body-pose-animation/train_pose.py
2021-02-17 13:18:15 +01:00

303 lines
8.5 KiB
Python

from modules.intersect import IntersectLoss
from modules.body_prior import BodyPrior
from modules.angle_sum import AngleSumLoss
from camera_estimation import TorchCameraEstimate
from modules.angle_clip import AngleClipper
from modules.angle_prior import AnglePriorsLoss
import smplx
import torch
from tqdm import tqdm
import torchgeometry as tgm
# internal imports
from modules.pose import BodyPose
from modules.filter import JointFilter
from model import VPoserModel
from modules.camera import SimpleCamera
from renderer import Renderer
def train_pose(
model: smplx.SMPL,
# current datapoints
keypoints,
keypoint_conf,
# 3D to 2D camera layer
camera: SimpleCamera,
# model type
model_type="smplx",
# pytorch config
device=torch.device('cuda'),
dtype=torch.float32,
# optimizer settings
optimizer=None,
optimizer_type="LBFGS",
learning_rate=1e-3,
iterations=60,
patience=10,
# configure loss function
# useBodyPrior=False,
# body_prior_weight=2,
# useAnglePrior=False,
# angle_prior_weight=0.5,
# use_angle_sum_loss=False,
# angle_sum_weight=0.1,
# body_mean_loss=False,
# body_mean_weight=0.01,
# useConfWeights=False,
# renderer options
renderer: Renderer = None,
render_steps=True,
# vposer=None,
extra_loss_layers=[],
use_progress_bar=True,
loss_analysis=True
):
if use_progress_bar:
print("[pose] starting training")
print("[pose] dtype=", dtype, device)
offscreen_step_output = []
loss_layer = torch.nn.MSELoss(reduction="sum").to(
device=device, dtype=dtype) # MSELoss()
# make sure camera module is on the correct device
camera = camera.to(device=device, dtype=dtype)
# setup keypoint data
keypoints = torch.tensor(keypoints).to(device=device, dtype=dtype)
keypoint_filter = JointFilter(
model_type=model_type, filter_dims=3).to(device=device, dtype=dtype)
# filter keypoints
keypoints = keypoint_filter(keypoints)
# get a list of openpose conf values
# keypoints_conf = torch.tensor(keypoint_conf).to(device=device, dtype=dtype)
# create filter layer to ignore unused joints, keypoints during optimization
filter_layer = JointFilter(
model_type=model_type, filter_dims=3).to(device=device, dtype=dtype)
# setup torch modules
pose_layer = BodyPose(model, dtype=dtype, device=device,
useBodyMeanAngles=False).to(device=device, dtype=dtype)
parameters = [pose_layer.body_pose]
# setup all loss layers
for l in extra_loss_layers:
# make sure layer is running on the correct device
l.to(device=device, dtype=dtype)
# register parameters if present
if l.has_parameters:
parameters = parameters + list(l.parameters())
if optimizer is None:
if optimizer_type.lower() == "lbfgs":
optimizer = torch.optim.LBFGS
elif optimizer_type.lower() == "adam":
optimizer = torch.optim.Adam
optimizer = optimizer(parameters, learning_rate)
if use_progress_bar:
pbar = tqdm(total=iterations)
def predict():
# return joints based on current model state
body_joints, cur_pose = pose_layer()
# compute homogeneous coordinates and project them to 2D space
points = tgm.convert_points_to_homogeneous(body_joints)
points = camera(points).squeeze()
points = filter_layer(points)
# compute loss between 2D joint projection and OpenPose keypoints
loss = loss_layer(points, keypoints) # * 100
# apply extra losses
for l in extra_loss_layers:
cur_loss = l(cur_pose, body_joints, points,
keypoints, pose_layer.cur_out)
if loss_analysis:
print(l.__class__.__name__, ":loss ->", cur_loss)
loss = loss + cur_loss
return loss
def optim_closure():
if torch.is_grad_enabled():
optimizer.zero_grad()
loss = predict()
if loss.requires_grad:
loss.backward()
