Haobo Xu

[Haobo Xu @xuhb20]

@Shadowfetch Thanks so much! Really appreciate your interest in our work!😃

[Haobo Xu @xuhb20]

Excited to share our new work, STEPQuant, a post-training quantization framework for Delta-rule recurrent states. Comments and discussions are welcome! 😃

📄 Paper: https://t.co/KyKYPCQ9vl
💻 Code: https://t.co/IVb93L5AVW

#LLM #Quantization

[Papers with Code @HuggingPapers]

Tencent researchers just released STEPQuant on Hugging Face

A spatial-temporal post-training quantization framework for Delta-rule recurrent states, matching FP32 accuracy with 6-bit states while cutting total serving memory by up to 68.7%

[Haobo Xu @xuhb20]

@Takanas0517 @iclr_conf

[Haobo Xu @xuhb20]

@Amber_CHN @ylecun @iclr_conf @risteski_a @jacobandreas @elonmusk @cvondrick @iclr_conf same question here!!