微調與部署 Gemma-4 的試煉與磨難 [P]

Hey all, Our ML team spent some time this week getting training and deployments working for Gemma-4, and wanted to document all the things we ran into along the way. PEFT doesn't recognize Gemma 4's custom layers. Google wrapped vision/audio projections in a new ClippableLinear class that doesn't inherit from nn.Linear, so PEFT refuses to attach LoRA, even for text-only fine-tuning. Fix: unwrap the wrappers after loading weights but before calling PEFT. SFTTrainer killed training silently.