facebookresearch / nocturne Goto Github PK

View Code? Open in Web Editor NEW

254.0 14.0 29.0 33.04 MB

A data-driven, fast driving simulator for multi-agent coordination under partial observability.

License: MIT License

Python 57.40% CMake 0.75% C++ 41.11% Jupyter Notebook 0.74%

nocturne's Issues

Dataset Build Instructions

Maybe let's continue this discussion (#11) here.

I ran into some issues during the dataset construction:

The files/scripts referenced in the README were moved to the scripts/ directory.
There are dependencies needed to run waymo_scenario_construction.py, which are not mentioned in environment.yml: tensorflow and waymo-open-dataset-tf-2-x-0.
Running make_solvable_files.py does not accept command line arguments (maybe because of hydra?)
Running make_solvable_files.py gives me this error: PermissionError: [Errno 13] Permission denied: '/checkpoint' . After some debugging I figured out that the paths in cfgs/config.yaml should be relative paths. But I first have to look more into hydra.
After fixing the paths, running make_solvable_files.py leads to the following error: omegaconf.errors.ConfigKeyError: Key 'start_time' is not in struct. I guess it should be cfg['scenario']['start_time].

[Question] Difference in active_masks copying in shared_buffer and separated_buffer files

Question

[Edit: I noticed an error in my post, but I modified my question appropriately]
I notice a difference in the way active_masks are being copied in the ppo buffers for insert() and chooseinsert():

In the chooseinsert() it is
self.active_masks[self.step] = active_masks.copy() line 184

While in the insert() it is
self.active_masks[self.step + 1] = active_masks.copy() line138

Thanks in advance!

[Bug] Problem installing it on Windows 10

Operating system

Windows 10

Bug description

I was trying to install it on windows 10 and the following error occurred. Can someone help? I ran the "python setup.py develop" command and the following error appeared on the screen

Steps to reproduce

No response

Relevant log output

D:\Users\User\anaconda3\envs\nocturne\lib\site-packages\setuptools\config\expand.py:142: UserWarning: File 'D:\\projects\\random_projects\\nocturne\\nocturne\\README.rst # todo' cannot be found
  warnings.warn(f"File {path!r} cannot be found")
D:\Users\User\anaconda3\envs\nocturne\lib\site-packages\setuptools\command\easy_install.py:144: EasyInstallDeprecationWarning: easy_install command is deprecated. Use build and pip and other standards-based tools.
  warnings.warn(
D:\Users\User\anaconda3\envs\nocturne\lib\site-packages\setuptools\command\install.py:34: SetuptoolsDeprecationWarning: setup.py install is deprecated. Use build and pip and other standards-based tools.
  warnings.warn(
setup.py:42: DeprecationWarning: distutils Version classes are deprecated. Use packaging.version instead.
  cmake_version = LooseVersion(
CMake Error at C:/Program Files/CMake/share/cmake-3.24/Modules/CMakeTestCXXCompiler.cmake:62 (message):
  The C++ compiler

    "C:/Program Files (x86)/Microsoft Visual Studio/2019/BuildTools/VC/Tools/MSVC/14.29.30133/bin/Hostx64/x64/cl.exe"

  is not able to compile a simple test program.

  It fails with the following output:

    Change Dir: D:/projects/random_projects/nocturne/nocturne/build/temp.win-amd64-cpython-38/Release/CMakeFiles/CMakeTmp

    Run Build Command(s):C:/Program Files (x86)/Microsoft Visual Studio/2019/BuildTools/MSBuild/Current/Bin/MSBuild.exe cmTC_a3a88.vcxproj /p:Configuration=Debug /p:Platform=x64 /p:VisualStudioVersion=16.0 /v:m && Microsoft (R) Build Engine version 16.11.2+f32259642 for .NET Framework
    Copyright (C) Microsoft Corporation. All rights reserved.

      Microsoft (R) C/C++ Optimizing Compiler Version 19.29.30146 for x64
      Copyright (C) Microsoft Corporation.  All rights reserved.
      cl /c /I"D:\gsl\vcpkg\installed\x64-windows\include" /Zi /Wall /WX- /diagnostics:column /Od /Ob0 /D _MBCS /D VERSION_INFO=0.0.1 /D WIN32 /D _WINDOWS /D "CMAKE_INTDIR=\"Debug\"" /Gm- /EHsc /RTC1 /MDd /GS /fp:precise /Zc:wchar_t /Zc:forScope /Zc:inline /GR /std:c++17 /Fo"cmTC_a3a88.dir\Debug\\" /Fd"cmTC_a3a88.dir\Debug\vc142.pdb" /external:W4 /Gd /TP /errorReport:queue  -O3 -Wextra -Wno-register -Wno-comment -fPIC -march=native -Wfatal-errors -fvisibility=hidden "D:\projects\random_projects\nocturne\nocturne\build\temp.win-amd64-cpython-38\Release\CMakeFiles\CMakeTmp\testCXXCompiler.cxx"
    cl : command line error D8021: invalid numeric argument '/Wextra' [D:\projects\random_projects\nocturne\nocturne\build\temp.win-amd64-cpython-38\Release\CMakeFiles\CMakeTmp\cmTC_a3a88.vcxproj]


