Table of Contents

Property TrainVisionProjector

Namespace
LMKit.Finetuning
Assembly
LM-Kit.NET.dll

TrainVisionProjector

Trains the vision projector (the module that maps the vision encoder's features into the language model's input space) together with the language adapters, on runs whose samples carry images. The projector receives LoRA pairs of rank VisionProjectorRank on its weight matrices, trained at VisionProjectorLearningRate; the vision encoder blocks before it stay frozen. The pairs are saved in the same adapter file, and every model that applies the adapter encodes its images through the trained projector. False (the default) keeps the projector frozen.

public bool TrainVisionProjector { get; set; }

Property Value

bool

Remarks

Supported for vision models whose projector is the Qwen3-VL family's patch merger (Qwen3-VL, Qwen3.5); training any other projector fails the run with a FinetuningException naming the cause. The vision weights must reside on the same device as the language model.

Share