MCPcopy Create free account
hub / github.com/MeiGen-AI/MultiTalk / _parse_args

Function _parse_args

generate_multitalk.py:60–252  ·  view source on GitHub ↗
()

Source from the content-addressed store, hash-verified

58
59
60def _parse_args():
61 parser = argparse.ArgumentParser(
62 description="Generate a image or video from a text prompt or image using Wan"
63 )
64 parser.add_argument(
65 "--task",
66 type=str,
67 default="multitalk-14B",
68 choices=list(WAN_CONFIGS.keys()),
69 help="The task to run.")
70 parser.add_argument(
71 "--size",
72 type=str,
73 default="multitalk-480",
74 choices=list(SIZE_CONFIGS.keys()),
75 help="The buckget size of the generated video. The aspect ratio of the output video will follow that of the input image."
76 )
77 parser.add_argument(
78 "--frame_num",
79 type=int,
80 default=81,
81 help="How many frames to be generated in one clip. The number should be 4n+1"
82 )
83 parser.add_argument(
84 "--ckpt_dir",
85 type=str,
86 default=None,
87 help="The path to the Wan checkpoint directory.")
88 parser.add_argument(
89 "--quant_dir",
90 type=str,
91 default=None,
92 help="The path to the Wan quant checkpoint directory.")
93 parser.add_argument(
94 "--wav2vec_dir",
95 type=str,
96 default=None,
97 help="The path to the wav2vec checkpoint directory.")
98 parser.add_argument(
99 "--lora_dir",
100 type=str,
101 nargs='+',
102 default=None,
103 help="The paths to the LoRA checkpoint files."
104 )
105 parser.add_argument(
106 "--lora_scale",
107 type=float,
108 nargs='+',
109 default=[1.2],
110 help="Controls how much to influence the outputs with the LoRA parameters. Accepts multiple float values."
111 )
112 parser.add_argument(
113 "--offload_model",
114 type=str2bool,
115 default=None,
116 help="Whether to offload the model to CPU after each model forward, reducing GPU memory usage."
117 )

Callers 1

Calls 1

_validate_argsFunction · 0.70

Tested by

no test coverage detected