↓ 2 callersMethodextract_video_flow Extract optical flow for an entire video sequence. Args: video_frames: Video tensor [B, T, C, H, W] or [T, C, H, W]
videox_fun/utils/optical_flow_utils.py:123
↓ 2 callersFunctionget_random_shape_mask(
shape, n_pts_range=[3, 10], rad_range=[0.0, 1.0], edgy_range=[0.0, 0.1], n_keyframes_range=[2, 25],
videox_fun/data/dataset_image_video.py:131
↓ 2 callersFunctionget_random_shape_mask(
shape, n_pts_range=[3, 10], rad_range=[0.0, 1.0], edgy_range=[0.0, 0.1], n_keyframes_range=[2, 25],
videox_fun/data/dataset_image_video_warped.py:131
↓ 2 callersFunctionlog_validation(vae, text_encoder, tokenizer, transformer3d, args, accelerator, weight_dtype, global_step)
scripts/cogvideox_fun/train.py:160
↓ 2 callersFunctionlog_validation(vae, text_encoder, tokenizer, transformer3d, args, accelerator, weight_dtype, global_step)
scripts/cogvideox_fun/train_warped_noise.py:161
↓ 2 callersFunctionmatrix_to_quaternionConvert 4x4 transformation matrix to quaternion [w, x, y, z, tx, ty, tz].
data_generation/blender_utils.py:50
↓ 2 callersFunctionquaternion_to_matrix_4x4 Convert quaternion to 4x4 transformation matrix. Quaternion format: [w, x, y, z, tx, ty, tz] or [w, x, y, z]
data_generation/blender_utils.py:22
↓ 1 callersMethodcheck_inputs(
self,
prompt,
height,
width,
negative_prompt,
callback_on_st
videox_fun/pipeline/pipeline_cogvideox_fun.py:467