clearml
Fine-tuning a tiny model to talk like a pirate, then running it on my Mac
Fine-tuning a 0.5B Qwen model to talk like a pirate on a homelab k3s cluster's GPU through ClearML, then running it locally with mlx-lm.
Fine-tuning a 0.5B Qwen model to talk like a pirate on a homelab k3s cluster's GPU through ClearML, then running it locally with mlx-lm.
Installing ClearML on a k3s cluster, adding the RTX 3090 as a GPU worker, and the problems along the way.