Jiaqi Li

jiaqili3

5 6 16

https://jiaqili3.github.io

jiaqili3

AI & ML interests

Spoken language models, Neural audio codec

Recent Activity

new activity about 5 hours ago

FlexiSLM/FlexiSLM-Data-2M-s2s-compact:Dataset Viewer issue

updated a dataset about 15 hours ago

FlexiSLM/FlexiSLM-Data-5M-t2t

liked a dataset about 15 hours ago

FlexiSLM/FlexiSLM-Data-2M-s2s-compact

View all activity

Organizations

New activity in FlexiSLM/FlexiSLM-Data-2M-s2s-compact about 5 hours ago

Dataset Viewer issue

#1 opened about 5 hours ago by

jiaqili3

updated a dataset about 15 hours ago

FlexiSLM/FlexiSLM-Data-5M-t2t

Viewer • Updated about 15 hours ago • 5.08M • 38 • 1

liked a dataset about 15 hours ago

FlexiSLM/FlexiSLM-Data-2M-s2s-compact

Updated about 11 hours ago • 86 • 1

updated a dataset about 16 hours ago

FlexiSLM/FlexiSLM-Data-2M-s2s-compact

Updated about 11 hours ago • 86 • 1

updated a dataset about 19 hours ago

FlexiSLM/FlexiSLM-Data-4M-s2s

Updated about 10 hours ago • 314 • 1

authored a paper 1 day ago

FlexiSLM: A Dynamic and Controllable Frame Rate Spoken Language Model

Paper • 2606.31247 • Published Jun 30 • 5

published a dataset 1 day ago

FlexiSLM/FlexiSLM-Data-2M-s2s-compact

Updated about 11 hours ago • 86 • 1

reacted to onekq's post with 🔥 1 day ago

Post

2302

Many developers discovered that the native DeepSeek API has higher cache-hit rate than neocloud APIs hosting the same DS models.

My speculation is that DS aggressively kills its old models. It has released 18 models thus far, and only 2 are being served now (v4 pro and flash).

This is tough to customers who don't want to upgrade (migrate or leave), but effectively boost the serving capacity to the same model, i.e. more woods behind fewer arrows.