  CMake will not be able to correctly generate this project.
Call Stack (most recent call first):
  CMakeLists.txt:12 (project)


Aborting due to errors when running command ['cmake', 'D:\\projects\\random_projects\\nocturne\\nocturne\\nocturne', '-DCMAKE_LIBRARY_OUTPUT_DIRECTORY=D:\\projects\\random_projects\\nocturne\\nocturne', '-DPYTHON_EXECUTABLE=D:\\Users\\User\\anaconda3\\envs\\nocturne\\python.exe', '-DCMAKE_BUILD_TYPE=Release']

Problems with mini dataset

Hi again,

It seems as though the mini-dataset folders are missing valid_files.json which throws an error when running create_env.py.
The full dataset has those files.

Thanks!

Will the other cars in the simulation react to the ego agent's actions?

Question

Hi,

Thank you for releasing the simulator. I'm trying to understand it and use it in my next paper. If I understood correctly, this simulator is based on a dataset, and other agents will not react to the ego agents decisions and movements, right? So this is kind of static environment? I think it's good to do other agent's motion prediction but not learning a policy for ego agent. Am I right?

Thank you

[Feature] Pretrained agents

Feature

Hi, thanks for your effort opensourcing this work!

I wonder if there is any plan on releasing the pretrained agents as reported in the paper?

Add Github Issues PR template

[Feature] Add support for Argoverse 2

Feature

Argoverse 2 also satisfies all the desiderata for what we want in a dataset so we should add support for it

[Bug] Mistake in continuous action space definition

Operating system

Ubuntu 20.04

Bug description

There seems to be a (small) mistake in the code here:

nocturne/nocturne/envs/base_env.py

Line 103 in 94fc821

high=np.array([

The high values should use the key 'x_upper_bound' instead of 'x_lower_bound'.

Steps to reproduce

No response

Relevant log output

No response

[Question] How can I get the VisibleObjects?

Question

Hi, i want to get the visible other objects from ones' perspective! It seems like VisibleObjects is that function in C++. Can I avail it in Python?
Thank u :)

[Question] Compatibility with VectorNet

Question

It mentions in the paper that the state vector follows conventions of the VectorNet.

I was wondering if there's any attempt/code of training a VectorNet-based policy with Nocturne? Thinking about the amount of engineering it takes to hook nocturne up with some latest trajectory prediction models.

Thanks,

Build Error: error: ld returned 1 exit status (libudev.so.0)

Running python setup.py develop failed and here's the error message:

