From 9e992b9550ff6ead33519973c4123303be626c7f Mon Sep 17 00:00:00 2001
From: Ethan Weber <eweb0124@gmail.com>
Date: Mon, 30 Jan 2023 19:49:17 +0000
Subject: [PATCH] fixing linting

---
 nerfstudio/cameras/camera_optimizers.py       |   2 +-
 .../data/datamanagers/base_datamanager.py     |   1 -
 .../data/dataparsers/mipnerf360_dataparser.py |  15 +-
 nerfstudio/data/pixel_samplers.py             |  64 ++++++--
 nerfstudio/engine/optimizers.py               |   3 +-
 nerfstudio/models/mipnerf.py                  |   2 +-
 nerfstudio/models/nerfacto.py                 |   2 +-
 nerfstudio/models/semantic_nerfw.py           |   2 +-
 nerfstudio/models/tensorf.py                  |   2 +-
 nerfstudio/models/vanilla_nerf.py             |   2 +-
 nerfstudio/pipelines/base_pipeline.py         |  30 ++--
 .../benchmark_nerfstudio_paper.py             | 149 ------------------
 scripts/downloads/download_data.py            |  13 +-
 scripts/render.py                             |   7 +-
 14 files changed, 84 insertions(+), 210 deletions(-)
 delete mode 100644 scripts/benchmarking/benchmark_nerfstudio_paper.py

diff --git a/nerfstudio/cameras/camera_optimizers.py b/nerfstudio/cameras/camera_optimizers.py
index 1d49a02df..7d0e045bc 100644
--- a/nerfstudio/cameras/camera_optimizers.py
+++ b/nerfstudio/cameras/camera_optimizers.py
@@ -81,7 +81,7 @@ class CameraOptimizer(nn.Module):
         self.initialize_parameters()
 
     def initialize_parameters(self) -> None:
-        # Initialize learnable parameters.
+        """Initialize learnable parameters."""
         if self.config.mode == "off":
             pass
         elif self.config.mode in ("SO3xR3", "SE3"):
diff --git a/nerfstudio/data/datamanagers/base_datamanager.py b/nerfstudio/data/datamanagers/base_datamanager.py
index 7fcc6b44a..c4a6d35e3 100644
--- a/nerfstudio/data/datamanagers/base_datamanager.py
+++ b/nerfstudio/data/datamanagers/base_datamanager.py
@@ -398,7 +398,6 @@ class VanillaDataManager(DataManager):  # pylint: disable=abstract-method
             num_cameras=self.eval_dataset.cameras.size, device=self.device
         )
         self.eval_ray_generator = RayGenerator(self.eval_dataset.cameras.to(self.device), self.eval_camera_optimizer)
-        # TODO(ethan): remove use of self.eval_dataloader from the spiral trajectory code
 
     def next_train(self, step: int) -> Tuple[RayBundle, Dict]:
         """Returns the next batch of data from the train dataloader."""
diff --git a/nerfstudio/data/dataparsers/mipnerf360_dataparser.py b/nerfstudio/data/dataparsers/mipnerf360_dataparser.py
index 5b4634dce..123ade13e 100644
--- a/nerfstudio/data/dataparsers/mipnerf360_dataparser.py
+++ b/nerfstudio/data/dataparsers/mipnerf360_dataparser.py
@@ -15,7 +15,6 @@
 
 from __future__ import annotations
 
-import math
 from dataclasses import dataclass, field
 from pathlib import Path, PurePath
 from typing import Optional, Type
@@ -60,15 +59,15 @@ class Mipnerf360DataParserConfig(DataParserConfig):
     """Whether to center the poses."""
     auto_scale_poses: bool = True
     """Whether to automatically scale the poses to fit in +/- 1 bounding box."""
-    train_split_percentage: float = 0.9
-    """The percent of images to use for training. The remaining images are for eval."""
+    eval_interval: int = 8
+    """The interval which eval images will be sampled."""
 
