Creating Inference Checkpoints
Overviews how to convert your inference checkpoint for various model servers
Tags: Experimentation
Fine-tuning Gemma 3-1B-it on L4
This tutorial guides you through fine-tuning the Gemma 3-1B-it language model on Google Kubernetes Engine (GKE) using L4 GPU, leveraging Parameter Efficient Fine Tuning (PEFT) and LoRA. It covers setting up a GKE cluster, containerizing the fine-tuning code, running the fine-tuning job, and uploading the resulting model to Hugging Face. Finally, it demonstrates how to deploy and interact with the fine-tuned model using vLLM on GKE.