/usr/bin/ld: warning: libudev.so.0, needed by /home/zdk/miniconda3/envs/nocturne/lib/libsfml-window.so.2.5.1, not found (try using -rpath or -rpath-link)
/home/zdk/miniconda3/envs/nocturne/lib/libsfml-window.so.2.5.1: undefined reference to `udev_device_get_devnode'
/home/zdk/miniconda3/envs/nocturne/lib/libsfml-window.so.2.5.1: undefined reference to `udev_monitor_enable_receiving'
/home/zdk/miniconda3/envs/nocturne/lib/libsfml-window.so.2.5.1: undefined reference to `udev_device_new_from_syspath'
/home/zdk/miniconda3/envs/nocturne/lib/libsfml-window.so.2.5.1: undefined reference to `udev_device_unref'
/home/zdk/miniconda3/envs/nocturne/lib/libsfml-window.so.2.5.1: undefined reference to `udev_monitor_new_from_netlink'
/home/zdk/miniconda3/envs/nocturne/lib/libsfml-window.so.2.5.1: undefined reference to `udev_enumerate_get_list_entry'
/home/zdk/miniconda3/envs/nocturne/lib/libsfml-window.so.2.5.1: undefined reference to `udev_device_get_syspath'
/home/zdk/miniconda3/envs/nocturne/lib/libsfml-window.so.2.5.1: undefined reference to `udev_monitor_unref'
/home/zdk/miniconda3/envs/nocturne/lib/libsfml-window.so.2.5.1: undefined reference to `udev_enumerate_add_match_subsystem'
/home/zdk/miniconda3/envs/nocturne/lib/libsfml-window.so.2.5.1: undefined reference to `udev_monitor_filter_add_match_subsystem_devtype'
/home/zdk/miniconda3/envs/nocturne/lib/libsfml-window.so.2.5.1: undefined reference to `udev_device_get_property_value'
/home/zdk/miniconda3/envs/nocturne/lib/libsfml-window.so.2.5.1: undefined reference to `udev_monitor_receive_device'
/home/zdk/miniconda3/envs/nocturne/lib/libsfml-window.so.2.5.1: undefined reference to `udev_enumerate_scan_devices'
/home/zdk/miniconda3/envs/nocturne/lib/libsfml-window.so.2.5.1: undefined reference to `udev_device_get_sysattr_value'
/home/zdk/miniconda3/envs/nocturne/lib/libsfml-window.so.2.5.1: undefined reference to `udev_monitor_get_fd'
/home/zdk/miniconda3/envs/nocturne/lib/libsfml-window.so.2.5.1: undefined reference to `udev_device_get_parent_with_subsystem_devtype'
/home/zdk/miniconda3/envs/nocturne/lib/libsfml-window.so.2.5.1: undefined reference to `udev_unref'
/home/zdk/miniconda3/envs/nocturne/lib/libsfml-window.so.2.5.1: undefined reference to `udev_device_get_action'
/home/zdk/miniconda3/envs/nocturne/lib/libsfml-window.so.2.5.1: undefined reference to `udev_list_entry_get_name'
/home/zdk/miniconda3/envs/nocturne/lib/libsfml-window.so.2.5.1: undefined reference to `udev_enumerate_new'
/home/zdk/miniconda3/envs/nocturne/lib/libsfml-window.so.2.5.1: undefined reference to `udev_list_entry_get_next'
/home/zdk/miniconda3/envs/nocturne/lib/libsfml-window.so.2.5.1: undefined reference to `udev_new'
/home/zdk/miniconda3/envs/nocturne/lib/libsfml-window.so.2.5.1: undefined reference to `udev_enumerate_unref'
collect2: error: ld returned 1 exit status
cpp/tests/CMakeFiles/nocturne_test.dir/build.make:234: recipe for target 'cpp/tests/nocturne_test' failed
make[2]: *** [cpp/tests/nocturne_test] Error 1
CMakeFiles/Makefile2:378: recipe for target 'cpp/tests/CMakeFiles/nocturne_test.dir/all' failed
make[1]: *** [cpp/tests/CMakeFiles/nocturne_test.dir/all] Error 2
make[1]: *** Waiting for unfinished jobs....
[100%] Built target nocturne_cpp
Makefile:135: recipe for target 'all' failed
make: *** [all] Error 2
Aborting due to errors when running command ['cmake', '--build', '.', '--config', 'Release', '--', '-j16']

I'm using cmake 3.24.0-rc1 and sfml 2.5.1.

[Bug] Unexpected behavior when setting padding=True in scenario.visible_state()

Operating system

Ubuntu 20.04

Bug description

Hi,

The description for the visible_state method in examples/nocturne_functions.py says that setting padding=True will add extra objects to the dict to hit the maximum number of objects for each type.

So for the default scenario config (where max_visible_objects is 16) we'd expect visible_dict['objects'].shape == (16,13).

But if you set padding=True, you actually get (7,13). The original unpadded shape is (5,13).

Additionally, visible_dict['traffic_lights'] and visible_dict['stop_signs'] have shapes (0,12) and (0,3) respectively when padding=True even though the maxes are nonzero.

The final shapes will also change depending on the vehicles and scenarios even when padding=True.

If I'm misunderstanding the expected behavior, please let me know.

Thanks!

Steps to reproduce

Here's a short script that reproduces the issue (run the same as examples/nocturne_functions.py)

import hydra
import numpy as np

from cfgs.config import PROJECT_PATH, get_scenario_dict, set_display_window
from nocturne import Simulation, Action


@hydra.main(config_path="../cfgs/", config_name="config")
def main(cfg):
    # load scenario. by default this won't have pedestrians or cyclists
    scenario_config = get_scenario_dict(cfg)
    sim = Simulation(scenario_path=str(PROJECT_PATH / 'examples' /
                                       'example_scenario.json'),
                     config=scenario_config)
    scenario = sim.getScenario()
    vehs = scenario.getObjectsThatMoved()

    visible_dict = scenario.visible_state(object=vehs[0],
                                          view_dist=80,
                                          view_angle=120 * (np.pi / 180),
                                          padding=True)
    
    print('Expecting {} objects, got {}'.format(scenario_config['max_visible_objects'], visible_dict['objects'].shape[0]))
    print('Expecting {} road points, got {}'.format(scenario_config['max_visible_road_points'], visible_dict['road_points'].shape[0]))
    print('Expecting {} traffic lights, got {}'.format(scenario_config['max_visible_traffic_lights'], visible_dict['traffic_lights'].shape[0]))
    print('Expecting {} stop signs, got {}'.format(scenario_config['max_visible_stop_signs'], visible_dict['stop_signs'].shape[0]))

main()

Relevant log output

No response

Comparison of missing driving simulators in the nocturne paper[Question]

Question

Hi,

I was reading the nocturne paper and came across Table 1 which describes a comparison between different autonomous driving open-source simulators. I have two queries below:

I wanted to mention a few of the simulators I know (due to my ongoing work in Ph.D.) which are missing in the comparison such as Flow, LGSVL, and AirSim. They are quite famous so If there is any specific reason for not comparing them please mention it here since I am also performing comparisons of such simulators in my research.
For Highway-env, I think they also support multi-agent driving (at least in a cooperative manner) just like the Flow simulator, but in Table 1 it is not shown as multi-agent.