 
 @dataclass
 class Mipnerf360(DataParser):
     """Mipnerf360 DatasetParser"""
 
-    config: NerfstudioDataParserConfig
+    config: Mipnerf360DataParserConfig
     downscale_factor: Optional[int] = None
 
     def _generate_dataparser_outputs(self, split="train"):
@@ -106,6 +105,7 @@ class Mipnerf360(DataParser):
         distort = []
 
         # sort the frames by fname
+        # they do this in mipnerf360 code
         fnames = []
         for frame in meta["frames"]:
             filepath = PurePath(frame["file_path"])
@@ -181,11 +181,10 @@ class Mipnerf360(DataParser):
         # filter image_filenames and poses based on train/eval split percentage
         num_images = len(image_filenames)
 
-        # mipnerf360 code...
-        # TODO: add parameters for this...
+        # this chunk of code is very similar to the mipnerf360 code
         all_indices = np.arange(num_images)
-        train_indices = all_indices[all_indices % 8 != 0]
-        eval_indices = all_indices[all_indices % 8 == 0]
+        train_indices = all_indices[all_indices % self.config.eval_interval != 0]
+        eval_indices = all_indices[all_indices % self.config.eval_interval == 0]
         i_train = train_indices
         i_eval = eval_indices
 
diff --git a/nerfstudio/data/pixel_samplers.py b/nerfstudio/data/pixel_samplers.py
index 29db22dd5..07b72c88e 100644
--- a/nerfstudio/data/pixel_samplers.py
+++ b/nerfstudio/data/pixel_samplers.py
@@ -70,7 +70,9 @@ class PixelSampler:  # pylint: disable=too-few-public-methods
 
         return indices
 
-    def collate_image_dataset_batch(self, batch: Dict, num_rays_per_batch: int, keep_full_image: bool = False):
+    def collate_image_dataset_batch(
+        self, batch: Dict, num_rays_per_batch: int, keep_full_image: bool = False, sample_all_pixels: bool = False
+    ):
         """
         Operates on a batch of images and samples pixels to use for generating rays.
         Returns a collated batch which is input to the Graph.
@@ -79,35 +81,52 @@ class PixelSampler:  # pylint: disable=too-few-public-methods
             batch: batch of images to sample from
             num_rays_per_batch: number of rays to sample per batch
             keep_full_image: whether or not to include a reference to the full image in returned batch
+            sample_all_pixels: if True, will sample all the pixels in the image
         """
 
         device = batch["image"].device
         num_images, image_height, image_width, _ = batch["image"].shape
 
-        if "mask" in batch:
-            indices = self.sample_method(
-                num_rays_per_batch, num_images, image_height, image_width, mask=batch["mask"], device=device
+        if sample_all_pixels:
+            image_indices = (torch.ones(num_images, device=device) * batch["image_idx"].to(device)).to(torch.int64)
+            collated_batch = {key: value for key, value in batch.items() if key != "image_idx" and value is not None}
+            collated_batch["indices"] = torch.stack(
+                torch.meshgrid(
+                    image_indices,
+                    torch.arange(image_height, device=device),
+                    torch.arange(image_width, device=device),
+                    indexing="ij",
+                ),
+                dim=-1,
             )
         else:
-            indices = self.sample_method(num_rays_per_batch, num_images, image_height, image_width, device=device)
+            # if the mask is in the batch, then only sample within the mask
+            if "mask" in batch:
+                indices = self.sample_method(
+                    num_rays_per_batch, num_images, image_height, image_width, mask=batch["mask"], device=device
+                )
+            else:
+                indices = self.sample_method(num_rays_per_batch, num_images, image_height, image_width, device=device)
 
-        c, y, x = (i.flatten() for i in torch.split(indices, 1, dim=-1))
-        collated_batch = {
-            key: value[c, y, x] for key, value in batch.items() if key != "image_idx" and value is not None
-        }
+            c, y, x = (i.flatten() for i in torch.split(indices, 1, dim=-1))
+            collated_batch = {
+                key: value[c, y, x] for key, value in batch.items() if key != "image_idx" and value is not None
+            }
 
-        assert collated_batch["image"].shape == (num_rays_per_batch, 3), collated_batch["image"].shape
+            assert collated_batch["image"].shape == (num_rays_per_batch, 3), collated_batch["image"].shape
 
