A platform for developing AI systems as described in A Roadmap towards Machine Intelligence - http://arxiv.org/abs/1511.08130
Home Page: https://research.fb.com/projects/commai
Hello facebookresearchers,
As you may be aware, I am expanding the foundations built here in the Revised-CommAI-env at https://github.com/stephenbhope/Revised-CommAI-env. I would like to fully recognise the facebookresearch team and their work.
I would like to know what link and art you prefer, in addition if any of the CommAI-env group (committers or advisors) would like to include a personal link with their name. A developer page, personal project, or third party link you support related to the effort. Alternately, a press pack would be great if available.
.
The beta will be released and promoted in my local Machine Learning Meetup group, a 1700+ member Ottawa based interdisciplinary group with members from all levels of government (Canada), CSIS (Canadian security), Industry Leaders, Academics, Students, and other Enthusiasts.
Replies will be received if posted to this issue on this github repository, at my repository, or directly to [email protected],
[I believe some cross-talk was echoed into your mail-chimp list, I am not a subscriber]
In the Running section, the command should be
cp tasks_config.sample.json tasks_config.json
and not
cp task_config.sample.json tasks_config.json
(cross post to facebook group)
Hello, I have been reviewing the commAI-env code on github, and would like to work on bringing it up to full PEP8 compliance, as preparation for CI automated testing with travis-ci.org or a similar service.
I feel it would help bring the code base to a single linted, and tested standard and state as required by the contributor guidelines.
Updated 19 April: After a line by line examination of the all the code I would like to propose altering the code to Python >=3.5 for numerous reasons. Most importantly for the longevity of what looks like a very solid concept as Python 2.7 EOL is 2020.
There are no working examples in the documentation.
While all the unit tests pass, I'm not able to achieve any score greater than 0.
for example:
_______________________cucumber in john's basket is healthy. how is cucumber? [ ] Total time: 511
________________________ healthy[0010000 ] Total reward: 0
input:
So how about giving some examples that actually work (in the documentation)?
Also what kind of grammar do you have for your languages? or do I have to dig through your code to figure that out also?
When I write an answer and press enter, the answer is appended to the 2nd line and nothing else happens.
___which objects are juicy in mary's basket? [ ] Total time: 687
____ berries I don't know ? ans1 ans2 ans3 ans4[0010000 ] Total reward: 0
input:
I get this when I run any/all of the sample learners. It answers a few questions, does pretty crap and then half way through it throws this error. (el capitan, anaconda 3.5, ?)
Traceback (most recent call last):
File "run.py", line 191, in <module>
main()
File "run.py", line 83, in main
session.run()
File "/Volumes/ExtraDiskSpace/Documents/CommAI-env/src/core/session.py", line 52, in run
token, reward = self._env.next(token)
File "/Volumes/ExtraDiskSpace/Documents/CommAI-env/src/core/environment.py", line 113, in next
self._switch_new_task()
File "/Volumes/ExtraDiskSpace/Documents/CommAI-env/src/core/environment.py", line 279, in _switch_new_task
self._current_task.start(self)
File "/Volumes/ExtraDiskSpace/Documents/CommAI-env/src/core/task.py", line 451, in start
self._env.raise_event(Start())
File "/Volumes/ExtraDiskSpace/Documents/CommAI-env/src/core/environment.py", line 212, in raise_event
return self.event_manager.raise_event(event)
File "/Volumes/ExtraDiskSpace/Documents/CommAI-env/src/core/events.py", line 70, in raise_event
trigger.event_handler(observer, event)
File "/Volumes/ExtraDiskSpace/Documents/CommAI-env/src/tasks/competition/navigation.py", line 589, in on_start
pe(np, objects[self.dir2obj[i]], True, True)
TypeError: 'set' object does not support indexing```
Also, I tried the human-mode and couldn't get any reward. Am I just being stupid?
