Reinforcement Learning
Diffusers
Safetensors
English
image-quality-assessment
vision-language
image-editing
Instructions to use RobinY99/MR-IQA-2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use RobinY99/MR-IQA-2 with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("RobinY99/MR-IQA-2", dtype=torch.bfloat16, device_map="cuda") prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k" image = pipe(prompt).images[0] - Notebooks
- Google Colab
- Kaggle
| # Requirements | |
| Actor/Judge and Editor use separate environments. | |
| Install and verify the pinned environments through the public wrapper: | |
| ```bash | |
| bash scripts/setup_envs.sh --profile inference | |
| bash scripts/setup_envs.sh --profile test | |
| ``` | |
| Full training additionally requires `FLASH_ATTN_WHEEL`; see | |
| `environment/README.md` for the one-command form. | |
| | File | Purpose | GPU required | | |
| | --- | --- | --- | | |
| | `actor-judge.txt` | GRPO Actor training, Actor inference, and frozen E5 Judge | NVIDIA CUDA | | |
| | `editor.txt` | FLUX.2 image-editing service | NVIDIA CUDA | | |
| | `test.txt` | Static checks and CPU contract tests | No | | |
| | `publish.txt` | Hugging Face upload tooling | No | | |
| Install the compatible FlashAttention wheel into Actor/Judge and configure it | |
| through `.env.example`: | |
| ```bash | |
| python -m pip install /path/to/validated_flash_attn.whl | |
| ``` | |
| For a CPU-only development environment: | |
| ```bash | |
| python -m pip install --upgrade pip | |
| python -m pip install -r requirements/test.txt | |
| python -m pip install torch==2.11.0 \ | |
| --index-url https://download.pytorch.org/whl/cpu | |
| bash scripts/test_release.sh | |
| ``` | |
| See `environment/README.md` for the exact GPU environment setup. | |