-        # Needed to correct the random indices to their actual camera idx locations.
-        indices[:, 0] = batch["image_idx"][c]
-        collated_batch["indices"] = indices  # with the abs camera indices
+            # Needed to correct the random indices to their actual camera idx locations.
+            indices[:, 0] = batch["image_idx"][c]
+            collated_batch["indices"] = indices  # with the abs camera indices
 
         if keep_full_image:
             collated_batch["full_image"] = batch["image"]
 
         return collated_batch
 
-    def collate_image_dataset_batch_list(self, batch: Dict, num_rays_per_batch: int, keep_full_image: bool = False):
+    def collate_image_dataset_batch_list(
+        self, batch: Dict, num_rays_per_batch: int, keep_full_image: bool = False, sample_all_pixels: bool = False
+    ):
         """
         Does the same as collate_image_dataset_batch, except it will operate over a list of images / masks inside
         a list.
@@ -118,6 +137,7 @@ class PixelSampler:  # pylint: disable=too-few-public-methods
             batch: batch of images to sample from
             num_rays_per_batch: number of rays to sample per batch
             keep_full_image: whether or not to include a reference to the full image in returned batch
+            sample_all_pixels: if True, will sample all the pixels in the image
         """
 
         device = batch["image"][0].device
@@ -127,6 +147,9 @@ class PixelSampler:  # pylint: disable=too-few-public-methods
         all_indices = []
         all_images = []
 
+        if sample_all_pixels:
+            raise NotImplementedError("sample_all_pixels is not implemented for collate_image_dataset_batch_list")
+
         if "mask" in batch:
             num_rays_in_batch = num_rays_per_batch // num_images
             for i in range(num_images):
@@ -176,19 +199,26 @@ class PixelSampler:  # pylint: disable=too-few-public-methods
 
         return collated_batch
 
-    def sample(self, image_batch: Dict):
+    def sample(self, image_batch: Dict, sample_all_pixels: bool = False):
         """Sample an image batch and return a pixel batch.
         Args:
             image_batch: batch of images to sample from
+            sample_all_pixels: if True, will sample all pixels in the image
         """
         if isinstance(image_batch["image"], list):
             image_batch = dict(image_batch.items())  # copy the dictionary so we don't modify the original
             pixel_batch = self.collate_image_dataset_batch_list(
-                image_batch, self.num_rays_per_batch, keep_full_image=self.keep_full_image
+                image_batch,
+                self.num_rays_per_batch,
+                keep_full_image=self.keep_full_image,
+                sample_all_pixels=sample_all_pixels,
             )
         elif isinstance(image_batch["image"], torch.Tensor):
             pixel_batch = self.collate_image_dataset_batch(
-                image_batch, self.num_rays_per_batch, keep_full_image=self.keep_full_image
+                image_batch,
+                self.num_rays_per_batch,
+                keep_full_image=self.keep_full_image,
+                sample_all_pixels=sample_all_pixels,
             )
         else:
             raise ValueError("image_batch['image'] must be a list or torch.Tensor")
diff --git a/nerfstudio/engine/optimizers.py b/nerfstudio/engine/optimizers.py
index c27ca4751..1622206c6 100644
--- a/nerfstudio/engine/optimizers.py
+++ b/nerfstudio/engine/optimizers.py
@@ -79,7 +79,8 @@ class Optimizers:
         for param_group_name, params in param_groups.items():
             if param_group_name not in config:
                 CONSOLE.print(
-                    f"[bold yellow]Warning: Skipping param group '{param_group_name}' because optimizer and scheduler are not in the config."
+                    f"[bold yellow]Warning: Skipping param group '{param_group_name}'",
+                    "[bold yellow]because optimizer and scheduler aren't in config.",
                 )
                 continue
             lr_init = config[param_group_name]["optimizer"].lr
diff --git a/nerfstudio/models/mipnerf.py b/nerfstudio/models/mipnerf.py
index b8c3bcc4f..cbc5139f2 100644
--- a/nerfstudio/models/mipnerf.py
+++ b/nerfstudio/models/mipnerf.py
@@ -86,7 +86,7 @@ class MipNerfModel(Model):
         # metrics
         self.psnr = PeakSignalNoiseRatio(data_range=1.0)
         self.ssim = structural_similarity_index_measure
-        self.lpips = LearnedPerceptualImagePatchSimilarity()
+        self.lpips = LearnedPerceptualImagePatchSimilarity(net_type="vgg", normalize=True)
 
