Skip to content
View H0rvex's full-sized avatar
1 follower · 0 following
  • Budapest, Hungary

Block or report H0rvex

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
H0rvex/README.md

Hi, I am Győző!

Budapest, HU → ELTE IK · CS BSc

Going deep on RL, computer vision, and the engineering side of deep learning systems — building foundations for embodied AI work in MuJoCo, Isaac Lab, and ROS 2.

❯ projects

  • SIM/ROS  isaaclab-quadruped-robust-locomotion — Isaac Lab Go2 locomotion with PPO/RSL-RL training, robustness/domain-randomization sweeps, policy export, ROS 2 bridge, fake robot-state publisher, and hardware-interface validation.
  • RL  mujoco-continuous-control — PPO for MuJoCo Walker2d/Ant continuous-control locomotion with squashed Gaussian policies, multi-seed evaluation, rollout videos, diagnostics, Docker
  • RL  rl-gymnasium — REINFORCE, DQN, PPO in PyTorch with reproducible configs, multi-seed eval, Docker
  • CV  unet-pet-segmentation — U-Net semantic segmentation on Oxford-IIIT Pet, 0.7422 mIoU
  • DL  transformer-from-scratch — Attention, embeddings, training loop, from scratch in PyTorch
  • CV  resnet-cifar10 — ResNet on CIFAR-10, ~90% accuracy, clean training pipeline

❯ learning_next

  • SIM  MuJoCo + Isaac Lab — embodied control simulation
  • RL   Offline RL, Decision Transformers
  • CV   Vision-based control, perception-to-action pipelines
  • ROS  ROS 2 middleware fundamentals

Pinned Loading

  1. isaaclab-quadruped-robust-locomotion isaaclab-quadruped-robust-locomotion Public

    Reproducible Isaac Lab quadruped locomotion project with robustness evaluation, policy export contracts, and ROS 2 hardware-interface.

    Python

  2. mujoco-continuous-control mujoco-continuous-control Public

    From-scratch PyTorch PPO for MuJoCo continuous control: Walker2d-v5, Ant-v5, and Humanoid-v5 with multi-seed evaluation, rollout GIFs, curves, and Docker.

    Python

  3. rl-gymnasium rl-gymnasium Public

    From-scratch reinforcement learning algorithms in PyTorch and Gymnasium, including REINFORCE, DQN, and PPO with evaluation, logging, and clean training scripts.

    Python

  4. transformer-from-scratch transformer-from-scratch Public

    From-scratch PyTorch Transformer project with IMDb classification and Tiny Shakespeare GPT-style language modeling experiments.

    Python

  5. unet-pet-segmentation unet-pet-segmentation Public

    From-scratch PyTorch U-Net for Oxford-IIIT Pet semantic segmentation with augmentation, Dice+CE loss, AMP, IoU metrics, qualitative grids, curves, and tests.

    Python

  6. resnet-cifar10 resnet-cifar10 Public

    PyTorch ResNet for CIFAR-10 image classification with reproducible training, evaluation metrics, curves, and clean experiment structure.

    Python