Dev gzf exp (#1624)

zhifu gao

2024-04-17 eaf9dda9e4d970af3d09db695e9e10c83ef94e25

 setup.py

@@ -14,16 +14,14 @@
        "librosa",
        "jamo",  # For kss
        "PyYAML>=5.1.2",
        # "soundfile>=0.12.1",
        "soundfile>=0.12.1",
        "kaldiio>=2.17.0",
        "torch_complex",
        # "nltk>=3.4.5",
        # ASR
        "sentencepiece", # train
        "jieba",
        # "rotary_embedding_torch",
        "rotary_embedding_torch",
        # "ffmpeg-python",
        # TTS
        # "pypinyin>=0.44.0",
        # "espnet_tts_frontend",
        # ENH
@@ -37,16 +35,16 @@
        # "textgrid",
        # "protobuf",
        "tqdm",
        "hdbscan",
        "umap",
        "umap_learn",
        "jaconv",
        "hydra-core>=1.3.2",
        "tensorboardX",
        "rotary_embedding_torch",
        "openai-whisper",
    ],
    # train: The modules invoked when training only.
    "train": [
        "editdistance",
        "wandb",
        "pip install tensorboardX",
    ],
    # all: The modules should be optionally installled due to some reason.
    #      Please consider moving them to "install" occasionally
@@ -55,6 +53,7 @@
        "torch_optimizer",
        "fairscale",
        "transformers",
        "openai-whisper"
    ],
    "setup": [
        "numpy",
@@ -84,8 +83,20 @@
        "sphinx-markdown-tables>=0.0.12",
        "configargparse>=1.2.1"
    ],
    "llm":[
        "transformers>=4.32.0",
        "accelerate",
        "tiktoken",
        "einops",
        "transformers_stream_generator>=0.0.4",
        "scipy",
        "torchvision",
        "pillow",
        "matplotlib",
    ],
}
requirements["all"].extend(requirements["train"])
requirements["all"].extend(requirements["llm"])
requirements["test"].extend(requirements["train"])

install_requires = requirements["install"]
@@ -130,5 +141,11 @@
    ],
    entry_points={"console_scripts": [
        "funasr = funasr.bin.inference:main_hydra",
        "funasr-train = funasr.bin.train:main_hydra",
        "funasr-export = funasr.bin.export:main_hydra",
        "scp2jsonl = funasr.datasets.audio_datasets.scp2jsonl:main_hydra",
        "jsonl2scp = funasr.datasets.audio_datasets.jsonl2scp:main_hydra",
        "funasr-scp2jsonl = funasr.datasets.audio_datasets.scp2jsonl:main_hydra",
        "funasr-jsonl2scp = funasr.datasets.audio_datasets.jsonl2scp:main_hydra",
    ]},
)