{
"dataset_param": {
"dataset_type": "t2v",
"basic_parameters": {
"data_path": "./examples/opensoraplan1.5/data.txt",
"data_storage_mode": "combine"
},
"preprocess_parameters": {
"video_processor_type": "OpensoraplanVideoProcessor",
"video_reader_type": "DecordVideo",
"image_reader_type": "Image",
"num_frames": 57,
"min_num_frames": 29,
"frame_interval": 1,
"max_height": 288,
"max_width": 512,
"max_hxw": 147456,
"min_hxw": 147456,
"train_fps": 12,
"speed_factor": 1.0,
"drop_short_ratio": 0.0,
"vae_scale_factor": [8,8,8],
"cfg": 0.1,
"hw_stride": 32,
"force_resolution": false,
"seed": 42,
"hw_aspect_thr": 2.0,
"use_aesthetic": true,
"train_pipeline": {
"video": [{
"trans_type": "ToTensorVideo"
},
{
"trans_type": "CenterCropResizeVideo",
"param": {
"transform_size": {
"max_height": 288,
"max_width": 512
},
"interpolation_mode": "bicubic"
}
},
{
"trans_type": "ae_norm"
}
],
"image": [{
"trans_type": "ToTensorVideo"
},
{
"trans_type": "CenterCropResizeVideo",
"param": {
"transform_size": {
"max_height": 288,
"max_width": 512
},
"interpolation_mode": "bicubic"
}
},
{
"trans_type": "ae_norm"
}
]
}
},
"use_text_processer": true,
"enable_text_preprocess": true,
"tokenizer_config":[{
"hub_backend": "hf",
"autotokenizer_name": "AutoTokenizer",
"from_pretrained": "t5/t5-v1_1-xl",
"low_cpu_mem_usage": false,
"model_max_length": 512
},
{
"hub_backend": "hf",
"autotokenizer_name": "AutoTokenizer",
"from_pretrained": "laion/CLIP-ViT-bigG-14-laion2B-39B-b160k",
"low_cpu_mem_usage": false,
"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,
"collate_param": {
"model_name": "open_sora_plan"
}
}
}