xavierpuigf / watch_and_help Goto Github PK
View Code? Open in Web Editor NEWCode for the paper Watch-And-Help: A Challenge for Social Perception and Human-AI Collaboration
License: Other
Code for the paper Watch-And-Help: A Challenge for Social Perception and Human-AI Collaboration
License: Other
Hi,
When I install the dependency of this repo, I got an error:
Processing /tmp/build/80754af9/numpy_and_numpy_base_1596233707986/work
ERROR: Could not install packages due to an OSError: [Errno 2] No such file or directory: '/tmp/build/80754af9/numpy_and_numpy_base_1596233707986/work'
How should I resolve this? Thank you!
It seems like the watch phase training scripts are not pushed to the repository. I assume they run the predicate-train script with some custom arguments and hyperparameters, but I'm not sure what they are. Do let me know if you are able to upload them! Thanks lots.
Hi,
First thanks for your great work! But I have three questions:
You said that "using the v2.2.0 version of the executable". However, there isn't a version that matches to v2.2.0 in http://virtual-home.org/documentation/v2.2.0/downloads/downloads.html. And some versions are not available. When I click "Download Linux x86-64 version" in the README of this repo ,it comes up with "The requested URL was not found on this server".
When opening an simulator in the Linux environment, it can't connect successfully.
My codes are:
parser.add_argument('--exec_file', type=str,
default='/home/featurize/work/test/watch_and_helpmain/virtualhome/simulation/unity_simulator/linux_exec.v2.3.0/linux_exec.v2.3.0.x86_64',
help='Use unity editor')
comm = comm_unity.UnityCommunication(port=args.port,
file_name=args.exec_file,
no_graphics=True,
logging=False,
x_display=args.display)`
The errors are:
platform
linux
yes
['/home/featurize/work/test/watch_and_help-main/virtualhome/simulation/unity_simulator/linux_exec.v2.3.0/linux_exec.v2.3.0.x86_64', '-batchmode', '-nographics', '-http-port=8080', '-logFile /home/featurize/work/test/watch_and_help-main/gen_data/Player_8080.log']
Getting connection...
<Response [502]>
error: XDG_RUNTIME_DIR not set in the environment.
<Response [502]>
<Response [502]>
<Response [502]>
<Response [502]>
CLOSING PROC
'Error processing input graph: Object reference not set to an instance of an object'
The errors are:
Using backend: pytorch
Init Env
Getting connection...
Resetting... Envid: 3. Taskid: 6. Index: 6
E:\Anacoda_file\Anacoda3\lib\site-packages\pyreadline\py3k_compat.py:8: DeprecationWarning:
Using or importing the ABCs from 'collections' instead of from 'collections.abc' is deprecated since Python 3.3, and in 3.9 it will stop working
d:\watch_and_help-main\envs\unity_environment.py(158)reset()
157 ipdb.set_trace()
--> 158 print("Error expanding scene")
Resetting... Envid: 3. Taskid: 6. Index: 6
E:\Anacoda_file\Anacoda3\lib\site-packages\pyreadline\py3k_compat.py:8: DeprecationWarning:
Using or importing the ABCs from 'collections' instead of from 'collections.abc' is deprecated since Python 3.3, and in 3.9 it will stop working
d:\watch_and_help-main\envs\unity_environment.py(158)reset()
157 ipdb.set_trace()
--> 158 print("Error expanding scene")
159 ipdb.set_trace()
ipdb>
Hi,
When I run vh_init.py to create my own Dataset, it got the following error:
HTTPConnectionPool(host='127.0.0.1', port=8092): Max retries exceeded with URL: / (Caused by NewConnectionError('<urllib3.connection.HTTPConnection object at 0x7f4bc4d2c090>: Failed to establish a new connection: [Errno 111] Connection refused'))
I guess there could be an error with my subprocess_args in simulation/unity_simulator/communication.py. The following is the value of my subprocess_args:
0: '/home/.../v2/virtualhome/simulation/linux_exec.v2.2.4.x86_64
1: -batchmode
2: -nographics
3: -http-port=8092
4: -logFile /home/.../v2/watch_and_help/Player_8092.log
I can start a communication with the simulator manually, not automatically
Hi, can you share which linux executable to use exactly, I tried to use a couple of versions (v2.30, v2.2.4, v2.0.1), but I get this error: 'Error processing input graph: Object reference not set to an instance of an object'
.
The link on the README seems to be broken.
I am trying to run test_single_agent.py
with the given dataset test_env_set_help.pik
Hello, whenever I run the test_single_agent.py
with the recording option set as True for the self.comm.render_script()
function that gets called inside the step()
function in envs/unity_environment.py
, the program hangs for a while until a time out error happens:
Time out for action: {0: '[grab] <cupcake> (378)'}
Is there a way to record what is happening during the simulation when using the code in here?
Somehow if I only use virtual-home with the jupyter notebooks provided this doesn't happen!
Hi~Whether the v2.2.2 of virtualhome is legal for your program?
I tried to find v2.2.0 on the virtualhome site but it is unaccessible now or can you share the executable of v2.2.0 ?
Thanks very much!
Hi, I'm really interested in your works. Recently, I tried to understand the details of Alice's policy by reading your code. However, due to the lack of annotations in the code, I still have some problems:
When testing an hp agent, how to initialize the belief? Is it initialized by the ground truth meta-data across all the rooms in the dataset?
How does belief initialize unseen node's position and rotation e.t.c.? Intuitively, belief should only initialize edges. But it seems that each node in belief.sampled_state['nodes'] has 'position' and 'rotation' attributes.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.