fabxx48/GPT_OSS_20B_ArduinoExpert_4bit_v0.3

GPT OSS 20B ArduinoExpert v0.3 - 4bit / LoRA

A GPT-OSS 20B Arduino-focused LoRA/QLoRA adapter trained with Unsloth.

This model is specialized for Arduino, electronics, embedded systems, C/C++ debugging, sensor wiring, board-specific differences, and beginner-friendly troubleshooting.

Model details

• Base model: unsloth/gpt-oss-20b
• Version: v0.3
• Format: Hugging Face / PEFT / Unsloth LoRA adapter
• Language: English and Italian
• License: Apache-2.0

What this model is for

ArduinoExpert is intended to help with:

• Arduino Uno, Nano, Mega and Leonardo projects
• ESP32 and ESP8266 differences
• C/C++ Arduino compiler errors
• sensor wiring and debugging
• 3.3V vs 5V constraints
• I2C, SPI, OneWire and digital sensor protocols
• beginner-friendly explanations
• safer electronics advice

Recommended adapter

Use:

lora-adapter-latest/

Pinned v0.3 adapter:

lora-adapter-v0.3/

GGUF version

For LM Studio, llama.cpp or Ollama-style usage, use the GGUF repository:

fabxx48/GPT_OSS_20B_ArduinoExpert_v0.3_GGUF

Recommended GGUF file:

gpt-oss-20b.MXFP4.gguf

Example usage with Transformers / PEFT

Install dependencies:

pip install transformers peft accelerate

Minimal loading example:

from transformers import AutoTokenizer, AutoModelForCausalLM
from peft import PeftModel

base_model = "unsloth/gpt-oss-20b"
adapter_repo = "fabxx48/GPT_OSS_20B_ArduinoExpert_4bit_v0.3"
adapter_path = "lora-adapter-latest"

tokenizer = AutoTokenizer.from_pretrained(base_model)

model = AutoModelForCausalLM.from_pretrained(
    base_model,
    device_map="auto",
    trust_remote_code=True,
)

model = PeftModel.from_pretrained(
    model,
    adapter_repo,
    subfolder=adapter_path,
)

Example prompt

Ho un ESP32 DevKit e un sensore DS18B20.
Dammi cablaggio corretto, resistenza necessaria e codice minimo funzionante.

Training notes

Version v0.3 was trained on a cleaned Arduino dataset with stronger filtering for:

• board-specific correctness;
• safer wiring advice;
• 3.3V / 5V constraints;
• compiler and upload errors;
• ESP32 / ESP8266 toolchain differences;
• common Arduino sensors and modules.

Disclaimer

This model can make mistakes. Always verify wiring, voltage levels, current limits and component datasheets before powering real hardware.