Mason-McGough

import functools

import torch
import psutil


def process_memory() -> float:
    """
    Estimate total memory usage of the current process.
    """

Mason-McGough

"""
Crop all images in a given directory to the same size
"""

import argparse
import os
import concurrent.futures
from glob import glob
from enum import Enum

Mason-McGough

from typing import Callable

import torch
from torch import nn
import torch.nn.functional as F


class TNet(nn.Module):
    """
    Transformation network module of PointNet

Mason-McGough

ModelNet10/sofa/test/sofa_0780.off,sofa
ModelNet10/bed/test/bed_0545.off,bed
ModelNet10/dresser/test/dresser_0264.off,dresser
ModelNet10/dresser/test/dresser_0239.off,dresser
ModelNet10/table/test/table_0420.off,table
ModelNet10/table/test/table_0425.off,table
ModelNet10/monitor/test/monitor_0535.off,monitor
ModelNet10/table/test/table_0428.off,table
ModelNet10/dresser/test/dresser_0211.off,dresser
ModelNet10/monitor/test/monitor_0553.off,monitor

Mason-McGough

ModelNet10/chair/train/chair_0381.off,chair
ModelNet10/bed/train/bed_0463.off,bed
ModelNet10/bathtub/train/bathtub_0031.off,bathtub
ModelNet10/monitor/train/monitor_0101.off,monitor
ModelNet10/sofa/train/sofa_0128.off,sofa
ModelNet10/dresser/train/dresser_0115.off,dresser
ModelNet10/sofa/train/sofa_0250.off,sofa
ModelNet10/dresser/train/dresser_0104.off,dresser
ModelNet10/sofa/train/sofa_0579.off,sofa
ModelNet10/bed/train/bed_0062.off,bed

Mason-McGough

"""
Remove a person from an image using a stable diffusion server

This short demo accompanies the Medium article "Stable Diffusion as an API: Make a Person-Removing Microservice".
The full article can be found here: https://towardsdatascience.com/stable-diffusion-as-an-api-5e381aec1f6

Example usage:
    python inpaint-person.py my-image.jpg -W 768 -H 768 -o my-output.png
"""

Mason-McGough

def sample_hierarchical(
  rays_o: torch.Tensor,
  rays_d: torch.Tensor,
  z_vals: torch.Tensor,
  weights: torch.Tensor,
  n_samples: int,
  perturb: bool = False
) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor]:
  r"""
  Apply hierarchical sampling to the rays.

Mason-McGough

def sample_stratified(
  rays_o: torch.Tensor,
  rays_d: torch.Tensor,
  near: float,
  far: float,
  n_samples: int,
  perturb: Optional[bool] = True,
  inverse_depth: bool = False
) -> Tuple[torch.Tensor, torch.Tensor]:
  r"""

Mason-McGough

class NeRF(nn.Module):
  r"""
  Neural radiance fields module.
  """
  def __init__(
    self,
    d_input: int = 3,
    n_layers: int = 8,
    d_filter: int = 256,
    skip: Tuple[int] = (4,),

Mason-McGough

def raw2outputs(
  raw: torch.Tensor,
  z_vals: torch.Tensor,
  rays_d: torch.Tensor,
  raw_noise_std: float = 0.0,
  white_bkgd: bool = False
) -> Tuple[torch.Tensor, torch.Tensor, torch.Tensor, torch.Tensor]:
  r"""
  Convert the raw NeRF output into RGB and other maps.
  """