dvlab-research / LLaMA-VID

Official Implementation for LLaMA-VID: An Image is Worth 2 Tokens in Large Language Models
Apache License 2.0
622 stars 39 forks source link

[h264 @ 0x871b380] mmco: unref short failure during stage-2 training #89

Open Nastu-Ho opened 1 month ago

Nastu-Ho commented 1 month ago

hey,I noticed this happens during stage-2, is this normal?

image