{
"dataset_param": {
"dataset_type": "t2v",
"basic_parameters": {
"data_path": "./examples/hunyuanvideo/feature_extract/data.txt",
"data_storage_mode": "combine"
},
"preprocess_parameters": {
"video_processor_type": "OpensoraplanVideoProcessor",
"video_reader_type": "DecordVideo",
"image_reader_type": "Image",
"num_frames": 93,
"frame_interval": 1,
"max_height": 480,
"max_width": 848,
"max_hxw": 407040,
"train_fps": 24,
"speed_factor": 1.0,
"drop_short_ratio": 0.0,
"cfg": 0.0,
"hw_stride": 8,
"force_resolution": false,
"hw_aspect_thr": 2.0,
"train_pipeline": {
"video": [
{
"trans_type": "ToTensorVideo"
},
{
"trans_type": "CenterCropResizeVideo",
"param": {
"use_short_edge": false,
"top_crop": false,
"align_corners": false,
"antialias": false,
"transform_size": "auto"
}
},
{
"trans_type": "norm_fun",
"param": {
"mean": 0.5,
"std": 0.5
}
}
]
}
},
"use_text_processer": true,
"enable_text_preprocessing": false,
"tokenizer_config": [
{
"autotokenizer_name": "AutoTokenizer",
"hub_backend": "hf",
"from_pretrained": "llava-llama-3-8b-text-encoder-tokenizer",
"model_max_length": 256
},
{
"autotokenizer_name": "CLIPTokenizer",
"hub_backend": "hf",
"from_pretrained": "clip-vit-large-patch14",
"model_max_length": 77
}
]
},
"dataloader_param": {
"dataloader_mode": "sampler",
"sampler_type": "LengthGroupedSampler",
"shuffle": true,
"drop_last": true,
"pin_memory": true,
"group_frame": false,
"group_resolution": false,
"group_data": true,
"initial_global_step_for_sampler": 0
}
}