Commits

Branch selector

 master

User selector

All users

Datepicker

All time

Commit History

Commits on Mar 16, 2021

• Merge pull request #4 from codingfabi/improve-qLearning-agent

  Show description for 996d270
  codingfabi

  authored
  996d270

  Copy full SHA for 996d270

• model that at least collects coins without crates

  

  Fabian Kneissl

  committed
  a165210

  Copy full SHA for a165210

• Merge pull request #3 from codingfabi/add-plotting-to-custom-agent

  Show description for 88d143a
  codingfabi

  authored
  88d143a

  Copy full SHA for 88d143a

• add plotting function for game results

  

  Fabian Kneissl

  committed
  9f9a21b

  Copy full SHA for 9f9a21b

Commits on Mar 15, 2021

• Merge pull request #2 from codingfabi/prepare_trainable_custom_agent

  Show description for adab9bd
  codingfabi

  authored
  adab9bd

  Copy full SHA for adab9bd

• another desperate attempt to improve model

  

  Fabian Kneissl

  committed
  a27b08c

  Copy full SHA for a27b08c

• add coins to surroundings

  

  Fabian Kneissl

  committed
  11e0ae4

  Copy full SHA for 11e0ae4

• commit intermediate state

  

  Fabian Kneissl

  committed
  f2a14d8

  Copy full SHA for f2a14d8

Commits on Mar 14, 2021

• add all coins to features

  

  Fabian Kneissl

  committed
  bf25020

  Copy full SHA for bf25020

• adjust prediction function to work if training mode is not activated

  

  Fabian Kneissl

  committed
  0c5c582

  Copy full SHA for 0c5c582

• trained modell that doesn't bomb yet

  

  Fabian Kneissl

  committed
  6a19474

  Copy full SHA for 6a19474

• change q table structure to dict

  

  Fabian Kneissl

  committed
  a9dcd7e

  Copy full SHA for a9dcd7e

• try to fix more bugs, model gets trained now but cannot load pickled data

  

  Fabian Kneissl

  committed
  c3afc66

  Copy full SHA for c3afc66

Commits on Mar 12, 2021

• fix all errors that prevent agent from working

  

  Fabian Kneissl

  committed
  72fdf23

  Copy full SHA for 72fdf23

• transfrom state_to_feature function into something that works, even though it doesn't make sense anymore

  

  Fabian Kneissl

  committed
  848acd1

  Copy full SHA for 848acd1

• fix bug from original game environment

  

  Fabian Kneissl

  committed
  3c8601a

  Copy full SHA for 3c8601a

• update game environment

  

  Fabian Kneissl

  committed
  351fd40

  Copy full SHA for 351fd40

• transfer alpha and gamma to model

  

  Fabian Kneissl

  committed
  0a618a7

  Copy full SHA for 0a618a7

• map actions and state to numeric stuff

  

  Fabian Kneissl

  committed
  00ca9ad

  Copy full SHA for 00ca9ad

• function to map actions to numeric, unsure if needed

  

  Fabian Kneissl

  committed
  17ea83d

  Copy full SHA for 17ea83d

• naive try to implement updating of q table

  

  Fabian Kneissl

  committed
  a72a59e

  Copy full SHA for a72a59e

Commits on Mar 10, 2021

• setup basic state_to_features function

  

  Fabian Kneissl

  committed
  9c90c27

  Copy full SHA for 9c90c27

Commits on Feb 26, 2021

• add git ignore file to avoid commiting log files and python cash

  

  Fabian Kneissl

  committed
  a0efba1

  Copy full SHA for a0efba1

• shift code structure into different classes

  

  Fabian Kneissl

  committed
  0e75cb5

  Copy full SHA for 0e75cb5

• fix missnamed variable

  

  Fabian Kneissl

  committed
  9116c96

  Copy full SHA for 9116c96

• initialize basic trainable agent

  

  Fabian Kneissl

  committed
  8772388

  Copy full SHA for 8772388

Commits on Feb 19, 2021

• Merge pull request #1 from codingfabi/implement_first_custom_agent

  Show description for fbd1746
  codingfabi

  authored
  fbd1746

  Copy full SHA for fbd1746

• implement first custom random agent

  

  Fabian Kneissl

  committed
  c8a80b2

  Copy full SHA for c8a80b2

Commits on Feb 16, 2021

• add client implementation from prof koethe to project

  

  Fabian Kneissl

  committed
  4d3c987

  Copy full SHA for 4d3c987

Commits on Feb 14, 2021

• Create README.md

  codingfabi

  authored
  5327c0f

  Copy full SHA for 5327c0f