snipercell_RL / README.md
webxos's picture
Update README.md
2c31113 verified
|
Raw
History Blame Contribute Delete
3.25 kB
metadata
license: mit
task_categories:
  - reinforcement-learning
  - robotics
  - tabular-classification
  - time-series-forecasting
tags:
  - timeseries
  - tabular
  - robotics
  - RL
  - RLHF
  - reinforcement-learning
  - experience-replay
  - state-action-reward
  - gaming
  - gym
  - video-games
  - games

Website GitHub Hugging Face Follow on X

 _____ _   _ ___________ ___________   _____  _____ _      _     
/  ___| \ | |_   _| ___ \  ___| ___ \ /  __ \|  ___| |    | |    
\ `--.|  \| | | | | |_/ / |__ | |_/ / | /  \/| |__ | |    | |    
 `--. \ . ` | | | |  __/|  __||    /  | |    |  __|| |    | |    
/\__/ / |\  |_| |_| |   | |___| |\ \  | \__/\| |___| |____| |____
\____/\_| \_/\___/\_|   \____/\_| \_|  \____/\____/\_____/\_____/

SNIPER CELL - SAMPLE TRAINING DATA EXPORT

This dataset sample was created with SNIPER CELL by webXOS located in the /gym/ folder, download it to create your own datasets. Sniper cell is a first person shooter based gym that runs minimal in browsers. Users can play the "gym" to record and export training data as seen below.

INFO

Sample of Sniper Cell Game Data: Exported: 2026-01-08T02:20:14.260Z Player Level: 9 Total XP: 15843 Waves Completed: 3 Drones Destroyed: 22 Accuracy: 21%

FILES INCLUDED:

  1. xforc3d_training_data.json - Complete dataset in JSON format
  2. shots_data.csv - All shots fired by player and drones
  3. hits_data.csv - All hits and drone destructions
  4. level_ups_data.csv - Level progression data
  5. waves_data.csv - Wave completion data

DATA STRUCTURE:

  • Each action is timestamped for time-series analysis
  • XP rewards and penalties are recorded
  • Position data for spatial analysis
  • Level-based difficulty scaling included

TIERED DRONE SYSTEM:

Level 1-10: 4-5 drones max Level 11-20: 6 drones max Level 21-30: 7 drones max Level 31-40: 8 drones max Level 41-50: 9 drones max Level 51-99: 10 drones max

USE FOR TRAINING:

This data can be used to train reinforcement learning models for drone combat AI, player behavior prediction, or difficulty balancing algorithms.

COMPRESSION:

This ZIP file supports GZIP, ZSTD, BZ2, LZ4, and LZMA decompression.