FreeStyle_Bench / README.md
Blue2Giant's picture
Add benchmark README
11d1116 verified
|
Raw
History Blame Contribute Delete
3.28 kB
---
pretty_name: FreeStyle Bench
task_categories:
- image-to-image
- text-to-image
language:
- en
tags:
- benchmark
- style-transfer
- image-editing
- cref
- sref
---
# FreeStyle Bench
This repository contains benchmark inputs for evaluating FreeStyle reference-based image editing.
It includes two benchmark subsets:
- `sref_bench/`: a style-transfer benchmark.
- `cref_sref_bench/`: a content-and-style dual-reference benchmark.
Both subsets use the same basic layout: a content-reference folder, a style-reference folder, and a `prompts.json` file that links each pair to its text prompt.
## Repository Layout
```text
<repo-root>/
README.md
sref_bench/
cref/
... content reference images ...
sref/
... style reference images ...
prompts.json
cref_sref_bench/
cref/
... content reference images ...
sref/
... style reference images ...
prompts.json
```
## `sref_bench`
`sref_bench` is the style-transfer benchmark.
For each benchmark pair:
- `sref_bench/cref/` contains the content reference image.
- `sref_bench/sref/` contains the style reference image.
- `sref_bench/prompts.json` contains the text prompt or instruction for the pair.
The keys in `prompts.json` are pair basenames. A key can be used to match the corresponding content and style reference images across the two folders.
For example, if `prompts.json` contains a key:
```text
example_pair
```
then the corresponding files should be found by matching the same basename under:
```text
sref_bench/cref/example_pair.*
sref_bench/sref/example_pair.*
```
where the image extension may depend on the actual uploaded files.
## `cref_sref_bench`
`cref_sref_bench` is the content-and-style dual-reference benchmark.
It has the same structure as `sref_bench`:
- `cref_sref_bench/cref/` contains the content reference image.
- `cref_sref_bench/sref/` contains the style reference image.
- `cref_sref_bench/prompts.json` contains the text prompt or instruction for the pair.
Again, each key in `prompts.json` is a pair basename and can be used to align the content reference image in `cref/` with the style reference image in `sref/`.
## How To Read A Pair
```python
import json
from pathlib import Path
from PIL import Image
bench_dir = Path("/path/to/FreeStyle_Bench/sref_bench") # or cref_sref_bench
with open(bench_dir / "prompts.json", "r", encoding="utf-8") as f:
prompts = json.load(f)
pair_name, prompt = next(iter(prompts.items()))
# Find images by matching the pair basename. The actual extension may vary.
cref_path = next((bench_dir / "cref").glob(pair_name + ".*"))
sref_path = next((bench_dir / "sref").glob(pair_name + ".*"))
content_ref = Image.open(cref_path).convert("RGB")
style_ref = Image.open(sref_path).convert("RGB")
print("pair:", pair_name)
print("prompt:", prompt)
print("content reference:", cref_path)
print("style reference:", sref_path)
```
## Notes
- `cref` means content reference.
- `sref` means style reference.
- `prompts.json` is the index file for each subset.
- The key of each `prompts.json` entry is the pair basename used to match images in `cref/` and `sref/`.
- Use `sref_bench` for style-transfer evaluation.
- Use `cref_sref_bench` for dual-reference content-and-style evaluation.