     def get_param_groups(self) -> Dict[str, List[Parameter]]:
         param_groups = {}
diff --git a/nerfstudio/models/nerfacto.py b/nerfstudio/models/nerfacto.py
index 5e5ff03e9..04fff130d 100644
--- a/nerfstudio/models/nerfacto.py
+++ b/nerfstudio/models/nerfacto.py
@@ -55,7 +55,7 @@ from nerfstudio.model_components.renderers import (
 )
 from nerfstudio.model_components.scene_colliders import AABBBoxCollider, NearFarCollider
 from nerfstudio.models.base_model import Model, ModelConfig
-from nerfstudio.utils import colormaps, colors
+from nerfstudio.utils import colormaps
 
 
 @dataclass
diff --git a/nerfstudio/models/semantic_nerfw.py b/nerfstudio/models/semantic_nerfw.py
index 941716dc8..1361ff1af 100644
--- a/nerfstudio/models/semantic_nerfw.py
+++ b/nerfstudio/models/semantic_nerfw.py
@@ -137,7 +137,7 @@ class SemanticNerfWModel(Model):
         # metrics
         self.psnr = PeakSignalNoiseRatio(data_range=1.0)
         self.ssim = structural_similarity_index_measure
-        self.lpips = LearnedPerceptualImagePatchSimilarity()
+        self.lpips = LearnedPerceptualImagePatchSimilarity(net_type="vgg", normalize=True)
 
     def get_param_groups(self) -> Dict[str, List[Parameter]]:
         param_groups = {}
diff --git a/nerfstudio/models/tensorf.py b/nerfstudio/models/tensorf.py
index 60fec9706..9e0c28289 100644
--- a/nerfstudio/models/tensorf.py
+++ b/nerfstudio/models/tensorf.py
@@ -201,7 +201,7 @@ class TensoRFModel(Model):
         # metrics
         self.psnr = PeakSignalNoiseRatio(data_range=1.0)
         self.ssim = structural_similarity_index_measure
-        self.lpips = LearnedPerceptualImagePatchSimilarity()
+        self.lpips = LearnedPerceptualImagePatchSimilarity(net_type="vgg", normalize=True)
 
         # colliders
         if self.config.enable_collider:
diff --git a/nerfstudio/models/vanilla_nerf.py b/nerfstudio/models/vanilla_nerf.py
index 3e731524f..2970c106a 100644
--- a/nerfstudio/models/vanilla_nerf.py
+++ b/nerfstudio/models/vanilla_nerf.py
@@ -118,7 +118,7 @@ class NeRFModel(Model):
         # metrics
         self.psnr = PeakSignalNoiseRatio(data_range=1.0)
         self.ssim = structural_similarity_index_measure
-        self.lpips = LearnedPerceptualImagePatchSimilarity()
+        self.lpips = LearnedPerceptualImagePatchSimilarity(net_type="vgg", normalize=True)
 
         if getattr(self.config, "enable_temporal_distortion", False):
             params = self.config.temporal_distortion_params
diff --git a/nerfstudio/pipelines/base_pipeline.py b/nerfstudio/pipelines/base_pipeline.py
index 76e57cbb4..01ba182a5 100644
--- a/nerfstudio/pipelines/base_pipeline.py
+++ b/nerfstudio/pipelines/base_pipeline.py
@@ -33,12 +33,10 @@ from rich.progress import (
     Progress,
     TextColumn,
     TimeElapsedColumn,
-    track,
 )
 from torch import nn
 from torch.nn import Parameter
 from torch.nn.parallel import DistributedDataParallel as DDP
-from tqdm import tqdm
 from typing_extensions import Literal
 
 from nerfstudio.configs import base_config as cfg
@@ -331,8 +329,8 @@ class VanillaPipeline(Pipeline):
         self.train()
         return metrics_dict, images_dict
 
