CoVT-3expert-Stage1-LoRA

Phase1 6K-step LoRA adapter for our 3-expert CoVT reproduction on Qwen2.5-VL-7B. Clean โ€” used as the Phase1 starting point by both the strict and non-strict Phase2 runs.

Companion repos

Caveat

This is a 3-expert variant; the CoVT paper's main configuration uses 4 experts.

Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support