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
File size: 2,113 Bytes
d13a83d | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 | CONDA_SH=/path/to/miniconda3/etc/profile.d/conda.sh
CONDA_ENV_NAME=mr_iqa_actor_judge
ACTOR_PYTHON=/path/to/miniconda3/envs/mr_iqa_actor_judge/bin/python
ACTOR_MODEL_PATH=/path/to/Qwen3.5-4B
NATIVE_ACTOR_MODEL_PATH=/path/to/Qwen3.5-4B
NO_MASK_ACTOR_MODEL_PATH=/path/to/no_mask/checkpoint-30
MASK_ACTOR_MODEL_PATH=/path/to/mask/checkpoint-30
ACTOR_PROCESSOR_PATH=/path/to/Qwen3.5-4B
# Prepared initial-Actor experiment identity recorded in mode_matrix.json.
ACTOR_MODEL_TREE_SHA256=eb67981666c5fc926c829b0d2b49b180c46dc6a74c1c8df895834a54c0fe33f3
TRAIN_IMAGE_ROOT=/path/to/training_images
EVAL_IMAGE_ROOT=/path/to/evaluation_images
DIFFUSERS_VENV=/path/to/editor_venv
DIFFUSERS_MODEL_PATH=/path/to/huggingface_download/editor
DIFFUSERS_WARMUP_IMAGE=
JUDGER_PYTHON=/path/to/miniconda3/envs/mr_iqa_actor_judge/bin/python
JUDGE_MODEL_ID=source-e5-judge-step725
JUDGE_MODEL_PATH=/path/to/huggingface_download/judge
JUDGE_MANIFEST_PATH=/path/to/huggingface_download/judge/provenance.json
# Full promoted source-checkpoint identity used by the reward/cache protocol.
JUDGE_MODEL_TREE_SHA256=e25415173aacf515e97d5d561c6647a7a84f586061f3a9b2ab3fc079fe21be0a
# Relocatable digest of the 10 public inference files downloaded from the Hub.
JUDGE_MODEL_EXPORT_TREE_SHA256=21b232a1a30dc765f3e7cf16c00fd270e4be354615fea0120e32f975e2777e5c
JUDGER_PROMPT_SCHEMA=e5_training_reasoning_v5
JUDGE_PROMPT_HASH=fa78a4ccfd2194a2026ff0b6b722bf22b28f8fa060389c57c4adb1618ac280f6
JUDGER_BACKEND=e5_qwen35_4b_vllm_judge
ORIGINAL_SCORE_CACHE_PATH=/path/to/locally_generated_original_scores.sqlite
ORIGINAL_SCORE_CACHE_SHA256=
ORIGINAL_SCORE_CACHE_EXPECTED_ROW_COUNT=7000
ORIGINAL_SCORE_CACHE_EXPECTED_SAMPLE_COUNT=7000
ORIGINAL_SCORE_CACHE_EXPECTED_ACTOR_IDS=source-e5-judge-step725-original-score
ORIGINAL_SCORE_CACHE_PAYLOAD_SCHEMA=vf_original_score_cache_e5_judge_v1
ORIGINAL_SCORE_CACHE_EXPECTED_RATING_MIN=0.0
ORIGINAL_SCORE_CACHE_EXPECTED_RATING_MAX=5.0
FLASH_ATTN_WHEEL=/path/to/validated_flash_attn.whl
FLASH_ATTN_WHEEL_SHA256=
WANDB_MODE=offline
WANDB_PROJECT=mr-iqa-2
WANDB_ENTITY=
VALIDATION_LOG_WANDB=0
VF_MIN_FREE_GIB=500
|