Unverified50% confidenceFactExact time
NVIDIA released two quantized weights: nvidia/GLM-5.3-Flash-NVFP4 (conventional post-training quantization, aimed at data centers) and nvidia/GLM-5.3-Flash-FP4-Local (quantization-aware distillation, aimed at local use).
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/18/2026
First Seen
Valid until: 1/16/2027
Sources
Related Entities
Related Claims
UnverifiedOwnership Lens deployed the nvidia/llama-3.3-nemotron-super-49b-v1.5 model via a hosted NVIDIA endpoint, benchmarked multiple models across 10 standardized tool-calling scenarios, and passed 15/15 verdict fidelity checks across three multi-turn investigation runs.74% similarUnverified混合精度训练由 NVIDIA 在 2018 年提出，同时使用 FP32 和 FP16 两种精度74% similarUnverifiedNVIDIA provides native support for nvfp4 in its TensorRT-LLM framework.74% similarUnverifiedNVIDIA Device Plugin registers the nvidia.com/gpu resource type with the Kubernetes scheduler, enabling Pods to declare GPU requirements via resources.limits.73% similarUnverifiedNVIDIA提出了非均匀张量并行（Nonuniform Tensor Parallelism）技术，用于在GPU故障时最大化LLM训练的有效产出（Goodput）73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1064223API
curl https://kongchang.com/api/v1/knowledge/claims/1064223MCP
get_claim(id=1064223)