How to obtain clearer rendered images with higher resolution？

[xinding-sys]

• I am running the code on a headless machine. The complete code is as follows,

`
def main(random_selection=False, headless=False, short_exec=False):

og.log.info(f"Demo {__file__}\n    " + "*" * 80 + "\n    Description:\n" + main.__doc__ + "*" * 80)

# Ask the user whether they want online object sampling or not
sampling_options = {
    False: "Use a pre-sampled cached BEHAVIOR activity scene",
    True: "Sample the BEHAVIOR activity in an online fashion",
}
should_sample = choose_from_options(
    options=sampling_options, name="online object sampling", random_selection=random_selection
)

# Load the pre-selected configuration and set the online_sampling flag
config_filename = os.path.join(og.example_config_path, "r1pro_behavior.yaml")
cfg = yaml.load(open(config_filename, "r"), Loader=yaml.FullLoader)
cfg["task"]["online_object_sampling"] = should_sample
cfg["task"]["use_presampled_robot_pose"] = not should_sample

# Load the environment
env = og.Environment(configs=cfg)

# Move camera to a good position
og.sim.viewer_camera.set_position_orientation(
    position=[1.6, 6.15, 1.5], orientation=[-0.2322, 0.5895, 0.7199, -0.2835]
)

# Allow user to move camera more easily
og.sim.enable_viewer_camera_teleoperation()

# Run a simple loop and reset periodically
max_iterations = 10 if not short_exec else 1
for j in range(max_iterations):
    og.log.info("Resetting environment")
    env.reset()
    for i in range(100):
        action = env.robots[0].action_space.sample()
        import pdb;pdb.set_trace()
        state, reward, terminated, truncated, info = env.step(action * 0.1)
        saved_files = save_obs_rgb(state, j, out_dir="./obs_rgb")
        if terminated or truncated:
            og.log.info("Episode finished after {} timesteps".format(i + 1))
            break

# Always close the environment at the end
og.shutdown()`

• but after “env.step”, when saving the RGB images,

saved_files = save_obs_rgb(state, j, out_dir="./obs_rgb")

• I noticed that the image quality is very low:

• I noticed the following two warnings:

`
[Warning] [rtx.postprocessing.plugin] DLSS increasing input dimensions: Render resolution of (64, 64) is below minimal input resolution of 300.

[Warning] [rtx.neuraylib.plugin] [IRAY:RENDER] 1.1 IRAY rend warn : CUDA device 0 (NVIDIA A100-SXM4-80GB): ECC is enabled, this will reduce rendering performance (it is strongly recommended to disable ECC with iray)
`
Is the problem related to them? and how to get higher resolution images?

[wensi-ai]

Hi @xinding-sys , the noisy rendering is a known issue for A100 GPUs, @cgokmen probably can say more about this.

In terms of resolution, you can change it when loading the robot in the config, something like this:

robots: 
  - type: R1Pro
    sensor_config:
      VisionSensor:
        sensor_kwargs:
          image_height: 480
          image_width: 480

You can also do something like this:

robot = env.robots[0]
# Update robot sensors:
for sensor_name in robot.sensors:
      robot.sensors[sensor_name].image_height = 480
      robot.sensors[sensor_name].image_width = 480
# reload observation space
env.load_observation_space()

[xinding-sys]

Thanks for your reply, @cgokmen because i want to perform RL training via omnigibson simulator, is it possible to fix noisy rendering question with A100?