-    def get_eval_optimize_camera_camera_ray_bundle(self, idx):
-        # start optimizing camera pose
+    def eval_optimize_camera(self, idx):
+        """Optimize the camera pose."""
         # TODO: raise error if the method is not Nerfacto
 
         # get the eval camera optimizer's parameters
@@ -362,11 +360,11 @@ class VanillaPipeline(Pipeline):
         self.eval()
         # done optimizing camera pose
 
-        return camera_ray_bundle
-
-    def get_eval_optimize_appearance_camera_ray_bundle(self, idx):
-        # start optimizing appearance embedding
+    def get_eval_optimize_appearance(self, idx):
+        """Optimize the appearance embedding."""
         # TODO: raise error if the method is not Nerfacto
+
+        # start optimizing appearance embedding
         self.datamanager.eval_ray_generator.cameras.rescale_output_resolution(self.config.eval_image_scale_factor)
         self.train()
 
@@ -376,7 +374,6 @@ class VanillaPipeline(Pipeline):
         optimizer = torch.optim.Adam(appearance_parameters, lr=1e-3, eps=1e-15)
 
         side = random.choice([0, 1])
-        half_width = width // 2
         rgb_loss = MSELoss()
 
         for _ in range(self.config.eval_num_appearance_iters):
@@ -387,6 +384,9 @@ class VanillaPipeline(Pipeline):
             )
             outputs = self.model.get_outputs_for_camera_ray_bundle(camera_ray_bundle)
 
+            _, width = camera_ray_bundle.shape
+            half_width = width // 2
+
             # compute loss on half the image
             image = batch["image"].to(self.device)
             if side == 0:
@@ -422,17 +422,17 @@ class VanillaPipeline(Pipeline):
         ) as progress:
             task = progress.add_task("[green]Evaluating all eval images...", total=num_images)
             for idx in range(num_images):
-                camera_ray_bundle, batch = self.datamanager.get_eval_image(idx)
                 # time this the following line
                 inner_start = time()
-                height, width = camera_ray_bundle.shape
-                num_rays = height * width
 
                 if self.config.eval_optimize_cameras:
-                    camera_ray_bundle = self.get_eval_optimize_camera_camera_ray_bundle(idx)
+                    self.eval_optimize_camera(idx)
                 if self.config.eval_optimize_appearance:
-                    camera_ray_bundle = self.get_eval_optimize_appearance_camera_ray_bundle(idx)
-                # at this point, camera_ray_bundle might be optimized for pose and/or appearance embedding
+                    self.get_eval_optimize_appearance(idx)
+
+                camera_ray_bundle, batch = self.datamanager.get_eval_image(idx)
+                height, width = camera_ray_bundle.shape
+                num_rays = height * width
 
