Search results

Stretch experiments: behavior cloning, pretrain, randomization, recurrent, distillation

Status: Open.

#353 In omuryildirim/before-evening;

· omuryildirim-fenomark opened on May 22, 2026

Test pure-pixel policy: drop speed and action-history scalars

Status: Open.

#352 In omuryildirim/before-evening;

· omuryildirim-fenomark opened on May 22, 2026

Implement hybrid CNN + scalar policy network

Status: Open.

#351 In omuryildirim/before-evening;

· omuryildirim-fenomark opened on May 22, 2026

Add headless frame-capture pipeline for vision-based policies

Status: Open.

#350 In omuryildirim/before-evening;

· omuryildirim-fenomark opened on May 22, 2026

Upgrade Q-learning to DQN with target network and persistent replay buffer

Status: Open.

#349 In omuryildirim/before-evening;

· omuryildirim-fenomark opened on May 22, 2026

Add deterministic evaluation harness and baseline bot

Status: Open.

#348 In omuryildirim/before-evening;

· omuryildirim-fenomark opened on May 22, 2026