Any response would be highly appreciated.

Build Error: sf_utils.h

Hi,
when running python setup.py develop the build process fails. The error message is the following:
... include/utils/sf_utils.h:29:6: error: ‘unique_ptr’ in namespace ‘std’ does not name a template type ...

It seems there is a missing include in this file. The error can be resolved by adding
#include <memory> to sf_utils.h.

[Bug] eval_sample_factory.py

Operating system

Ubuntu

Bug description

eval_sample_factory.py

Steps to reproduce

run python eval_sample_factory.py

Relevant log output

(nocturne) root@ryan:/home/ryan/nocturne/scripts/paper_plots# python eval_sample_factory.py 
[2023-11-09 11:49:08,314][26147] Default env families supported: ['doom_*', 'atari_*', 'dmlab_*', 'mujoco_*', 'MiniGrid*']
[2023-11-09 11:49:08,314][26147] Env registry entry created: my_custom_multi_env_
[2023-11-09 11:49:08,314][26147] Adding model class <class 'examples.sample_factory_files.run_sample_factory.CustomEncoder'> to registry (with name custom_env_encoder)
Traceback (most recent call last):
  File "eval_sample_factory.py", line 1317, in <module>
    sys.exit(main())
  File "eval_sample_factory.py", line 1045, in main
    for key in data_dicts[0].keys():
IndexError: list index out of range

[Question] How to import the provided expert trajectory and integrate it into the state?

Question

How to import the provided expert trajectory and integrate it into the state for IL?

[Feature] Scenario method to return all visible vehicles / pedestrians / cyclists

Feature

In title

Add support for running Nocturne entirely on GPU

@xiaomengy we've discussed this before and I'm just throwing it up on ths issue tracker so we don't forget about it

Upload the traffic light files

We have the no-tl files, might as well make the tl files too

[Feature] Don't install `cfgs` module as a top level Python import

Feature

In the process of packaging nocturne for Nixpkgs (NixOS/nixpkgs#208847 (comment)), it was discovered that nocturne doesn't just install "nocturne" as a top-level package -- eg import nocturne -- but also "cfgs". In fact, having cfgs available on the top-level is required in order to import nocturne:

nocturne/nocturne/__init__.py

Line 27 in ae0a4e3

 from cfgs.config import PROCESSED_TRAIN_NO_TL, PROCESSED_VALID_NO_TL, PROJECT_PATH 

Is this intentional? Polluting the top-level scope with "cfgs" seems undesirable. Would it be possible to move cfgs out of the top-level package namespace?

Add a ReadTheDocs

Problems with rendering.py

Hi,
Thanks for sharing this awesome environment.

I'm trying to run rendering.py and end up with IndexError.
It's happening at the third video creation with the following error:
source=scenario.getObjectsThatMoved()[3], IndexError: list index out of range

Same thing happens with the fourth video creation. It seems as though changing the index to -1 for both works but I'm not sure if this is a sound change.

Thanks!

[Question] Usage issues

Question

I am a beginner, my mentor asked me to configure Nocturne, so far I have cloned the code and downloaded the corresponding dataset, what do I need to do next to start running test cases and start using them

[Question]

Question

I am a beginner, sorry to disturb, I use conda env create -f environment.yml command when there is an error:
Pip subprocess error:
ERROR: Command errored out with exit status 1:
command: 'E:\Anaconda\envs\nocturne\python.exe' -c 'import io, os, sys, setuptools, tokenize; sys.argv[0] = '"'"'C'
cwd: C:\Users\FengmoZhang\AppData\Local\Temp\pip-install-3pa_xqy4\gym_c4c7742b0436454389c4ecc2a5b6b8ae
Complete output (1 lines):
error in gym setup command: 'extras_require' must be a dictionary whose values are strings or lists of strings cont.
----------------------------------------
WARNING: Discarding https://pypi.tuna.tsinghua.edu.cn/packages/f1/16/a421155206e7dc41b3a79d4e9311287b88c20140d567182839.
ERROR: Could not find a version that satisfies the requirement gym==0.20.0 (from versions: 0.0.2, 0.0.3, 0.0.4, 0.0.5, )
ERROR: No matching distribution found for gym==0.20.0

failed

CondaEnvException: Pip failed

Is there a solution? I tried to change multiple pip download sources, but I couldn't download them.