                 with torch.no_grad():
                     outputs = self.model.get_outputs_for_camera_ray_bundle(camera_ray_bundle)
diff --git a/scripts/benchmarking/benchmark_nerfstudio_paper.py b/scripts/benchmarking/benchmark_nerfstudio_paper.py
deleted file mode 100644
index cd3cd730c..000000000
--- a/scripts/benchmarking/benchmark_nerfstudio_paper.py
+++ /dev/null
@@ -1,149 +0,0 @@
-"""
-Benchmarking script for nerfstudio paper.
-
-- nerfacto and instant-ngp methods on mipnerf360 data
-- nerfacto ablations
-"""
-
-import threading
-import time
-from pathlib import Path
-
-import GPUtil
-
-from nerfstudio.utils.scripts import run_command
-
-# for the mipnerf360 experiments
-mipnerf360_capture_names = ["bicycle", "garden", "stump", "room", "counter", "kitchen", "bonsai"]  # 7 splits
-mipnerf360_table_rows = [
-    # nerfacto method
-    (
-        "nerfacto-w/o-pose-app",
-        "nerfacto",
-        "--pipeline.eval_optimize_cameras False --pipeline.eval_optimize_appearance False --pipeline.datamanager.camera-optimizer.mode off --pipeline.model.use-appearance-embedding False nerfstudio-data --downscale-factor 4 --train-split-percentage 0.875 mipnerf360-data",
-    ),
-]
-
-# for the ablation experiments
-ablations_capture_names = [
-    "Egypt",
-    "person",
-    "kitchen",
-    "plane",
-    "dozer",
-    "floating-tree",
-    "aspen",
-    "stump",
-    "sculpture",
-    "Giannini-Hall",
-]
-
-ablations_table_rows = [
-    ("nerfacto", "nerfacto", "--pipeline.eval_optimize_cameras True --pipeline.eval_optimize_appearance True"),
-    (
-        "w/o-pose",
-        "nerfacto",
-        "--pipeline.eval_optimize_cameras True --pipeline.eval_optimize_appearance True --pipeline.datamanager.camera-optimizer.mode off",
-    ),
-    (
-        "w/o-app",
-        "nerfacto",
-        "--pipeline.eval_optimize_cameras True --pipeline.eval_optimize_appearance False --pipeline.model.use-appearance-embedding False",
-    ),
-    (
-        "w/o-pose-app",
-        "nerfacto",
-        "--pipeline.eval_optimize_cameras True --pipeline.eval_optimize_appearance False --pipeline.datamanager.camera-optimizer.mode off --pipeline.model.use-appearance-embedding False",
-    ),
-    (
-        "1-prop-network",
-        "nerfacto",
-        '--pipeline.eval_optimize_cameras True --pipeline.eval_optimize_appearance True --pipeline.model.num-proposal-samples-per-ray "256" --pipeline.model.num_proposal_iterations 1',
-    ),
-    (
-        "l2-contraction",
-        "nerfacto",
-        "--pipeline.eval_optimize_cameras True --pipeline.eval_optimize_appearance True --pipeline.model.scene-contraction-norm l2",
-    ),
-    (
-        "shared-prop-network",
-        "nerfacto",
-        "--pipeline.eval_optimize_cameras True --pipeline.eval_optimize_appearance True --pipeline.model.use-same-proposal-network True",
-    ),
-    (
-        "random-background-color",
-        "nerfacto",
-        "--pipeline.eval_optimize_cameras True --pipeline.eval_optimize_appearance True --pipeline.model.background-color random",
-    ),
-    (
-        "no-contraction",
-        "nerfacto",
-        "--pipeline.eval_optimize_cameras True --pipeline.eval_optimize_appearance True --pipeline.model.use-bounded True --pipeline.model.use-scene-contraction False nerfstudio-data --scale_factor 0.125",
-    ),
-    (
-        "synthetic-on-real",
-        "nerfacto",
-        "--pipeline.eval_optimize_cameras True --pipeline.eval_optimize_appearance False --pipeline.datamanager.camera-optimizer.mode off --pipeline.model.use-appearance-embedding False --pipeline.model.use-bounded True --pipeline.model.use-scene-contraction False nerfstudio-data --scale_factor 0.125",
-    ),
-]
-
-
-def main(capture_names, table_rows, data_path: Path = Path("data/nerfstudio")):
-    """Main method."""
-    # 30K iterations
-
-    # make a list of all the jobs that need to be fun
-    jobs = []
-    for capture_name in capture_names:
-
-        for table_row_name, method, table_row_command in table_rows:
-            command = " ".join(
-                (
-                    f"ns-train {method}",
-                    "--vis wandb",
-                    f"--data { data_path / capture_name}",
-                    "--output-dir outputs/nerfacto-ablations",
-                    "--trainer.steps-per-eval-batch 0 --trainer.steps-per-eval-image 0",
-                    "--trainer.steps-per-eval-all-images 5000 --trainer.max-num-iterations 30001",
-                    f"--wandb-name {capture_name}_{table_row_name}",
-                    f"--experiment-name {capture_name}_{table_row_name}",
-                    # extra_string,
-                    table_row_command,
-                )
-            )
-            jobs.append(command)
-
-    while jobs:
-        # get GPUs that capacity to run these jobs
-        gpu_devices_available = GPUtil.getAvailable(order="first", limit=10, maxMemory=0.1)
-
-        print("Available GPUs: ", gpu_devices_available)
-
-        # thread list
-        threads = []
-        while gpu_devices_available and jobs:
-            gpu = gpu_devices_available.pop(0)
-            command = f"CUDA_VISIBLE_DEVICES={gpu} " + jobs.pop(0)
-
-            def task():
-                print("Starting command: ", command)
-                # out = run_command(command, verbose=False)
-                # print("Finished command: ", command)
-
-            threads.append(threading.Thread(target=task))
-            threads[-1].start()
-
-            # NOTE(ethan): here we need a delay, otherwise the wandb/tensorboard naming is messed up... not sure why
-            time.sleep(5)
-
-        # wait for all threads to finish
-        for t in threads:
-            t.join()
-
-        print("Finished all threads")
-
-
-if __name__ == "__main__":
-    # pass
-    # main(mipnerf360_capture_names, mipnerf360_table_rows, data_path=Path("data/nerfstudio-data-mipnerf360"))
-    main(ablations_capture_names, ablations_table_rows)
diff --git a/scripts/downloads/download_data.py b/scripts/downloads/download_data.py
index aea0ddf59..0afdae14f 100755
--- a/scripts/downloads/download_data.py
+++ b/scripts/downloads/download_data.py
@@ -65,8 +65,6 @@ class FriendsDownload(DatasetDownload):
     """Download the friends dataset."""
 
