| | --- |
| | library_name: tensorrt-rtx |
| | license: apache-2.0 |
| | base_model: black-forest-labs/FLUX.1-dev |
| | tags: |
| | - tensorrt-rtx |
| | - flux1 |
| | - fp4 |
| | - dev |
| | - optimized |
| | inference: false |
| | --- |
| | |
| | # FLUX1 TensorRT-RTX: DEV-Fp4 🔨 Building |
| |
|
| | Optimized TensorRT-RTX engines for **FLUX1** on **Fp4** architecture with **DEV** quantization. |
| |
|
| | ## 🎯 This Repository |
| |
|
| | **One variant, one download** - only get exactly what you need! |
| |
|
| | - **Model**: FLUX1 |
| | - **Architecture**: Fp4 (Compute Capability 8.0+) |
| | - **Quantization**: DEV |
| | - **Memory**: TBD |
| | - **Speed**: TBD for 1024x1024 generation |
| |
|
| | ## 🚀 Quick Start |
| |
|
| | ### Automatic (Recommended) |
| |
|
| | ```bash |
| | # ImageAI server downloads automatically |
| | curl -X POST "http://localhost:8001/generate" \ |
| | -H "Content-Type: application/json" \ |
| | -d '{ |
| | "prompt": "a beautiful landscape", |
| | "model": "flux1-tensorrt_rtx:dev", |
| | "width": 1024, |
| | "height": 1024 |
| | }' |
| | ``` |
| |
|
| | ### Manual Download |
| |
|
| | ```python |
| | from huggingface_hub import snapshot_download |
| | |
| | # Download this specific variant only |
| | engines_path = snapshot_download( |
| | repo_id="imgailab/flux1-trtx-dev-fp4-blackwell" |
| | ) |
| | |
| | # Engines are in: engines_path/engines/*.plan |
| | ``` |
| |
|
| | ### Direct Integration |
| |
|
| | ```python |
| | from imageai_server.tensorrt.nvidia_sdxl_pipeline import NVIDIASDXLPipeline |
| | |
| | pipeline = NVIDIASDXLPipeline() |
| | pipeline.load_engines( |
| | engine_dir=f"{engines_path}/engines", |
| | framework_model_dir=f"{engines_path}/framework", |
| | onnx_dir=f"{engines_path}/onnx" |
| | ) |
| | pipeline.activate_engines() |
| | |
| | images, time_ms = pipeline.infer( |
| | prompt="a serene mountain landscape", |
| | height=1024, |
| | width=1024 |
| | ) |
| | ``` |
| |
|
| | ## 📊 Performance |
| |
|
| | | Metric | Value | |
| | |--------|-------| |
| | | **Memory Usage** | TBD | |
| | | **Inference Speed** | TBD | |
| | | **Resolution** | 1024x1024 (optimized) | |
| | | **Batch Size** | 1 (optimized) | |
| | | **Precision** | DEV | |
| |
|
| | ## 🔧 Requirements |
| |
|
| | ### Hardware |
| | - **GPU**: Fp4 architecture |
| | - Ampere: RTX 3090, A100, etc. |
| | - Ada Lovelace: RTX 4090, etc. |
| | - Blackwell: H200, etc. |
| | - **VRAM**: TBD minimum |
| | - **Compute Capability**: 8.0+ |
| |
|
| | ### Software |
| | - **TensorRT-RTX**: 1.0.0.21+ |
| | - **CUDA**: 12.0+ |
| | - **Python**: 3.8+ |
| |
|
| | ## 📁 Repository Structure |
| |
|
| | ``` |
| | flux1-trtx-dev-fp4-blackwell/ |
| | ├── engines/ # TensorRT engine files |
| | │ ├── *.plan # Optimized engines |
| | ├── config.json # Configuration metadata |
| | └── README.md # This file |
| | ``` |
| |
|
| | ## 🌐 Related Repositories |
| |
|
| | Other variants for FLUX1: |
| | - [Ampere BF16](https://huggingface.co/imgailab/flux1-trtx-bf16-ampere)\n- [Ada FP8](https://huggingface.co/imgailab/flux1-trtx-fp8-ada)\n- [Ada BF16](https://huggingface.co/imgailab/flux1-trtx-bf16-ada)\n- [Blackwell FP4](https://huggingface.co/imgailab/flux1-trtx-fp4-blackwell)\n- [Blackwell FP8](https://huggingface.co/imgailab/flux1-trtx-fp8-blackwell)\n- [Blackwell BF16](https://huggingface.co/imgailab/flux1-trtx-bf16-blackwell)\n |
| |
|
| | ## 📝 License |
| |
|
| | Inherits license from base model: [black-forest-labs/FLUX.1-dev](https://huggingface.co/black-forest-labs/FLUX.1-dev) |
| |
|
| | ## 🔄 Updates |
| |
|
| | - **2025-08-12**: Initial release |
| | - Optimized for single-variant downloads |
| |
|
| | --- |
| |
|
| | *Part of the ImageAI TensorRT-RTX engine collection* |
| |
|
