Unverified50% confidenceBenchmarkExact time
On 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.
1
Sources
50%
Confidence
Long-term
Relevance
9/11/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedB200采用Blackwell架构，引入第二代Transformer Engine和FP4精度支持，推理吞吐量较H100提升最高30倍71% similarVerifiedBlackwell's second-generation Transformer Engine adds FP4 precision support, which in theory can improve Transformer model training throughput by 2–4x compared to the H100.66% similarUnverifiedTransformer-based layout analysis models such as LayoutLMv3 and DiT (Document Image Transformer) have reached the accuracy level of commercial systems.62% similarVerifiedNVIDIA's H100 features a Transformer Engine specifically designed for AI workloads, delivering nearly 4,000 TFLOPS of compute at FP8 precision, approximately 3 times the improvement over the A100.60% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/908612API
curl https://kongchang.com/api/v1/knowledge/claims/908612MCP
get_claim(id=908612)