https://github.com/kraglik/htmcl

HTM implementation on the GPU
https://github.com/kraglik/htmcl

gpgpu hierarchical-temporal-memory opencl

Last synced: about 1 month ago
JSON representation

HTM implementation on the GPU

• Host: GitHub
• URL: https://github.com/kraglik/htmcl
• Owner: kraglik
• License: lgpl-2.1
• Created: 2021-01-24T15:36:52.000Z (over 4 years ago)
• Default Branch: master
• Last Pushed: 2021-05-15T17:16:49.000Z (about 4 years ago)
• Last Synced: 2025-02-14T12:16:12.837Z (3 months ago)
• Topics: gpgpu, hierarchical-temporal-memory, opencl
• Language: C
• Homepage:
• Size: 138 KB
• Stars: 1
• Watchers: 3
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
## HTMCL

This repo contains an implementation of Hierarchical Temporal Memory in OpenCL.

TODO List:

- [x] Adapt [KMA](https://github.com/RSpliet/KMA) for use with PyOpenCL and test it

    - [x] Test it (something near 6 millions malloc calls per second)

    - [x] Fix that bug with code failing on Nvidia GPU. It was caused by wrong objects alignment in memory in case of 64-bit GPU.

    - [x] Fix that error with being unable to allocate more than 128 megabytes on x64 GPUs (probably fixed).

- [x] Implement random

    - [x] Test it

- [x] Implement generic list

    - [x] Test it

- [x] Implement input layer

- [x] Implement spatial pooler

- [ ] Implement temporal pooler

- [ ] Implement classifier

- [ ] Implement multi-layer architecture

- [ ] Write a proper documentation