kenjyoung / mctx_learning_demo Goto Github PK
View Code? Open in Web Editor NEWLicense: MIT License
License: MIT License
Is there any visualization of the task like I can use the trained model?
Thanks for preparing the whole training script and the JAX environment.
I have minor suggestion to improve the target for the value network.
When using sequential halving, the search does exploration to find the best action.
The root value is then possibly underestimated, because the average value includes the values of the explored actions.
mctx_learning_demo/basic_tree_search.py
Line 189 in e79a655
A better target for the value network would be the Q-value of the selected action:
search_value = policy_output.search_tree.qvalues(policy_output.action)
I install Haiku through this command:
pip install git+https://github.com/deepmind/dm-haiku
But I got this error:
Traceback (most recent call last):
File "basic_tree_search.py", line 2, in <module>
import haiku as hk
File "/home/percyp/anaconda3/envs/isaac/lib/python3.8/site-packages/haiku/__init__.py", line 19, in <module>
from haiku import data_structures
File "/home/percyp/anaconda3/envs/isaac/lib/python3.8/site-packages/haiku/data_structures.py", line 18, in <module>
from haiku._src.data_structures import to_haiku_dict
File "/home/percyp/anaconda3/envs/isaac/lib/python3.8/site-packages/haiku/_src/data_structures.py", line 30, in <module>
from haiku._src import utils
File "/home/percyp/anaconda3/envs/isaac/lib/python3.8/site-packages/haiku/_src/utils.py", line 42, in <module>
def auto_repr(cls: type[Any], *args, **kwargs) -> str:
TypeError: 'type' object is not subscriptable
Am I installing from the wrong place? Could you help me solve this problem?
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.