FastLoRA v4.2 — Fine-tuning library that never crashes (pip installable)

Hi everyone! I built FastLoRA , a drop-in alternative for fine-tuning LLMs.

bash

pip install "fastlora[full]"

Why I built it:

• Unsloth kept breaking on install and crashing mid-training

• Wanted something that just works, no matter what

What it does:

• Never crashes — every error is caught, reported and recovered automatically

• Auto hardware detection — scans your GPU and applies best settings

• Unlimited model size — 1B to 1T+, automatic strategy selection

• Every feature is a True/False toggle + 0.0–1.0 power control

• Compiled kernel cache — 3min compile once, 5 seconds after

Quick start:

python

from fastlora import FastLoRA

fl = FastLoRA("meta-llama/Llama-3.2-3B", lora=True, quantization="4bit")
model, tokenizer = fl.load()

Benchmarked on Tesla T4. Also benchmarked Unsloth on the same setup — Unsloth didn’t run.

Would love feedback from the communit