dataloader logic, new folder structure, flatter configs (#26)

* renaming dataloader to data. adding some dataloader classes

* restructuring configs

* fixed instant ngp tcnn direction rescaling

* working through test cases

* black and lint and tests
This commit is contained in:
Ethan Weber authored and GitHub committed 2022-06-08 09:01:56 -07:00
1 parent 24c1c77ef1
commit 457788d689
105 files changed
+8828 -682

No files matched your search

+1 -1
View File
@@ -12,7 +12,7 @@ from hydra.core.global_hydra import GlobalHydra
from omegaconf import DictConfig
from tqdm import tqdm
from pyrad.nerf.trainer import Trainer
from pyrad.engine.trainer import Trainer
BENCH = {
"method": "vanilla_nerf",
+1 -1
View File
@@ -7,7 +7,7 @@ Script to process different dataset types into our format.
from hydra import compose, initialize
import json
from pyrad.nerf.dataset.utils import get_dataset_inputs
from pyrad.data.utils import get_dataset_inputs
from pyrad.utils.io import get_absolute_path, make_dir, write_to_pkl
from pyrad.utils.misc import get_hash_str_from_dict
+88
View File
@@ -0,0 +1,88 @@
"""
run_eval.py
"""
import json
import logging
import os
from datetime import date
import torch
from hydra import compose, initialize
from hydra.core.global_hydra import GlobalHydra
from omegaconf import DictConfig
from tqdm import tqdm
from pyrad.engine.trainer import Trainer
from pyrad.cameras.cameras import Camera
BENCH = {
"method": "vanilla_nerf",
"hydra_base_dir": "outputs/",
"benchmark_date": "05-26-2022",
"object_list": ["mic", "ficus", "chair", "hotdog", "materials", "drums", "ship", "lego"],
}
def load_best_ckpt(hydra_dir: str, config: DictConfig) -> str:
"""helper function to update config with latest checkpoint in specified hydra_dir"""
model_dir = os.path.join(hydra_dir, config.graph.model_dir)
latest_ckpt = os.listdir(model_dir)
latest_ckpt.sort()
latest_ckpt = latest_ckpt[-1]
step = os.path.splitext(latest_ckpt)[0].split("-")[-1]
config.graph.resume_train.load_dir = model_dir
config.graph.resume_train.load_step = int(step)
config.logging.writer.TensorboardWriter.save_dir = "/tmp/"
return os.path.join(model_dir, latest_ckpt)
def main():
"""Main function."""
# benchmarks = {}
# hydra_base_dir = BENCH["hydra_base_dir"]
# method = BENCH["method"]
# benchmark_date = BENCH["benchmark_date"]
# for dataset in tqdm(BENCH["object_list"]):
# hydra_dir = f"{hydra_base_dir}/blender_{dataset}_{benchmark_date}/{method}/"
# basename = os.listdir(hydra_dir)[0]
# hydra_dir = f"{hydra_dir}/{basename}"
# initialize(config_path=os.path.join("../../", hydra_dir, ".hydra/"))
# config = compose("configs/flat.yaml")
with initialize(version_base=None, config_path="../configs"):
config = compose(config_name="flat.yaml")
trainer = Trainer(config, local_rank=0, world_size=1)
trainer.setup(test_mode=False)
# create a camera trajectory
print("Running eval.")
# camera = Camera()
# trainer.graph.eval()
# camera_ray_bundle = camera.get_camera_ray_bundle(device=trainer.device)
image_idx = 0
camera_ray_bundle, batch = trainer.dataloader_eval.get_data_from_image_idx(image_idx)
image_height, image_width = camera_ray_bundle.origins.shape[:2]
print(image_height, image_width)
outputs = trainer.graph.get_outputs_for_camera_ray_bundle(camera_ray_bundle)
# print(outputs)
# avg_psnr = 0
# for step, image_idx in enumerate(config.data.val_image_indices):
# with torch.no_grad():
# psnr = trainer.test_image(image_idx=image_idx, step=step)
# avg_psnr = (step * avg_psnr + psnr) / (step + 1)
# benchmarks[dataset] = (avg_psnr, ckpt)
# GlobalHydra.instance().clear()
# benchmark_info = {"bench": BENCH, "results": benchmarks}
# timestamp = date.today().strftime("%b-%d-%Y")
# json_file = os.path.join(BENCH["hydra_base_dir"], f"{timestamp}.json")
# with open(json_file, "w", encoding="utf8") as f:
# json.dump(benchmark_info, f, indent=2)
# logging.info("saved benchmark results to %s", json_file)
if __name__ == "__main__":
main()
+4 -4
View File
@@ -19,7 +19,7 @@ import torch.multiprocessing as mp
from omegaconf import DictConfig
import yaml
from pyrad.nerf.trainer import Trainer
from pyrad.engine.trainer import Trainer
from pyrad.utils import comms, profiler
logging.basicConfig(format="[%(filename)s:%(lineno)d] %(message)s", level=logging.DEBUG)
@@ -102,7 +102,7 @@ def _distributed_worker(
assert num_gpus_per_machine <= torch.cuda.device_count()
torch.cuda.set_device(local_rank)
_set_random_seed(config.seed + global_rank)
_set_random_seed(config.machine_config.seed + global_rank)
comms.synchronize(world_size)
output = main_func(local_rank, world_size, config)
@@ -151,11 +151,11 @@ def launch(
world_size = num_machines * num_gpus_per_machine
if world_size == 0:
# Using only CPU and one process.
_set_random_seed(config.seed)
_set_random_seed(config.machine_config.seed)
main_func(local_rank=0, world_size=0, config=config)
elif world_size == 1:
# Using one gpu and one process.
_set_random_seed(config.seed)
_set_random_seed(config.machine_config.seed)
try:
main_func(local_rank=0, world_size=1, config=config)
except KeyboardInterrupt: