Installing dpg bench with latest python

[arijit-hub]

Hye,

I really thank the authors for their awesome repo. In case anyone is wandering how to install dpg bench with the latest python3.13; here is the dump of my .toml file. I used uv for my package/dependency management so if you are already using uv it might be useful for you to directly add it to your toml file. If you are not using uv you can just copy paste the package installations.

[project]
name = "ella"
version = "0.1.0"
description = "Add your description here"
readme = "README.md"
requires-python = ">=3.13"
dependencies = [
    "accelerate>=1.10.1",
    "addict>=2.4.0",
    "cloudpickle>=3.1.1",
    "datasets==2.21.0",
    "decord>=0.6.0",
    "diffusers>=0.35.1",
    "fairseq>=0.12.2",
    "ftfy>=6.0.3",
    "librosa==0.10.1",
    "modelscope>=1.30.0",
    "numpy>=2.3.3",
    "opencv-python>=4.11.0.86",
    "pandas>=2.3.2",
    "pillow>=11.3.0",
    "rapidfuzz>=3.14.1",
    "rouge-score<=0.0.4",
    "safetensors>=0.6.2",
    "simplejson>=3.20.1",
    "sortedcontainers>=2.4.0",
    "soundfile>=0.13.1",
    "taming-transformers-rom1504>=0.0.6",
    "tiktoken>=0.11.0",
    "timm>=1.0.19",
    "tokenizers>=0.22.1",
    "torchvision>=0.23.0",
    "tqdm>=4.67.1",
    "transformers>=4.56.2",
    "transformers-stream-generator>=0.0.5",
    "unicodedata2>=16.0.0",
    "zhconv>=1.4.3",
]

[tool.uv.sources]
fairseq = { git = "https://github.com/One-sixth/fairseq.git" }

Again I really appreciate the effort of the authors and the benchmark. Thanks a lot for your contribution.