Datasets:
Run pqn/seed1 trained
Browse files- reports/experiments/pqn.md +10 -2
- reports/figures/curves_pqn.png +3 -0
- reports/figures/curves_pqn.svg +0 -0
- reports/summary.md +9 -8
reports/experiments/pqn.md
CHANGED
|
@@ -14,13 +14,15 @@ Metric: success-weighted speed (SWS) and finish rate; intervals are 95 % two-lev
|
|
| 14 |
|
| 15 |
| Run | Status | Phase | Agent steps | Wall hours | Physics | Commit | Evaluated on |
|
| 16 |
|---|---|---|---|---|---|---|---|
|
| 17 |
-
| `pqn/seed1` |
|
| 18 |
| `pqn/seed2` | not started | 6-seed-sensitivity | — | — | v1 | — | — |
|
| 19 |
| `pqn/seed3` | not started | 6-seed-sensitivity | — | — | v1 | — | — |
|
| 20 |
|
| 21 |
## Configuration (logged in each run.json)
|
| 22 |
|
| 23 |
-
|
|
|
|
|
|
|
| 24 |
|
| 25 |
## Results
|
| 26 |
|
|
@@ -32,6 +34,12 @@ _No run of this group has logged its configuration yet (the experiment files are
|
|
| 32 |
|
| 33 |
Rows from other groups are shown for comparison. Track counts: Test (unseen generated): 971 tracks; Validation: 197 tracks; Training sample: 1,000 tracks; Own training tracks: 1,000 tracks; Out-of-range test: 187 tracks; Campaign (hand-made): 5 tracks; Unfiltered test: 200 tracks.
|
| 34 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 35 |
## Learnings
|
| 36 |
|
| 37 |
_No hand-written notes yet (add `reports/notes/pqn.md`)._
|
|
|
|
| 14 |
|
| 15 |
| Run | Status | Phase | Agent steps | Wall hours | Physics | Commit | Evaluated on |
|
| 16 |
|---|---|---|---|---|---|---|---|
|
| 17 |
+
| `pqn/seed1` | trained | 2-single-seed | 1,000,341,504 | 1.10 | v1 | 70a35809b4 | — |
|
| 18 |
| `pqn/seed2` | not started | 6-seed-sensitivity | — | — | v1 | — | — |
|
| 19 |
| `pqn/seed3` | not started | 6-seed-sensitivity | — | — | v1 | — | — |
|
| 20 |
|
| 21 |
## Configuration (logged in each run.json)
|
| 22 |
|
| 23 |
+
| Run | total_steps | num_envs | num_steps | lr | gamma | gae_lambda | ent_coef | clip_coef | hidden | action_mode | weight_decay | loaded_tracks | window_k | blind | time_cost | fall_penalty | random_start_prob | max_steps | hint_weight | scripted_starts | training tracks |
|
| 24 |
+
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
| 25 |
+
| `pqn/seed1` | 1000000000 | 16384 | 32 | 0.00025 | 0.99 | 0.95 | 0.01 | 0.2 | [256, 256] | factored | 0.0 | 4096 | 20 | False | 0.0 | 0.0 | 0.5 | 1500 | 0.0 | False | 48949 |
|
| 26 |
|
| 27 |
## Results
|
| 28 |
|
|
|
|
| 34 |
|
| 35 |
Rows from other groups are shown for comparison. Track counts: Test (unseen generated): 971 tracks; Validation: 197 tracks; Training sample: 1,000 tracks; Own training tracks: 1,000 tracks; Out-of-range test: 187 tracks; Campaign (hand-made): 5 tracks; Unfiltered test: 200 tracks.
