autotrust/GLM5.3-Flash-E224-DGX-Spark

Hugging Face Models Trending Models
multimodal moe quantization vllm open-source expert-pruning

Summary

autotrust 发布的 GLM5.3-Flash-E224-DGX-Spark 模型仓库，基于 zai-org/GLM-5.3-Flash 的图像-文本到文本（image-text-to-text）量化版本，支持 NVFP4/8-bit 量化，可在 DGX Spark (GB10/Blackwell) 上通过 vLLM 运行。

Task: image-text-to-text Tags: vllm, safetensors, glm5_next, autotrust, moe, nvfp4, modelopt, dgx-spark, gb10, blackwell, expert-pruned, image-text-to-text, conversational, en, zh, base_model:zai-org/GLM-5.3-Flash, base_model:quantized:zai-org/GLM-5.3-Flash, license:mit, 8-bit, region:us
Original Article

Similar Articles

zai-org/GLM-5.3-Flash

Hugging Face Models Trending

GLM-5.3-Flash is the first natively multimodal model in the GLM-5 series, with 320B total parameters and 18B active parameters, outperforming previous versions and approaching Claude Opus 4.8 through a redesigned hybrid architecture for improved efficiency.

unsloth/GLM-5.3-Flash-GGUF

Hugging Face Models Trending

GLM-5.3-Flash is the first natively multimodal model in the GLM-5 series, with 320B total parameters and 18B active, utilizing a hybrid sparse-linear attention architecture to reduce costs while outperforming previous versions and approaching Claude Opus 4.8 on benchmarks.