at full precision down to 3.9GB72% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/869548API
curl https://kongchang.com/api/v1/knowledge/claims/869548MCP
get_claim(id=869548)