HN Explorer

genji970

↗ HN profile [ 69.3 ms ] full profile
Karma
0
Created
()
Submissions
0
  1. Applying zigzag algebra in abstract algebra to improve LLM training (zenodo.org)
    1 points 0
  2. 1 points 0
  3. 1 points 0
  4. 1 points 0
  5. 1 points 0
  6. 1 points 0
  7. 1 points 0
  8. 1 points 0
  9. 1 points 0
  10. 1 points 0
  11. 1 points 0
  12. 1 points 0
  13. Fixing hallucination in LLM prediction with only one 48gib GPU (zenodo.org)
    1 points 0
  14. 1 points 0
  15. 1 points 0
  16. 1 points 0
  17. 1 points 0
  18. 1 points 0
  19. Show HN: Fine tuning and RLHF mistralai 7B using DeepSpeed (github.com)
    1 points 0
  20. Show HN: RLHF and Lora finetuning to mistralai 7B with DeepSpeed learning (github.com)

    Using multiple gpus, training 7B model with lora and RLHF with external dataset.

    1 points 0
  21. Show HN: Mistralai-7B distributed learning using DeepSpeed pipeline (github.com)

    Currently, I built basic pipeline to do lora fine tuning with multiple gpus. Samples with alpaca dataset works fine. data pipeline is in progress.

    4 points 0
  22. Show HN: Vlm in 3D PC, 16 shot scanobjectnn top1 acc: 99.91 (github.com)

    I'm not sure this accuracy is possible. But when i do 4 shot then val acc for scanobjectnn drops to about 90%. And for modelnet40 16 shot, it shows 96 val acc in top1. So I share my experiment result.

    1 points 0
  23. Show HN: I achieved over 10% improvement on 3D vision PointCLIP (github.com)

    Hello everyone! This project is about achieving more than 96 % top1 accuracy on modelnet40 test dataset with light weighted 3d custome models. projecting 3d pointcloud dataset(with gaussian splatting method) into…

    2 points 0