[Bug]

Operating system

Ubuntu 2204

Bug description

Hi there,

I encountered an error when executing nocturne/examples/create_env.py.

After the execution in a terminal, nothing showed up.

Before that, I strictly followed instructions and made sure every step was just the same as the readme file. And I've changed the
DATA_ FOLDER to the place where the valid and train data are stored.

Any help would be greatly appreciated! Thanks in advance

Steps to reproduce

strictly followed instructions and made sure every step was just the same as the readme file

Relevant log output

[Question]

Question

failed

CondaEnvException: Pip failed

Is there a solution? I tried to change multiple pip download sources, but I couldn't download them.

Disable pyvirtualdisplay for mac users

We've been doing this on a headless machine but pyvirtualdisplay seems to cause errors for mac users

Issues Running MAPPO

Hi Folks,

I'm trying to run "on-policy PPO" using python examples/on_policy_files/nocturne_runner.py algorithm=ppo and there are a couple of issues I'm encountering.

algo vs. algorithm: The config.yml file uses algorithm whereas the script uses cfg.algo. Switching algo to algorithm seems to fix the issue.
wandb_name seems to be missing from the cfg. To make it work, I just disabled use of wandb.
The wrapper environment calls len(self.vehicles) on line 30 which throws AttributeError: 'BaseEnv' object has no attribute 'vehicles'. Replacing self.vehicles with self.controlled_vehicles seems to solve the issue. Is this the correct way to fix it?

Thanks for your help.

The Dropbox abort the download

Operating system

Ubuntu

Bug description

The Dropbox abort the download of the 40GB package.

Steps to reproduce

No response

Relevant log output

No response

[Feature] Add installation instructions for M1 chips

Feature

The current version of SFML downloaded with Brew (2.5.1) is not compatible with M1 chips and we need to upgrade the installation instructions to account for this.

[Question] Optimal hyperparameters and scripts to reach 2000 steps/sec training speed

Question

Hello I am wondering which script and hyperparameters could achieve the 2000+ step/sec training speed as mentioned in the paper. So I have tried the following:

run_sample_factory.py algorithm=APPO
Problem: When using sample_factory library: parameters lr_schedule and max_entropy_coeff are missing, not sure what are the optimal numbers I should use.
run_rllib.py
Problem: same run time error for every worker, attached below:

nocturne_runner.py
Problem: the training speed is not that fast (around 100 steps/sec with around 40 fps). I have tried #38 and fps improved to around 80fps but the steps are still around the same.

My settings:
Code: newest code from main branch
OS: Ubuntu 20.04
GPU: RTX 3080 with CUDA 11.6
sample_factory: I have tried latest and aed6cc92a7eb3510c4d4bcfac083ced07b5222f9 (as mentioned in paper)

Please let me know if I made anything wrong when running the scripts. Thanks so much for answering!

[Bug] APPO

Operating system

Ubuntu
No response

Bug description

The buffer load without an end.

2023-11-08 09:27:34,872][06994] Worker 39 awakens!
[2023-11-08 09:27:34,879][06994] Waiting for 20 trajectory buffers...
[2023-11-08 09:27:35,186][06936] Waiting for 14 trajectory buffers...
[2023-11-08 09:27:35,194][06916] Waiting for 20 trajectory buffers..

Steps to reproduce

Execute: python examples/sample_factory_files/run_sample_factory.py algorithm=APPO

Relevant log output

2023-11-08 09:27:25,050][06984] Worker 35, sleep for 8.750 sec to decorrelate experience collection
Process Process-1:
Traceback (most recent call last):
  File "/root/anaconda3/envs/nocturne/lib/python3.8/multiprocessing/process.py", line 315, in _bootstrap
    self.run()
  File "/root/anaconda3/envs/nocturne/lib/python3.8/multiprocessing/process.py", line 108, in run
    self._target(*self._args, **self._kwargs)
  File "/root/anaconda3/envs/nocturne/lib/python3.8/site-packages/sample_factory/algorithms/appo/learner.py", line 1414, in _run
    rollouts = self._process_rollouts(rollouts, timing)
  File "/root/anaconda3/envs/nocturne/lib/python3.8/site-packages/sample_factory/algorithms/appo/learner.py", line 565, in _process_rollouts
    self._process_macro_batch(rollouts_to_process, batch_size, timing)
  File "/root/anaconda3/envs/nocturne/lib/python3.8/site-packages/sample_factory/algorithms/appo/learner.py", line 495, in _process_macro_batch
    assert macro_batch_size % self.cfg.rollout == 0
AssertionError

