User seeks alternatives to Unsloth for post-training custom datasets

A Reddit user is asking for methods to fine-tune local models with custom datasets without using Unsloth. The user currently runs the llama.cpp server and wants to tune a model for specific work areas.

• The user has experience running local models on Raspberry Pi.
• They are looking for pointers on post-training techniques other than Unsloth.