Instructions to use Thox-ai/thox-micro-125m with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Thox-ai/thox-micro-125m with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="Thox-ai/thox-micro-125m")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Thox-ai/thox-micro-125m")
model = AutoModelForCausalLM.from_pretrained("Thox-ai/thox-micro-125m", device_map="auto")

llama-cpp-python

How to use Thox-ai/thox-micro-125m with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="Thox-ai/thox-micro-125m",
	filename="thox-micro-125m.Q4_0.gguf",
)

output = llm(
	"Once upon a time,",
	max_tokens=512,
	echo=True
)
print(output)

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use Thox-ai/thox-micro-125m with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf Thox-ai/thox-micro-125m:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf Thox-ai/thox-micro-125m:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf Thox-ai/thox-micro-125m:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf Thox-ai/thox-micro-125m:Q4_K_M

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf Thox-ai/thox-micro-125m:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf Thox-ai/thox-micro-125m:Q4_K_M

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf Thox-ai/thox-micro-125m:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf Thox-ai/thox-micro-125m:Q4_K_M

Use Docker

docker model run hf.co/Thox-ai/thox-micro-125m:Q4_K_M

LM Studio
Jan

vLLM

How to use Thox-ai/thox-micro-125m with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Thox-ai/thox-micro-125m"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Thox-ai/thox-micro-125m",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/Thox-ai/thox-micro-125m:Q4_K_M

SGLang

How to use Thox-ai/thox-micro-125m with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "Thox-ai/thox-micro-125m" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Thox-ai/thox-micro-125m",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "Thox-ai/thox-micro-125m" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Thox-ai/thox-micro-125m",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Ollama
How to use Thox-ai/thox-micro-125m with Ollama:
```
ollama run hf.co/Thox-ai/thox-micro-125m:Q4_K_M
```

Unsloth Studio

How to use Thox-ai/thox-micro-125m with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for Thox-ai/thox-micro-125m to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for Thox-ai/thox-micro-125m to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for Thox-ai/thox-micro-125m to start chatting

Atomic Chat new
Docker Model Runner
How to use Thox-ai/thox-micro-125m with Docker Model Runner:
```
docker model run hf.co/Thox-ai/thox-micro-125m:Q4_K_M
```

Lemonade

How to use Thox-ai/thox-micro-125m with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull Thox-ai/thox-micro-125m:Q4_K_M

Run and chat with the model

lemonade run user.thox-micro-125m-Q4_K_M

List all available models

lemonade list

ThoxMicro-125M

ThoxMicro-125M is THOX.ai's smallest from-scratch base language model: a 123,587,328-parameter decoder-only transformer, pretrained on a single RTX 4060 Ti. It is sized for the THOX 512 MB device tier (ThoxMini / ThoxMini Air / ThoxClip, Raspberry Pi Zero class), where chat-class models do not fit but a small local next-token predictor is useful.

This repository holds the PyTorch / safetensors weights. Quantized GGUF builds for on-device inference live in Thox-ai/thox-micro-125m-GGUF.

License: Apache-2.0 · IP: THOX IP-017 · Author: THOX.ai LLC

Honest description

This is a base / completion model. It is not instruction-tuned, has no chat template, carries no THOX persona or curated facts, and will not reliably answer general-knowledge questions. It is a foundation for SFT/LoRA work and a local building block at the edge. Prompt it in completion style and use a repetition penalty of ~1.2+ to avoid degeneration.

Model details

Field	Value
Parameters	123,587,328
Architecture	Llama-family decoder (RMSNorm, SwiGLU, RoPE)
Layers / hidden / heads	12 / 768 / 12 (full MHA, 12 KV heads)
Head dim / FFN	64 / 2048
Context length	1024
Tokenizer	GPT-2 / `r50k_base` BPE, vocab 50,304 (padded)
Tied embeddings	yes
RoPE theta	10,000
Precision	fp16 weights (BF16 mixed-precision training)
Training steps	100,000
Best validation loss	2.6099
Final validation loss	2.6459
Hardware	1 × NVIDIA RTX 4060 Ti 16 GB

Usage

import torch
from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("Thox-ai/thox-micro-125m", dtype=torch.float32)
tok = AutoTokenizer.from_pretrained("Thox-ai/thox-micro-125m")

ids = tok("The capital of France is", return_tensors="pt")
out = model.generate(**ids, max_new_tokens=32, do_sample=True,
                     temperature=0.8, top_k=40, repetition_penalty=1.3)
print(tok.decode(out[0], skip_special_tokens=True))

For device / CPU inference use the GGUF builds:

llama-completion -m thox-micro-125m.q4_0.gguf \
  -p "The capital of France is" -n 32 --repeat-penalty 1.3 --temp 0.8

Quantization

Published in Thox-ai/thox-micro-125m-GGUF:

Quant	Size	Note
Q4_0	77.6 MB	Device target — scalar-safe on ARMv6 (no NEON needed)
Q4_K_M	80.7 MB	Higher-quality alternate (ARMv7+/x86)
Q8_0	127.0 MB	Higher quality, ~1.6× the weights
F16	237.5 MB	Re-quant source

Q4_0 perplexity 22.60 vs F16 21.41 (~6%) on a neutral English sample — clean quantization. Unlike THOX's 327M (whose FFN dim of 3413 forces K-quants only), this model's FFN dim of 2048 divides evenly, so the ARM-friendly legacy Q4_0 and Q8_0 formats both apply.

Note: some files named thox-micro-125m.*.gguf with uppercase quant suffixes also exist in this repository from an earlier upload. The canonical GGUFs are the lowercase-named files in Thox-ai/thox-micro-125m-GGUF — that is what the THOX device images pin by digest.

Provenance

Trained from scratch by THOX.ai LLC — no base model, no distillation. THOX IP-017. Inventors of record are listed in NOTICE.

Downloads last month: 2

Safetensors

Model size

0.1B params

Tensor type

F16