[Bug] Dependencies not listed in environment file

Operating system

No response

Bug description

There are some dependencies that are not listed in environment.yml, e.g:

hydra
gym
PyVirtualDisplay

Steps to reproduce

git clone https://github.com/facebookresearch/nocturne.git
cd nocturne
git submodule sync
git submodule update --init --recursive

# create the environment and install the dependencies
conda env create -f environment.yml

# activate the environment where the Python library should be installed
conda activate nocturne

# run the C++ build and install Nocturne into the simulation environment
python setup.py develop

# run example
python examples/create_env.py

Relevant log output

Traceback (most recent call last):
  File "examples/create_env.py", line 10, in <module>
    from nocturne.envs.wrappers import create_env
  File "/Users/julian/Desktop/Projects/RL/nocturne/nocturne/envs/__init__.py", line 2, in <module>
    from nocturne.envs.base_env import BaseEnv
  File "/Users/julian/Desktop/Projects/RL/nocturne/nocturne/envs/base_env.py", line 13, in <module>
    from gym import Env
ModuleNotFoundError: No module named 'gym'

[Feature] Is it possible to support user-specific map?

Feature

I would like to run some simulations in Toronto's road network; I am wondering if it is possible to add user-specific road networks.

Impossible to create render texture (failed to link the target texture to the frame buffer)

When running examples/create_env.py, I get this error:

Impossible to create render texture (unsupported anti-aliasing level) Requested: 4 Maximum supported: 1. Impossible to create render texture (failed to link the target texture to the frame buffer).

It seems that it might be caused by SFML package. To get SFML>=2.5, I downloaded it from: https://www.sfml-dev.org/download.php (apt-get install libsfml-dev installs version 2.4.2). Has anyone faced this issue?

[Question] How is the masking achieved?

Question

Hi，

This is a very commendable work and I'm currently using it for my research, but I'm currently having some problems with masking

How is the masking achieved?

Add CircleCI badge

We already have tests run but no badge confirming that they have passed

[Question] Need Help Installing

Question

Hello, I am a student doing AI research and currently trying to set up Nocturne into my school pc:

My Visual Studio Code runs into this error after trying to run the files, please let me know if you have any suggestions/advice!

[Question] SFML/Graphics.hpp file not found

I am on macOS Monterey and I have cmake and sfml installed via brew. However when I run, python3 setup.py develop I am getting the fatal error of SFML/Graphics.hpp' file not found. It appears that brew is installing SFML as below:

/usr/local/Cellar/sfml/2.5.1_2
├── INSTALL_RECEIPT.json
├── changelog.md
├── include
│   └── SFML
│       ├── Audio
│       ├── Audio.hpp
│       ├── Config.hpp
│       ├── GpuPreference.hpp
│       ├── Graphics
│       ├── Graphics.hpp
│       ├── Main.hpp
│       ├── Network
│       ├── Network.hpp
│       ├── OpenGL.hpp
│       ├── SFML
│       ├── System
│       ├── System.hpp
│       ├── Window
│       └── Window.hpp
├── lib
│   ├── cmake
│   │   └── SFML
│   ├── libsfml-audio.2.5.1.dylib
│   ├── libsfml-audio.2.5.dylib -> libsfml-audio.2.5.1.dylib
│   ├── libsfml-audio.dylib -> libsfml-audio.2.5.dylib
│   ├── libsfml-graphics.2.5.1.dylib
│   ├── libsfml-graphics.2.5.dylib -> libsfml-graphics.2.5.1.dylib
│   ├── libsfml-graphics.dylib -> libsfml-graphics.2.5.dylib
│   ├── libsfml-network.2.5.1.dylib
│   ├── libsfml-network.2.5.dylib -> libsfml-network.2.5.1.dylib
│   ├── libsfml-network.dylib -> libsfml-network.2.5.dylib
│   ├── libsfml-system.2.5.1.dylib
│   ├── libsfml-system.2.5.dylib -> libsfml-system.2.5.1.dylib
│   ├── libsfml-system.dylib -> libsfml-system.2.5.dylib
│   ├── libsfml-window.2.5.1.dylib
│   ├── libsfml-window.2.5.dylib -> libsfml-window.2.5.1.dylib
│   ├── libsfml-window.dylib -> libsfml-window.2.5.dylib
│   └── pkgconfig
│       ├── sfml-all.pc
│       ├── sfml-audio.pc
│       ├── sfml-graphics.pc
│       ├── sfml-network.pc
│       ├── sfml-system.pc
│       └── sfml-window.pc
├── license.md
├── readme.md
└── share
    └── SFML
        ├── doc
        ├── license.md
        └── readme.md

I ensured /usr/local/lib has the .dylib files for sfml and I manually copied the include directory from the downloaded SFML to /usr/local/include so that SFML/Graphics.hpp exists in there. None of this has worked in getting the project to compile and I am unsure how to proceed. Any tips on how to proceed to get this working?