|
| 36 |
|
| 37 |
+
### Training curves
|
| 38 |
+
|
| 39 |
+

|
| 40 |
+
|
| 41 |
+
*Training curves of PQN; one line per seed (seeds 1). Return and finish rate are over training episodes (standing and random starts); validation SWS is the evaluator's score on the validation tracks.*
|
| 42 |
+
|
| 43 |
## Learnings
|
| 44 |
|
| 45 |
_No hand-written notes yet (add `reports/notes/pqn.md`)._
|
reports/figures/curves_pqn.png
ADDED
|
Git LFS Details
|
reports/figures/curves_pqn.svg
ADDED
|
|
reports/summary.md
CHANGED
|
@@ -10,11 +10,11 @@ _Generated by `uv run python scripts/07_make_report.py` from the files in `resul
|
|
| 10 |
|---|---|---|---|---|---|---|
|
| 11 |
| baselines | 0 | 0 | 0 | 0 | 15 | 15 |
|
| 12 |
| 1-first-models | 0 | 0 | 0 | 0 | 2 | 2 |
|
| 13 |
-
| 2-single-seed | 22 |
|
| 14 |
| 4-imitation-hybrids | 5 | 0 | 0 | 0 | 0 | 5 |
|
| 15 |
| 5-hyperparameters | 4 | 0 | 0 | 0 | 0 | 4 |
|
| 16 |
| 6-seed-sensitivity | 65 | 0 | 0 | 0 | 0 | 65 |
|
| 17 |
-
| **all** | 96 |
|
| 18 |
|
| 19 |
Statuses: **not started** = in the run matrix, not yet registered; **planned** = registered; code and config pushed before training; **training** = training in progress (or interrupted and resumable); **trained** = trained to its step budget, not yet evaluated; **evaluated** = evaluated; scores available.
|
| 20 |
|
|
@@ -65,7 +65,7 @@ Track counts: Test (unseen generated): 971 tracks; Validation: 197 tracks; Train
|
|
| 65 |
| Blind control | Control | PPO | MLP 2x256 tanh (actor and critic), two 3-way heads | centreline progress | car state + nearest centreline point only (blind) | 48,949 generated tracks | standing + random centreline (p=0.5) | 1 | 1 evaluated, 4 not started | Test 1.010 / 73 %; Val 1.003 / 73 %; Train 1.008 / 73 %; Own train 1.008 / 73 %; OOR 0.605 / 44 %; Campaign 1.443 / 100 %; Unfiltered 1.022 / 74 % |
|
| 66 |
| Generalist (PPO) [pilot, seed 0] | Main method | PPO | MLP 2x256 tanh (actor and critic), two 3-way heads | centreline progress | car state + 20 look-ahead points | 48,949 generated tracks | standing + random centreline (p=0.5) | 0 | 1 evaluated | Test 1.199 / 80 %; Val 1.242 / 83 %; Train 1.234 / 83 %; OOR 0.788 / 52 %; Campaign 1.596 / 100 %; Unfiltered 1.227 / 82 % |
|
| 67 |
| Generalist (PPO) | Main method | PPO | MLP 2x256 tanh (actor and critic), two 3-way heads | centreline progress | car state + 20 look-ahead points | 48,949 generated tracks | standing + random centreline (p=0.5) | 1 | 1 evaluated, 9 not started | Test 1.214 / 81 %; Val 1.236 / 83 %; Train 1.250 / 84 %; Own train 1.250 / 84 %; OOR 0.875 / 58 %; Campaign 1.590 / 100 %; Unfiltered 1.231 / 82 % |
|
| 68 |
-
| PQN | Baseline algorithm |
|
| 69 |
| Trained on 1 track | Ablation | PPO | MLP 2x256 tanh (actor and critic), two 3-way heads | centreline progress | car state + look-ahead points (default) | 1 generated tracks | standing + random centreline | — | 5 not started | — |
|
| 70 |
