Nebius Post-Training: Deploy LoRA Adapter

Serve your fine-tuned LoRA adapter as a serverless endpoint with per-token billing. Two deployment paths: from a Nebius fine-tuning job, or from a local archive / HuggingFace link.

Prerequisites

pip install requests openai
export NEBIUS_API_KEY="your-key"

API: https://api.tokenfactory.nebius.com

Supported base models for serverless LoRA deployment