     def download(self, save_dir: Path):
-        """Download the friends dataset."""
-
         # https://drive.google.com/file/d/1sgKr0ZO7BQC0FYinAnRSxobIWNucAST5/view?usp=sharing
         friends_file_id = "1sgKr0ZO7BQC0FYinAnRSxobIWNucAST5"
 
@@ -81,13 +79,10 @@ class FriendsDownload(DatasetDownload):
 
 @dataclass
 class Mipnerf360Download(DatasetDownload):
-    """Download the friends dataset."""
+    """Download the MIP-NeRF 360 dataset."""
 
     def download(self, save_dir: Path):
-        """Download the friends dataset."""
-
-        # https://drive.google.com/file/d/1sgKr0ZO7BQC0FYinAnRSxobIWNucAST5/view?usp=sharing
-        friends_file_id = "1sgKr0ZO7BQC0FYinAnRSxobIWNucAST5"
+        """Download the MIP-NeRF 360 dataset."""
 
         # Download the files
         url = "https://data.nerf.studio/nerfstudio-data-mipnerf360.zip"
@@ -275,13 +270,11 @@ PhototourismCaptureName = tyro.extras.literal_type_from_choices(phototourism_dow
 
 @dataclass
 class PhototourismDownload(DatasetDownload):
-    """Download the phototourism dataset."""
+    """Download the phototourism dataset: https://www.cs.ubc.ca/~kmyi/imw2020/data.html"""
 
     capture_name: PhototourismCaptureName = "brandenburg-gate"
 
     def download(self, save_dir: Path):
-        """Download a PhotoTourism dataset: https://www.cs.ubc.ca/~kmyi/imw2020/data.html"""
-
         if self.capture_name == "all":
             for capture_name in phototourism_downloads:
                 if capture_name != "all":
diff --git a/scripts/render.py b/scripts/render.py
index 4466c5385..78d8e5524 100644
--- a/scripts/render.py
+++ b/scripts/render.py
@@ -283,10 +283,11 @@ class RenderTrajectory:
         seconds = self.seconds
         crop_data = None
 
-        # TODO(ethan): use camera information from parsing args
         if self.traj == "spiral":
-            camera_start = pipeline.datamanager.eval_dataloader.get_camera(image_idx=0).flatten()
-            # TODO(ethan): pass in the up direction of the camera
+            # TODO(ethan): use camera information provided from arguments
+            # - camera index instead of defaulting to 0
+            # - up direction of the camera
+            camera_start = pipeline.datamanager.eval_dataset.cameras[0].flatten()
             camera_type = CameraType.PERSPECTIVE
             camera_path = get_spiral_path(camera_start, steps=30, radius=0.1)
         elif self.traj == "filename":