| Trained on 10 tracks | Ablation | PPO | MLP 2x256 tanh (actor and critic), two 3-way heads | centreline progress | car state + look-ahead points (default) | 10 generated tracks | standing + random centreline | — | 3 not started | — |
|
| 71 |
| Trained on 100 tracks | Ablation | PPO | MLP 2x256 tanh (actor and critic), two 3-way heads | centreline progress | car state + look-ahead points (default) | 100 generated tracks | standing + random centreline | — | 3 not started | — |
|
|
@@ -109,7 +109,7 @@ Track counts: Test (unseen generated): 971 tracks; Validation: 197 tracks; Train
|
|
| 109 |
| [Blind control](experiments/blind-control.md) | Does the agent use the look-ahead (H3)? | 1 evaluated, 4 not started |
|
| 110 |
| [Training-set size](experiments/training-set-size.md) | Is track diversity what generalises (H2)? | 17 not started |
|
| 111 |
| [Specialists](experiments/specialists.md) | How much time does a generalist lose against single-track specialists? | 30 not started |
|
| 112 |
-
| [Baseline algorithm: PQN](experiments/pqn.md) | Does the result depend on PPO? | 1
|
| 113 |
| [Ablations: inputs, actions and reward](experiments/ablations.md) | Which design choices matter (exploratory)? | 21 not started |
|
| 114 |
| [Imitation and hybrids (phase 4)](experiments/imitation-hybrids.md) | What does help from the scripted driver change (exploratory)? | 5 not started |
|
| 115 |
| [Hyperparameter sensitivity (phase 5)](experiments/hyperparameters.md) | How far does the test score move when one setting changes? | 8 not started |
|
|
@@ -132,7 +132,8 @@ Track counts: Test (unseen generated): 971 tracks; Validation: 197 tracks; Train
|
|
| 132 |
|---|---|---|
|
| 133 |
| oor | RandomDriver | 3 |
|
| 134 |
| oor | blind/seed1 | 1 |
|
| 135 |
-
| oor |
|
|
|
|
| 136 |
| oor | generalist/seed1 | 1 |
|
| 137 |
| oor | scripted | 1 |
|
| 138 |
| oor | untrained-seed0 | 1 |
|
|
@@ -148,10 +149,10 @@ Track counts: Test (unseen generated): 971 tracks; Validation: 197 tracks; Train
|
|
| 148 |
| test | RandomDriver | 3 |
|
| 149 |
| test | blind/seed1 | 1 |
|
| 150 |
| test | cem | 2 |
|
| 151 |
-
| test | export:unseen-track-demo |
|
| 152 |
-
| test | generalist/seed0 |
|
| 153 |
| test | generalist/seed1 | 1 |
|
| 154 |
-
| test | scripted |
|
| 155 |
| test | untrained-seed0 | 1 |
|
| 156 |
| test | untrained-seed1 | 1 |
|
| 157 |
| test | untrained-seed2 | 1 |
|
|
|
|
| 10 |
|---|---|---|---|---|---|---|
|
| 11 |
| baselines | 0 | 0 | 0 | 0 | 15 | 15 |
|
| 12 |
| 1-first-models | 0 | 0 | 0 | 0 | 2 | 2 |
|
| 13 |
+
| 2-single-seed | 22 | 0 | 0 | 1 | 2 | 25 |
|
| 14 |
| 4-imitation-hybrids | 5 | 0 | 0 | 0 | 0 | 5 |
|
| 15 |
| 5-hyperparameters | 4 | 0 | 0 | 0 | 0 | 4 |
|
| 16 |
| 6-seed-sensitivity | 65 | 0 | 0 | 0 | 0 | 65 |
|
| 17 |
+
| **all** | 96 | 0 | 0 | 1 | 19 | 116 |
|
| 18 |
|
| 19 |
Statuses: **not started** = in the run matrix, not yet registered; **planned** = registered; code and config pushed before training; **training** = training in progress (or interrupted and resumable); **trained** = trained to its step budget, not yet evaluated; **evaluated** = evaluated; scores available.
