triplaneturbo_executable/utils/general_utils.py

import torch
import torch.nn as nn
import torch.nn.functional as F
from torch import Tensor
import numpy as np
from dataclasses import asdict, is_dataclass
import gc

def config_to_primitive(config):
    """Convert a dataclass config to a dictionary recursively."""
    if is_dataclass(config):
        config_dict = asdict(config)
        return {k: config_to_primitive(v) for k, v in config_dict.items()}
    elif isinstance(config, dict):
        return {k: config_to_primitive(v) for k, v in config.items()}
    elif isinstance(config, list):
        return [config_to_primitive(v) for v in config]
    elif isinstance(config, tuple):
        return tuple(config_to_primitive(v) for v in config)
    else:
        return config

def scale_tensor(
    dat, 
    inp_scale, 
    tgt_scale
):
    if inp_scale is None:
        inp_scale = (0, 1)
    if tgt_scale is None:
        tgt_scale = (0, 1)
    if isinstance(tgt_scale, Tensor):
        assert dat.shape[-1] == tgt_scale.shape[-1]
    dat = (dat - inp_scale[0]) / (inp_scale[1] - inp_scale[0])
    dat = dat * (tgt_scale[1] - tgt_scale[0]) + tgt_scale[0]
    return dat

def contract_to_unisphere_custom(
    x, bbox, unbounded = False
):
    if unbounded:
        x = scale_tensor(x, bbox, (-1, 1))
        x = x * 2 - 1  # aabb is at [-1, 1]
        mag = x.norm(dim=-1, keepdim=True)
        mask = mag.squeeze(-1) > 1
        x[mask] = (2 - 1 / mag[mask]) * (x[mask] / mag[mask])
        x = x / 4 + 0.5  # [-inf, inf] is at [0, 1]
    else:
        x = scale_tensor(x, bbox, (-1, 1))
    return x

# bug fix in https://github.com/NVlabs/eg3d/issues/67
planes =  torch.tensor(
            [
                [
                    [1, 0, 0],
                    [0, 1, 0],
                    [0, 0, 1]
                ],
                [
                    [1, 0, 0],
                    [0, 0, 1],
                    [0, 1, 0]
                ],
                [
                    [0, 0, 1],
                    [0, 1, 0],
                    [1, 0, 0]
                ]
            ], dtype=torch.float32)


def grid_sample(input, grid):
    # if grid.requires_grad and _should_use_custom_op():
    #     return grid_sample_2d(input, grid, padding_mode='zeros', align_corners=False)
    return torch.nn.functional.grid_sample(input=input, grid=grid, mode='bilinear', padding_mode='zeros', align_corners=False)


def project_onto_planes(planes, coordinates):
    """
    Does a projection of a 3D point onto a batch of 2D planes,
    returning 2D plane coordinates.

    Takes plane axes of shape n_planes, 3, 3
    # Takes coordinates of shape N, M, 3
    # returns projections of shape N*n_planes, M, 2
    """
    N, M, C = coordinates.shape
    n_planes, _, _ = planes.shape
    coordinates = coordinates.unsqueeze(1).expand(-1, n_planes, -1, -1).reshape(N*n_planes, M, 3)
    inv_planes = torch.linalg.inv(planes).unsqueeze(0).expand(N, -1, -1, -1).reshape(N*n_planes, 3, 3)
    projections = torch.bmm(coordinates, inv_planes)
    return projections[..., :2]

def sample_from_planes(plane_features, coordinates, mode='bilinear', padding_mode='zeros', box_warp=2, interpolate_feat = None):
    assert padding_mode == 'zeros'
    N, n_planes, C, H, W = plane_features.shape
    _, M, _ = coordinates.shape
    plane_features = plane_features.view(N*n_planes, C, H, W)

    coordinates = (2/box_warp) * coordinates # add specific box bounds

    if interpolate_feat in [None, "v1"]:
        projected_coordinates = project_onto_planes(planes.to(coordinates), coordinates).unsqueeze(1)
        output_features = grid_sample(plane_features, projected_coordinates.float())
        output_features = output_features.permute(0, 3, 2, 1).reshape(N, n_planes, M, C)
        output_features = output_features.sum(dim=1, keepdim=True).reshape(N, M, C)

    elif interpolate_feat in ["v2"]:
        projected_coordinates = project_onto_planes(planes.to(coordinates), coordinates).unsqueeze(1)
        output_features = grid_sample(plane_features, projected_coordinates.float())
        output_features = output_features.permute(0, 3, 2, 1).reshape(N, n_planes, M, C)
        output_features = output_features.permute(0, 2, 1, 3).reshape(N, M, n_planes*C)        

    return output_features.contiguous()

def cleanup():
    """Cleanup torch memory."""
    gc.collect()
    torch.cuda.empty_cache()
    torch.cuda.ipc_collect()