Instructions to use igor-im/flux_prompt_expander with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use igor-im/flux_prompt_expander with Transformers:

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

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

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("igor-im/flux_prompt_expander", dtype="auto")

PEFT
How to use igor-im/flux_prompt_expander with PEFT:
```
Task type is invalid.
```

llama-cpp-python

How to use igor-im/flux_prompt_expander with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="igor-im/flux_prompt_expander",
	filename="unsloth.F16.gguf",
)

llm.create_chat_completion(
	messages = [
		{
			"role": "user",
			"content": "What is the capital of France?"
		}
	]
)

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use igor-im/flux_prompt_expander with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf igor-im/flux_prompt_expander:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf igor-im/flux_prompt_expander:Q4_K_M

Install from WinGet (Windows)

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

Use Docker

docker model run hf.co/igor-im/flux_prompt_expander:Q4_K_M

LM Studio
Jan

vLLM

How to use igor-im/flux_prompt_expander with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/igor-im/flux_prompt_expander:Q4_K_M

SGLang

How to use igor-im/flux_prompt_expander 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 "igor-im/flux_prompt_expander" \
    --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": "igor-im/flux_prompt_expander",
		"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 "igor-im/flux_prompt_expander" \
        --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": "igor-im/flux_prompt_expander",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use igor-im/flux_prompt_expander with Ollama:
```
ollama run hf.co/igor-im/flux_prompt_expander:Q4_K_M
```

Unsloth Studio new

How to use igor-im/flux_prompt_expander 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 igor-im/flux_prompt_expander 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 igor-im/flux_prompt_expander to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for igor-im/flux_prompt_expander to start chatting

Pi new

How to use igor-im/flux_prompt_expander with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf igor-im/flux_prompt_expander:Q4_K_M

Configure the model in Pi

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

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use igor-im/flux_prompt_expander with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama-server -hf igor-im/flux_prompt_expander: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 igor-im/flux_prompt_expander:Q4_K_M

Run Hermes

hermes

Docker Model Runner
How to use igor-im/flux_prompt_expander with Docker Model Runner:
```
docker model run hf.co/igor-im/flux_prompt_expander:Q4_K_M
```

Lemonade

How to use igor-im/flux_prompt_expander with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull igor-im/flux_prompt_expander:Q4_K_M

Run and chat with the model

lemonade run user.flux_prompt_expander-Q4_K_M

List all available models

lemonade list

v 0.2 Trained for 1 epoch ~2500 steps. Generates good prompt but doesn't fully follow input prompt

Uploaded model

Developed by: igor-im
License: apache-2.0
Finetuned from model : unsloth/llama-3.2-3b-instruct-bnb-4bit

This llama model was trained 2x faster with Unsloth and Huggingface's TRL library.

Samples

Jaggernaut XL - rundiffusion Foocus

Plain Prompt	Prompt exapander

Claude generated prompt	Llama 3B generated Prompt

Flux - huggingface inference

Plain Prompt	Prompt exapander

Claude generated prompt	Llama 3B generated Prompt

Prompts used to generated Samples

Plain Prompt

A portrait of a man looking off into a distance, he has long hair, and a short beard

Fine Tuned Prompt (generated by simply passing plain prompt to the model)

Imagine a super-realistic portrait of an enigmatic middle-aged man with a rugged appearance and a close-cropped beard that frames his intense, deep-set blue eyes. He has an air of wisdom and contemplation as he gazes pensively into the distance, exuding a sense of resilience and strength. From an eye-level perspective, observe his slightly furrowed brow and the subtle creases around his mouth, a testament to his life experience. The soft yet direct lighting accentuates the depth and texture of his features, creating an intimate and powerful connection with the subject.

Claude prompt (generated via following template: create a prompt to be used with an image generation ai to better guide it to generate based on user input: /n INPUT: {plain_prompt})

Create a highly detailed portrait photograph of a contemplative man gazing toward the horizon. Subject has flowing shoulder-length hair with natural movement and texture, paired with a neatly trimmed short beard. Capture in natural lighting during golden hour, with soft rim lighting highlighting his hair's texture. Shot with a high-end portrait lens at f/2.8 for pleasing bokeh. Composition follows rule of thirds, with subject positioned to direct attention toward his eyeline. Color palette emphasizes warm, muted tones. Style reference: Modern fashion photography meets classical portraiture. Technical details: Ultra high resolution, photorealistic quality, masterful depth of field, sharp focus on facial features. 8K, highly detailed, professional photography, award-winning portrait.

Llama 3.2 3B untuned prompt (generated via following template: create a prompt to be used with an image generation ai to better guide it to generate based on user input: /n INPUT: {plain_prompt})

Generate an image of a realistic portrait of a man in his 40s or 50s, standing in front of a background with a subtle blur effect on the distant landscape. The subject's facial features should be well-defined, and he should have medium-length (not too long nor too short) hair that falls around his shoulders. His beard should be short (e.g., V-shaped) and neatly trimmed, with a slight curl at the edges. Incorporate some subtle shading and texture to give the image depth and dimensionality

Downloads last month: 61

Safetensors

Model size

3B params

Tensor type

BF16

igor-im
/

flux_prompt_expander