Instructions to use gohumanize/gohumanize-open-humanizer with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use gohumanize/gohumanize-open-humanizer with Transformers:

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

pipe = pipeline("text-generation", model="gohumanize/gohumanize-open-humanizer")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("gohumanize/gohumanize-open-humanizer")
model = AutoModelForCausalLM.from_pretrained("gohumanize/gohumanize-open-humanizer", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use gohumanize/gohumanize-open-humanizer 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 gohumanize/gohumanize-open-humanizer:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf gohumanize/gohumanize-open-humanizer:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf gohumanize/gohumanize-open-humanizer:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf gohumanize/gohumanize-open-humanizer: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 gohumanize/gohumanize-open-humanizer:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf gohumanize/gohumanize-open-humanizer: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 gohumanize/gohumanize-open-humanizer:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf gohumanize/gohumanize-open-humanizer:Q4_K_M

Use Docker

docker model run hf.co/gohumanize/gohumanize-open-humanizer:Q4_K_M

LM Studio
Jan

vLLM

How to use gohumanize/gohumanize-open-humanizer with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "gohumanize/gohumanize-open-humanizer"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "gohumanize/gohumanize-open-humanizer",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/gohumanize/gohumanize-open-humanizer:Q4_K_M

SGLang

How to use gohumanize/gohumanize-open-humanizer 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 "gohumanize/gohumanize-open-humanizer" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "gohumanize/gohumanize-open-humanizer",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "gohumanize/gohumanize-open-humanizer" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "gohumanize/gohumanize-open-humanizer",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use gohumanize/gohumanize-open-humanizer with Ollama:
```
ollama run hf.co/gohumanize/gohumanize-open-humanizer:Q4_K_M
```
Unsloth Desktop

How to use gohumanize/gohumanize-open-humanizer with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf gohumanize/gohumanize-open-humanizer:Q4_K_M

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "gohumanize/gohumanize-open-humanizer:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use gohumanize/gohumanize-open-humanizer with Docker Model Runner:
```
docker model run hf.co/gohumanize/gohumanize-open-humanizer:Q4_K_M
```

Lemonade

How to use gohumanize/gohumanize-open-humanizer with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull gohumanize/gohumanize-open-humanizer:Q4_K_M

Run and chat with the model

lemonade run user.gohumanize-open-humanizer-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use gohumanize/gohumanize-open-humanizer with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf gohumanize/gohumanize-open-humanizer:Q4_K_M

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default gohumanize/gohumanize-open-humanizer:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use gohumanize/gohumanize-open-humanizer with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf gohumanize/gohumanize-open-humanizer:Q4_K_M

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "gohumanize/gohumanize-open-humanizer:Q4_K_M" \
  --custom-provider-id llama-cpp \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

GoHumanize Open Humanizer

A small open model that rewrites AI-styled English prose into more natural human writing. It is a QLoRA fine-tune of Qwen3-4B on 2,000 pairs of (AI-styled passage, human original) built from public-domain books (Project Gutenberg).

The Open Humanizer is a public research and educational model created to demonstrate the general approach used to develop AI text humanization systems. It is separate from the production models used by GoHumanize.ai, but it reflects many of the same high-level principles: careful dataset preparation, transformation of source text into training pairs, model fine-tuning, evaluation and iterative improvement. It is not intended to reproduce the architecture, data, configuration or performance of the production systems, and it makes no claim about AI detectors.

Everything about the project (dataset, code, evaluation, write-up): https://gohumanize.ai/research

Use

Chat format, thinking disabled. The system prompt below is the one used in training.

from transformers import AutoTokenizer, AutoModelForCausalLM
import torch

repo = "gohumanize/gohumanize-open-humanizer"
tok = AutoTokenizer.from_pretrained(repo)
model = AutoModelForCausalLM.from_pretrained(repo, torch_dtype=torch.bfloat16, device_map="auto")

SYSTEM = ("Rewrite the following text so that it reads as if a person wrote it: varied sentence "
          "length, concrete wording, natural rhythm, no filler transitions. Keep the meaning, the "
          "facts and the order of ideas. Return only the rewritten text.")
messages = [{"role": "system", "content": SYSTEM},
            {"role": "user", "content": "It is worth noting that the committee ultimately reached a consensus."}]
ids = tok.apply_chat_template(messages, add_generation_prompt=True, enable_thinking=False, return_tensors="pt").to(model.device)
out = model.generate(ids, max_new_tokens=600, temperature=0.7, top_p=0.9, do_sample=True)
print(tok.decode(out[0][ids.shape[1]:], skip_special_tokens=True))

Also available: a GGUF build in gguf/ for Ollama / LM Studio / llama.cpp, an MCP server (npx gohumanize-open-humanizer-mcp), a Python client (pip install gohumanize-open-humanizer) and a browser demo at https://gohumanize.ai/research.

Training


Base model	Qwen/Qwen3-4B (Apache-2.0)
Method	QLoRA with Unsloth: 4-bit base, LoRA rank 16, alpha 32, all attention and MLP projections
Data	2,000 train / 200 test pairs, see the dataset card
Loss	on the assistant (human target) tokens only
Epochs, LR, batch	2 epochs (250 steps), 2e-4 cosine, effective batch 16, max 1,024 tokens
Hardware	1x NVIDIA A10G (24 GB) on Modal, 21.7 minutes
Tracking	Weights & Biases run
Final losses	train 1.30, eval 1.39 (3.17 before training)

The LoRA adapter is in lora/; the main files are the merged 16-bit weights.

Evaluation

Base Qwen3-4B vs this model on the 200 held-out pairs, against the human original.

Measure (200 held-out pairs)	AI-styled input	Base Qwen3-4B	Open Humanizer	Human target
BERTScore F1 vs human (higher = closer meaning)	0.914	0.900	0.921
ROUGE-L vs human (higher = closer wording)	0.493	0.424	0.540
Names/capitalised tokens kept (recall)	0.587	0.594	0.623
Length ratio vs human	1.027	0.830	0.928	1.000
Contractions per 100 words	0.279	0.772	0.044	0.118
Transition words per 100 words	0.757	0.023	0.106	0.143
Stock LLM phrases per text	0.19	0.01	0.00	0.00
Average sentence length (words)	21.6	15.2	22.5	28.1

These measure how far the output moves from AI-styled prose towards the human target. They are not detector scores.

Limitations

The human targets are pre-1929 prose, so the model leans towards a literary, slightly old-fashioned register.
Trained on 80 to 300 word passages; rewrite long documents paragraph by paragraph.
English only. Can drop or alter details on inputs unlike its training data; check facts in the output.
Educational demo; no claims about AI detectors.

Licence and citation

Apache-2.0. Cite as:

GoHumanize team (2026). GoHumanize Open Humanizer: an open text-humanization model, dataset and pipeline built from public-domain data. Version 0.1.0. Zenodo. https://doi.org/10.5281/zenodo.22843083

Downloads last month: -

Safetensors

Model size

4B params

Tensor type

BF16

Model tree for gohumanize/gohumanize-open-humanizer

Base model

Qwen/Qwen3-4B-Base

Finetuned

Qwen/Qwen3-4B

Adapter

(1158)

this model

gohumanize
/

gohumanize-open-humanizer

GoHumanize Open Humanizer

Use

Training

Evaluation

Limitations

Links

Licence and citation

Model tree for gohumanize/gohumanize-open-humanizer

Dataset used to train gohumanize/gohumanize-open-humanizer