When trying to run with Python 2.7.12 from Anaconda on Windows 10, I get
Traceback (most recent call last):
File "run.py", line 18, in <module>
from core.environment import Environment
File "C:\Users\oldhe\Dropbox\CS\ML\CommAI-env\src\core\environment.py", line 13, in <module>
from core.task import StateChanged, MessageReceived, \
File "C:\Users\oldhe\Dropbox\CS\ML\CommAI-env\src\core\task.py", line 12, in <module>
from core.aux.observer import Observable
ImportError: No module named aux.observer
Hi,
I am trying to run the tests by running under the src folder:
python -m unittest discover
but I receive the following error:
======================================================================
ERROR: testTask11 (tasks.micro.tests.test_nano.TestRepetitionTasks)
----------------------------------------------------------------------
Traceback (most recent call last):
File "/Users/samurai/Documents/HKUST/CommAI-env/src/tasks/micro/tests/test_nano.py", line 103, in testTask11
correct=True)
File "/Users/samurai/Documents/HKUST/CommAI-env/src/tasks/micro/tests/test_nano.py", line 26, in trySolution
serializer=IdentitySerializer()) as m:
File "/usr/local/Cellar/python3/3.6.2/Frameworks/Python.framework/Versions/3.6/lib/python3.6/contextlib.py", line 81, in __enter__
return next(self.gen)
File "/Users/samurai/Documents/HKUST/CommAI-env/src/tasks/competition/tests/helpers.py", line 156, in task_messenger
m = EnvironmentMessenger(env, serializer)
File "/Users/samurai/Documents/HKUST/CommAI-env/src/tasks/competition/tests/helpers.py", line 26, in __init__
self.init()
File "/Users/samurai/Documents/HKUST/CommAI-env/src/tasks/competition/tests/helpers.py", line 30, in init
first_bit, reward = self._env.next(None)
File "/Users/samurai/Documents/HKUST/CommAI-env/src/core/environment.py", line 92, in next
self._switch_new_task(train_mode=not(test_mode))
File "/Users/samurai/Documents/HKUST/CommAI-env/src/core/environment.py", line 257, in _switch_new_task
self._current_task = self._task_scheduler.get_next_task(train_mode=train_mode)
TypeError: get_next_task() got an unexpected keyword argument 'train_mode'
======================================================================
ERROR: testReverseXTask (tasks.micro.tests.test_small_comp.TestSmallCompTasks)
----------------------------------------------------------------------
Traceback (most recent call last):
File "/Users/samurai/Documents/HKUST/CommAI-env/src/tasks/micro/tests/test_small_comp.py", line 53, in testReverseXTask
self._test_solution(small_comp.ReverseXTask, get_correct_answer, True)
File "/Users/samurai/Documents/HKUST/CommAI-env/src/tasks/micro/tests/test_small_comp.py", line 20, in _test_solution
serializer=IdentitySerializer()) as m:
File "/usr/local/Cellar/python3/3.6.2/Frameworks/Python.framework/Versions/3.6/lib/python3.6/contextlib.py", line 81, in __enter__
return next(self.gen)
File "/Users/samurai/Documents/HKUST/CommAI-env/src/tasks/competition/tests/helpers.py", line 156, in task_messenger
m = EnvironmentMessenger(env, serializer)
File "/Users/samurai/Documents/HKUST/CommAI-env/src/tasks/competition/tests/helpers.py", line 26, in __init__
self.init()
File "/Users/samurai/Documents/HKUST/CommAI-env/src/tasks/competition/tests/helpers.py", line 30, in init
first_bit, reward = self._env.next(None)
File "/Users/samurai/Documents/HKUST/CommAI-env/src/core/environment.py", line 92, in next
self._switch_new_task(train_mode=not(test_mode))
File "/Users/samurai/Documents/HKUST/CommAI-env/src/core/environment.py", line 257, in _switch_new_task
self._current_task = self._task_scheduler.get_next_task(train_mode=train_mode)
TypeError: get_next_task() got an unexpected keyword argument 'train_mode'
----------------------------------------------------------------------
Ran 52 tests in 0.022s
FAILED (errors=37)
the error is a bit longer, but I think you can guess what's going on.
I have tried with both python 2.7 and python 3.6, and with both Mac OS and Ubuntu OS.
Thanks in advance
Andrea
Python curses as implemented in this program side scrolls to the left.
it makes it unusable on anything less than a super wide screen,
there is also no history.
Also only one line from teacher is visible, in available examples it seems teacher sometimes is supposed to output more than one line. Net result is teacher asks things which are impossible to know, like which fruit are ripe in mary's basket, having no previous knowledge of mary's basket.
Please have the option of a normal console interface (using stdout and stdin) as default.
Then it would scroll up, like all normal console interfaces, and work on all sizes of console.
Are there any plans for creating a group/community around CommAI-env like the one we have for say fasttest
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
Django
The Web framework for perfectionists with deadlines.
Laravel
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
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.
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
Microsoft
Open source projects and samples from Microsoft.
Google
Google ❤️ Open Source for everyone.
Alibaba
Alibaba Open Source for everyone
D3
Data-Driven Documents codes.
Tencent
China tencent open source team.