Clarify pre-training vs fine-tuning phases and dev test purpose

- Pre-training: diffusion model on offline D4RL data (200 epochs)
- Fine-tuning: PPO fine-tune with online environment interaction
- Dev test: 2 epochs only for quick verification, not full training

EXPERIMENT_PLAN.md

@@ -8,25 +8,26 @@
 - All dependencies installed including PyTorch, d4rl, dm-control
 
 ### Initial Testing
-✅ **DPPO Confirmed Working on HoReKa**
+✅ **DPPO Confirmed Working on HoReKa with WandB**
-- Successfully completed dev test (Job ID 3445106)
+- Successfully completed dev test (Job ID 3445117)
-- Pre-training working: 2 epochs, loss reduction 0.2494→0.2010
+- Quick verification: 2 epochs only (not full training), loss reduction 0.2494→0.2010
-- Model checkpoints saved correctly
+- WandB logging working: https://wandb.ai/dominik_roth/gym-hopper-medium-v2-pretrain/runs/rztwqutf
-- Ready for full experiments
+- Model checkpoints and logging fully functional
+- Ready for full 200-epoch production runs
 
 ## Experiments To Run
 
 ### 1. Reproduce Paper Results - Gym Tasks
 
-**Pre-training Phase**:
+**Pre-training Phase** (Train diffusion model on offline D4RL datasets):
-- hopper-medium-v2
+- hopper-medium-v2 → diffusion model trained on offline data (200 epochs)
-- walker2d-medium-v2  
+- walker2d-medium-v2 → diffusion model trained on offline data (200 epochs)
-- halfcheetah-medium-v2
+- halfcheetah-medium-v2 → diffusion model trained on offline data (200 epochs)
 
-**Fine-tuning Phase**:
+**Fine-tuning Phase** (PPO fine-tune diffusion model with online interaction):
-- hopper-v2
+- hopper-v2 → fine-tune pre-trained hopper model with PPO + online env
-- walker2d-v2
+- walker2d-v2 → fine-tune pre-trained walker2d model with PPO + online env
-- halfcheetah-v2
+- halfcheetah-v2 → fine-tune pre-trained halfcheetah model with PPO + online env
 
 **Settings**: Paper hyperparameters, 3 seeds each
 
@@ -92,6 +93,18 @@ No issues with the DPPO repository - installation and setup completed successful
 
 ## Next Steps
 
-1. Run corrected dev test
+### Immediate Tasks (To Verify All Environments Work)
-2. Begin systematic pre-training experiments
+
-3. Document successful runs and results
+1. **Test remaining Gym environments**: 
+   - [ ] walker2d-medium-v2 (2 epochs dev test)
+   - [ ] halfcheetah-medium-v2 (2 epochs dev test)
+   
+2. **Test other environment types**:
+   - [ ] Robomimic: can task (basic test)
+   - [ ] D3IL: avoid_m1 (basic test)
+
+3. **Full production runs** (after confirming all work):
+   - [ ] Full pre-training: hopper, walker2d, halfcheetah (200 epochs each)
+   - [ ] Fine-tuning experiments
+
+**Status**: Only hopper-medium-v2 confirmed working. Need to verify other environments before production runs.