Add an example of a trajectory optimizer

RuntimeError: DataLoader worker is killed by signal: Floating point exception.

Operating system

Ubuntu 18.04

Bug description

When running the imitation learning baseline, I am sometimes getting a floating point exception. Unfortunately, It's not deterministic and I cannot always reproduce. It just happens sometimes. Has anyone experienced this bug before?

Steps to reproduce

python examples/imitation_learning/train.py

Relevant log output

Error executing job with overrides: ['device=cuda:1'] Traceback (most recent call last): 
File "scripts/train.py", line 204, in  main dist = model.dist(states) 
File "/home/bernard.lange/imitation-learning-agents-research/./src/algos/imitation_learning/model.py", line 83, in dist return MultivariateNormal( File "/home/bernard.lange/miniconda3/envs/nocturne/lib/python3.8/site-packages/torch/distributions/multivariate_normal.py", line 146, in init super(MultivariateNormal, self).init(batch_shape, event_shape, validate_args=validate_args) 
File "/home/bernard.lange/miniconda3/envs/nocturne/lib/python3.8/site-packages/torch/distributions/distribution.py", line 53, in init valid = constraint.check(value) 
File "/home/bernard.lange/miniconda3/envs/nocturne/lib/python3.8/site-packages/torch/distributions/constraints.py", line 509, in check sym_check = super().check(value) 
File "/home/bernard.lange/miniconda3/envs/nocturne/lib/python3.8/site-packages/torch/distributions/constraints.py", line 490, in check return torch.isclose(value, value.mT, atol=1e-6).all(-2).all(-1) 
File "/home/bernard.lange/miniconda3/envs/nocturne/lib/python3.8/site-packages/torch/utils/data/_utils/signal_handling.py", line 66, in handler _error_if_any_worker_fails() 
RuntimeError: DataLoader worker (pid 7036) is killed by signal: Floating point exception.

ERROR: Unexpected floating-point exception encountered in worker.

IndexError: _Map_base::at when running sim = Simulation()

Hi,

I've just installed your package. When running the example Python script from 'Installing Nocturne', I am getting the following error:

Python 3.8.13 (default, Mar 28 2022, 11:38:47)
[GCC 7.5.0] :: Anaconda, Inc. on linux
Type "help", "copyright", "credits" or "license" for more information.
from nocturne import Simulation
sim = Simulation()
Traceback (most recent call last):
File "", line 1, in
IndexError: _Map_base::at

Have I missed something during the installation process?

Thanks

Add examples of permutation invariant architectures

Our current architecture in sample-factory is just an MLP encoder; I suspect a permutation invariant or GNN-based architecture would be better

[Bug] python examples/imitation_learning/train.py

Operating system

python examples/imitation_learning/train.py

Bug description

python examples/imitation_learning/train.py

Steps to reproduce

run python examples/imitation_learning/train.py
than choose option 3

Relevant log output

(nocturne) root@ryan:/home/ryan/nocturne# python examples/imitation_learning/train.py
WaymoDataset: loading 1000 files.
ImitationAgent(
  (nn): Sequential(
    (0): MeanStdFilter()
    (1): Linear(in_features=35110, out_features=1024, bias=True)
    (2): Tanh()
    (3): Sequential(
      (0): Linear(in_features=1024, out_features=256, bias=True)
      (1): Tanh()
    )
    (4): Sequential(
      (0): Linear(in_features=256, out_features=128, bias=True)
      (1): Tanh()
    )
  )
  (heads): ModuleList(
    (0): Linear(in_features=128, out_features=15, bias=True)
    (1): Linear(in_features=128, out_features=43, bias=True)
  )
)
Wrote configs at /checkpoint/root/nocturne/test/2023.11.08/test/09.54.05/train_logs/2023_11_08_09_54_07/configs.json
wandb: (1) Create a W&B account
wandb: (2) Use an existing W&B account
wandb: (3) Don't visualize my results
wandb: Enter your choice: 3
wandb: You chose 'Don't visualize my results'
wandb: ERROR Error while calling W&B API: user is not logged in (<Response [401]>)
Thread SenderThread:
Traceback (most recent call last):
  File "/root/anaconda3/envs/nocturne/lib/python3.8/site-packages/wandb/sdk/lib/retry.py", line 102, in __call__
    result = self._call_fn(*args, **kwargs)
  File "/root/anaconda3/envs/nocturne/lib/python3.8/site-packages/wandb/sdk/internal/internal_api.py", line 140, in execute
    return self.client.execute(*args, **kwargs)
  File "/root/anaconda3/envs/nocturne/lib/python3.8/site-packages/wandb/vendor/gql-0.2.0/wandb_gql/client.py", line 52, in execute
    result = self._get_result(document, *args, **kwargs)
  File "/root/anaconda3/envs/nocturne/lib/python3.8/site-packages/wandb/vendor/gql-0.2.0/wandb_gql/client.py", line 60, in _get_result
    return self.transport.execute(document, *args, **kwargs)
  File "/root/anaconda3/envs/nocturne/lib/python3.8/site-packages/wandb/vendor/gql-0.2.0/wandb_gql/transport/requests.py", line 39, in execute
    request.raise_for_status()
  File "/root/anaconda3/envs/nocturne/lib/python3.8/site-packages/requests/models.py", line 1021, in raise_for_status
    raise HTTPError(http_error_msg, response=self)
requests.exceptions.HTTPError: 401 Client Error: Unauthorized for url: https://api.wandb.ai/graphql