|
| 20 |
|
|
|
|
| 65 |
| Blind control | Control | PPO | MLP 2x256 tanh (actor and critic), two 3-way heads | centreline progress | car state + nearest centreline point only (blind) | 48,949 generated tracks | standing + random centreline (p=0.5) | 1 | 1 evaluated, 4 not started | Test 1.010 / 73 %; Val 1.003 / 73 %; Train 1.008 / 73 %; Own train 1.008 / 73 %; OOR 0.605 / 44 %; Campaign 1.443 / 100 %; Unfiltered 1.022 / 74 % |
|
| 66 |
| Generalist (PPO) [pilot, seed 0] | Main method | PPO | MLP 2x256 tanh (actor and critic), two 3-way heads | centreline progress | car state + 20 look-ahead points | 48,949 generated tracks | standing + random centreline (p=0.5) | 0 | 1 evaluated | Test 1.199 / 80 %; Val 1.242 / 83 %; Train 1.234 / 83 %; OOR 0.788 / 52 %; Campaign 1.596 / 100 %; Unfiltered 1.227 / 82 % |
|
| 67 |
| Generalist (PPO) | Main method | PPO | MLP 2x256 tanh (actor and critic), two 3-way heads | centreline progress | car state + 20 look-ahead points | 48,949 generated tracks | standing + random centreline (p=0.5) | 1 | 1 evaluated, 9 not started | Test 1.214 / 81 %; Val 1.236 / 83 %; Train 1.250 / 84 %; Own train 1.250 / 84 %; OOR 0.875 / 58 %; Campaign 1.590 / 100 %; Unfiltered 1.231 / 82 % |
|
| 68 |
+
| PQN | Baseline algorithm | PQN | Q-network 2x256, LayerNorm, one 9-way head | centreline progress | car state + 20 look-ahead points | 48,949 generated tracks | standing + random centreline (p=0.5) | 1 | 1 trained, 2 not started | — |
|
| 69 |
| Trained on 1 track | Ablation | PPO | MLP 2x256 tanh (actor and critic), two 3-way heads | centreline progress | car state + look-ahead points (default) | 1 generated tracks | standing + random centreline | — | 5 not started | — |
|
| 70 |
| Trained on 10 tracks | Ablation | PPO | MLP 2x256 tanh (actor and critic), two 3-way heads | centreline progress | car state + look-ahead points (default) | 10 generated tracks | standing + random centreline | — | 3 not started | — |
|
| 71 |
| Trained on 100 tracks | Ablation | PPO | MLP 2x256 tanh (actor and critic), two 3-way heads | centreline progress | car state + look-ahead points (default) | 100 generated tracks | standing + random centreline | — | 3 not started | — |
|
|
|
|
| 109 |
| [Blind control](experiments/blind-control.md) | Does the agent use the look-ahead (H3)? | 1 evaluated, 4 not started |
|
| 110 |
| [Training-set size](experiments/training-set-size.md) | Is track diversity what generalises (H2)? | 17 not started |
|
| 111 |
| [Specialists](experiments/specialists.md) | How much time does a generalist lose against single-track specialists? | 30 not started |
|
| 112 |
+
| [Baseline algorithm: PQN](experiments/pqn.md) | Does the result depend on PPO? | 1 trained, 2 not started |
|
| 113 |
| [Ablations: inputs, actions and reward](experiments/ablations.md) | Which design choices matter (exploratory)? | 21 not started |
|
| 114 |
| [Imitation and hybrids (phase 4)](experiments/imitation-hybrids.md) | What does help from the scripted driver change (exploratory)? | 5 not started |
|
| 115 |
| [Hyperparameter sensitivity (phase 5)](experiments/hyperparameters.md) | How far does the test score move when one setting changes? | 8 not started |
|
|
|
|
| 132 |
|---|---|---|
|
| 133 |
| oor | RandomDriver | 3 |
|
| 134 |
| oor | blind/seed1 | 1 |
|
| 135 |
+
| oor | export:unseen-track-demo | 1 |
|
| 136 |
+
| oor | generalist/seed0 | 2 |
|
| 137 |
| oor | generalist/seed1 | 1 |
|
| 138 |
| oor | scripted | 1 |
|
| 139 |
| oor | untrained-seed0 | 1 |
|
|
|
|
| 149 |
| test | RandomDriver | 3 |
|
| 150 |
| test | blind/seed1 | 1 |
|
| 151 |
| test | cem | 2 |
|
| 152 |
+
| test | export:unseen-track-demo | 4 |
|
| 153 |
+
| test | generalist/seed0 | 2 |
|
| 154 |
| test | generalist/seed1 | 1 |
|
| 155 |
+
| test | scripted | 3 |
|
| 156 |
| test | untrained-seed0 | 1 |
|
| 157 |
| test | untrained-seed1 | 1 |
|
| 158 |
| test | untrained-seed2 | 1 |
|