AI NEWS.ccAI正在发生 未来正在更新
♡ 我的收藏 0
AI 模型库/AxionML/Qwen3.5-4B-NVFP4
MODEL PROFILE · 模型档案

Qwen3.5-4B-NVFP4

AxionML · Qwen3.5-4B-NVFP4

> Developed by AxionML for open-source serving and deployment use cases. Part of AxionML's effort to provide ready-to-serve quantized models for the community. This is an NVFP4-quantized version of Qwen/Qwen3.5-4B (4B parameters), quantized using NVIDIA TensorRT Model Optimizer. Weights and activations of linear layers are quantized to FP4, reducing disk size and GPU memory by ~4x compared to BF16. **About NVFP4 quantization:** NVFP4 on Blackwell couples a compact E2M1 FP4 codebook with blockwise FP8 (E4M3) scaling over 16-element micro-blocks, so that 4-bit stored values remain numerically useful for neural-network computation. The E2M1 codebook provides a small, nonuniform set of representable magnitudes up to ±6 and relies on saturating behavior rather than IEEE NaN/Inf encodings to maximize usable range per bit. Using an FP8 block scale (rather than power-of-two-only E8M0) enables fractional scales and error-minimizing scale selection strategies such as dual-pass evaluation comparing "map max to 6" versus "map max to 4 with clipping." On Blackwell Tensor Cores, native FP4 multipliers exploit E2M1 simplicity to reduce multiplier area while higher-precision FP32 accumulation protects dot-product accuracy.

译文排队中，当前展示来源原文。

视觉问答license:apache-2.0
上下文窗口262,144 Token
输入 → 输出文本、图像 → 文本
模型权重待核验
档案更新时间2026/11/2 19:15:46
OVERVIEW

模型介绍

> Developed by AxionML for open-source serving and deployment use cases. Part of AxionML's effort to provide ready-to-serve quantized models for the community. This is an NVFP4-quantized version of Qwen/Qwen3.5-4B (4B parameters), quantized using NVIDIA TensorRT Model Optimizer. Weights and activations of linear layers are quantized to FP4, reducing disk size and GPU memory by ~4x compared to BF16. **About NVFP4 quantization:** NVFP4 on Blackwell couples a compact E2M1 FP4 codebook with blockwise FP8 (E4M3) scaling over 16-element micro-blocks, so that 4-bit stored values remain numerically useful for neural-network computation. The E2M1 codebook provides a small, nonuniform set of representable magnitudes up to ±6 and relies on saturating behavior rather than IEEE NaN/Inf encodings to maximize usable range per bit. Using an FP8 block scale (rather than power-of-two-only E8M0) enables fractional scales and error-minimizing scale selection strategies such as dual-pass evaluation comparing "map max to 6" versus "map max to 4 with clipping." On Blackwell Tensor Cores, native FP4 multipliers exploit E2M1 simplicity to reduce multiplier area while higher-precision FP32 accumulation protects dot-product accuracy.

主要特点

特点摘要待依据来源补齐，不凭模型名称推断能力。

局限与注意事项

暂无足够来源支持具体缺点结论。请重点验证实际任务表现、时延、数据政策、许可与版本生命周期。

PRICING & COST

来源报价与费用分析

同一模型在不同渠道、地区和配置下可能采用不同价格。下表独立保留各渠道报价，不将渠道价格视为厂商官方统一价格。

来源 / 渠道输入输出计费单位采集时间
暂未取得可核验报价，请查阅来源渠道。

按月估算基础 Token 费用

—

预估不包含缓存优惠、阶梯定价、多模态附加费用、网络检索、税费及推理额外输出。

INDEPENDENT BENCHMARKS

公开评测与运行配置

暂无已采集的可核验公开评分。模型未上榜不代表能力较低。

查看完整已收录榜单 →

常见问题

开放权重是否代表可以免费商用？

不一定。应核对具体版本许可证、商用门槛和使用政策。自部署还会产生算力和维护成本。

这些价格是实时结算价格吗？

不是实时结算价格，而是带采集时间的来源报价快照。请到对应渠道核对当前版本、区域、折扣和计费方式。

资料来源与可追溯记录

huggingface · AxionML/Qwen3.5-4B-NVFP4 ↗2026/10/31 02:36:31
huggingface · enrich:f551e3fd-354e-4f5e-baad-3f0e77348c00 ↗2026/10/29 14:39:55

许可证：apache-2.0。模型介绍与第三方评测均以原来源的具体版本和方法为准。