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/root/anaconda3/envs/nocturne/lib/python3.8/site-packages/wandb/apis/normalize.py", line 22, in wrapper
    return func(*args, **kwargs)
  File "/root/anaconda3/envs/nocturne/lib/python3.8/site-packages/wandb/sdk/internal/internal_api.py", line 1318, in upsert_run
    response = self.gql(mutation, variable_values=variable_values, **kwargs)
  File "/root/anaconda3/envs/nocturne/lib/python3.8/site-packages/wandb/sdk/lib/retry.py", line 118, in __call__
    if not check_retry_fn(e):
  File "/root/anaconda3/envs/nocturne/lib/python3.8/site-packages/wandb/util.py", line 875, in no_retry_auth
    raise CommError("Invalid or missing api_key. Run `wandb login`")
wandb.errors.CommError: Invalid or missing api_key. Run `wandb login`

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/root/anaconda3/envs/nocturne/lib/python3.8/site-packages/wandb/sdk/internal/internal_util.py", line 51, in run
    self._run()
  File "/root/anaconda3/envs/nocturne/lib/python3.8/site-packages/wandb/sdk/internal/internal_util.py", line 102, in _run
    self._process(record)
  File "/root/anaconda3/envs/nocturne/lib/python3.8/site-packages/wandb/sdk/internal/internal.py", line 310, in _process
    self._sm.send(record)
  File "/root/anaconda3/envs/nocturne/lib/python3.8/site-packages/wandb/sdk/internal/sender.py", line 234, in send
    send_handler(record)
  File "/root/anaconda3/envs/nocturne/lib/python3.8/site-packages/wandb/sdk/internal/sender.py", line 692, in send_run
    self._init_run(run, config_value_dict)
  File "/root/anaconda3/envs/nocturne/lib/python3.8/site-packages/wandb/sdk/internal/sender.py", line 716, in _init_run
    server_run, inserted = self._api.upsert_run(
  File "/root/anaconda3/envs/nocturne/lib/python3.8/site-packages/wandb/apis/normalize.py", line 58, in wrapper
    raise CommError(message, err).with_traceback(sys.exc_info()[2])
  File "/root/anaconda3/envs/nocturne/lib/python3.8/site-packages/wandb/apis/normalize.py", line 22, in wrapper
    return func(*args, **kwargs)
  File "/root/anaconda3/envs/nocturne/lib/python3.8/site-packages/wandb/sdk/internal/internal_api.py", line 1318, in upsert_run
    response = self.gql(mutation, variable_values=variable_values, **kwargs)
  File "/root/anaconda3/envs/nocturne/lib/python3.8/site-packages/wandb/sdk/lib/retry.py", line 118, in __call__
    if not check_retry_fn(e):
  File "/root/anaconda3/envs/nocturne/lib/python3.8/site-packages/wandb/util.py", line 875, in no_retry_auth
    raise CommError("Invalid or missing api_key. Run `wandb login`")
wandb.errors.CommError: Invalid or missing api_key. Run `wandb login`
wandb: ERROR Internal wandb error: file data was not synced

facebookresearch / nocturne Goto Github PK

nocturne's Issues

Question

Operating system

Bug description

Steps to reproduce

Relevant log output

Question

Feature

Feature

Operating system

Bug description

Steps to reproduce

Relevant log output

Question

Question

Operating system

Bug description

Steps to reproduce

Relevant log output

Question

Operating system

Bug description

Steps to reproduce

Relevant log output

Question

Feature

Feature

Question

Question

Operating system

Bug description

Steps to reproduce

Relevant log output

Question

Operating system

Bug description

Steps to reproduce

Relevant log output

Feature

Question

Operating system

Bug description

Steps to reproduce

Relevant log output

Operating system

Bug description

Steps to reproduce

Relevant log output

Feature

Question

Question

Operating system

Bug description

Steps to reproduce

Relevant log output

Operating system

Bug description

Steps to reproduce

Relevant log output

Recommend Projects

Recommend Topics

Recommend Org

Jobs