return loss
# store results for optional plotting
cur_patience = patience
best_loss = None
best_output = None
loss_history = []
for t in range(iterations):
loss = optimizer.step(optim_closure)
# compute loss
cur_loss = loss.item()
loss_history.append(loss)
if best_loss is None:
best_loss = cur_loss
elif cur_loss < best_loss:
best_loss = cur_loss
best_output = pose_layer.cur_out
else:
cur_patience = cur_patience - 1
if patience == 0:
print("[train] aborted due to patience limit reached")
if use_progress_bar:
pbar.set_description("Error %f" % cur_loss)
pbar.update(1)
if renderer is not None and render_steps:
R = camera.trans.detach().cpu().numpy().squeeze()
renderer.render_model_with_tfs(
model, pose_layer.cur_out, keep_pose=True, transforms=R)
if renderer.use_offscreen:
offscreen_step_output.append(renderer.get_snapshot())
# renderer.set_group_pose("body", R)
if use_progress_bar:
pbar.close()
print("Final result:", loss.item())
return best_output, loss_history, offscreen_step_output
def get_loss_layers(config, model: smplx.SMPL, device, dtype):
""" Utility method to create loss layers based on a config file
Args:
config ([type]): [description]
device ([type]): [description]
dtype ([type]): [description]
"""
extra_loss_layers = []
if config['pose']['bodyPrior']['enabled']:
vmodel = VPoserModel.from_conf(config)
extra_loss_layers.append(BodyPrior(
device=device,
dtype=dtype,
vmodel=vmodel,
weight=config['pose']['bodyPrior']['weight']))
if config['pose']['anglePrior']['enabled']:
extra_loss_layers.append(AnglePriorsLoss(
device=device,
global_weight=config['pose']['anglePrior']['weight'],
dtype=dtype))
if config['pose']['angleSumLoss']['enabled']:
extra_loss_layers.append(AngleSumLoss(
device=device,
dtype=dtype,
weight=config['pose']['angleSumLoss']['weight']))
if config['pose']['angleLimitLoss']['enabled']:
extra_loss_layers.append(AngleClipper(
device=device,
dtype=dtype,
weight=config['pose']['angleLimitLoss']['weight']))
if config['pose']['intersectLoss']['enabled']:
extra_loss_layers.append(IntersectLoss(
model=model,
device=device,
dtype=dtype,
weight=config['pose']['intersectLoss']['weight'],
sigma=config['pose']['intersectLoss']['sigma'],
max_collisions=config['pose']['intersectLoss']['maxCollisions']
))
return extra_loss_layers
def train_pose_with_conf(
config,
camera: TorchCameraEstimate,
model: smplx.SMPL,
keypoints,
keypoint_conf,
device=torch.device('cpu'),
dtype=torch.float32,
renderer: Renderer = None,
render_steps=True,
use_progress_bar=True,
print_loss_layers=False
):
# configure PyTorch device and format
# dtype = torch.float64
if 'device' in config['pose'] is not None:
device = torch.device(config['pose']['device'])
else:
device = torch.device('cpu')
# create camera module
pose_camera, cam_trans, cam_int, cam_params = SimpleCamera.from_estimation_cam(
cam=camera,
use_intrinsics=config['pose']['useCameraIntrinsics'],
dtype=dtype,
device=device,
)
# apply transform to scene
if renderer is not None:
renderer.set_group_pose("body", cam_trans.cpu().numpy())
loss_layers = get_loss_layers(config, model, device, dtype)
if print_loss_layers:
print(loss_layers)
best_output, loss_history, offscreen_step_output = train_pose(
model=model.to(dtype=dtype),
keypoints=keypoints,
keypoint_conf=keypoint_conf,
camera=pose_camera,
device=device,
dtype=dtype,
renderer=renderer,
optimizer_type=config['pose']['optimizer'],
iterations=config['pose']['iterations'],
learning_rate=config['pose']['lr'],
render_steps=render_steps,
use_progress_bar=use_progress_bar,
extra_loss_layers=loss_layers
)
return best_output, cam_trans, loss_history, offscreen_step_output