Publish DungeonEscape step 70,800 bundle
Browse filesReplace the DungeonEscape safetensors bundle with training step 70,800, trained using simple + medium + expert data. KV64 closed-loop validation over seeds 100..108: 187/324 agent successes, 101/108 cooperative-group successes, and 324/324 natural terminations. Legacy ONNX artifacts are retained and explicitly labeled in the model card.
- README.md +34 -24
- dungeon-escape/config.json +3 -2
- dungeon-escape/metrics.json +11 -0
- dungeon-escape/model.safetensors +1 -1
README.md
CHANGED
|
@@ -76,11 +76,10 @@ of history, plus the ONNX variants used for Unity deployment.
|
|
| 76 |
| `crawler` | `Tuple(Box(126), Box(32))` | `Box(20, [-1, 1])` |
|
| 77 |
| `pushblock` | `Tuple(Box(105), Box(105))` | `Discrete(7)` |
|
| 78 |
|
| 79 |
-
The result column
|
| 80 |
-
|
| 81 |
-
|
| 82 |
-
|
| 83 |
-
length below.
|
| 84 |
|
| 85 |
**Unity behavior names:** `DungeonEscape`, `SoccerTwos`. HF bundle subfolders are
|
| 86 |
lowercase and case-sensitive.
|
|
@@ -90,8 +89,9 @@ Each bundle wraps its ego spaces in a named container β
|
|
| 90 |
scene. `Batch` is the fixed ONNX batch: 36 for DungeonEscape (12 arenas Γ 3
|
| 91 |
agents), 16 for SoccerTwos (8 fields Γ 2 controlled agents).
|
| 92 |
|
| 93 |
-
Training
|
| 94 |
-
`medium-v0`
|
|
|
|
| 95 |
|
| 96 |
Metrics differ by scene type and are not comparable across rows. DungeonEscape is
|
| 97 |
cooperative: **agent success** is the fraction of ego agents that finish with a
|
|
@@ -124,8 +124,8 @@ Results at 0.5Γ, 1Γ, and 2Γ the training length:
|
|
| 124 |
|
| 125 |
| Bundle | Runtime | 16 (0.5Γ) | 32 (1Γ) | 64 (2Γ) |
|
| 126 |
|---|---|---:|---:|---:|
|
| 127 |
-
| `dungeon-escape` | PyTorch (cached) | 57.41% | **63.89%** | 61.11% |
|
| 128 |
-
| `dungeon-escape` | ONNX (windowed) | 56.48% | **60.19%** | not published β‘ |
|
| 129 |
| `soccer-twos` | PyTorch (cached) | β β | **41.25%** | β β |
|
| 130 |
| `soccer-twos` | ONNX (windowed) | 40.00% | **40.00%** | 32.50% |
|
| 131 |
| `pushblock` | PyTorch (cached) | 159/160 | 159/160 | 159/160 |
|
|
@@ -146,22 +146,32 @@ axis is not unmeasured for this bundle β only the cached runtime at 0.5Γ and
|
|
| 146 |
<details>
|
| 147 |
<summary><strong>Per-environment protocol and full results</strong></summary>
|
| 148 |
|
| 149 |
-
**DungeonEscape** β
|
| 150 |
-
|
| 151 |
-
|
| 152 |
-
expert tier was not used.
|
| 153 |
|
| 154 |
-
|
| 155 |
-
|
| 156 |
-
|
| 157 |
-
|
| 158 |
-
|
|
| 159 |
-
|
|
| 160 |
-
|
|
| 161 |
-
|
| 162 |
-
|
| 163 |
-
|
| 164 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 165 |
|
| 166 |
**SoccerTwos** β matching release-23 model-removed build, with the stock
|
| 167 |
`SoccerTwos.onnx` from the same release as the opponent. Every row covers
|
|
|
|
| 76 |
| `crawler` | `Tuple(Box(126), Box(32))` | `Box(20, [-1, 1])` |
|
| 77 |
| `pushblock` | `Tuple(Box(105), Box(105))` | `Discrete(7)` |
|
| 78 |
|
| 79 |
+
The result column reports the published **ONNX at context length 32**. For
|
| 80 |
+
DungeonEscape, those ONNX files are retained legacy artifacts; the current
|
| 81 |
+
step-70,800 safetensors bundle and its cached-runtime results are documented
|
| 82 |
+
below. SoccerTwos ONNX remains derived from its sibling bundle.
|
|
|
|
| 83 |
|
| 84 |
**Unity behavior names:** `DungeonEscape`, `SoccerTwos`. HF bundle subfolders are
|
| 85 |
lowercase and case-sensitive.
|
|
|
|
| 89 |
scene. `Batch` is the fixed ONNX batch: 36 for DungeonEscape (12 arenas Γ 3
|
| 90 |
agents), 16 for SoccerTwos (8 fields Γ 2 controlled agents).
|
| 91 |
|
| 92 |
+
Training-data provenance is bundle-specific. The current DungeonEscape bundle
|
| 93 |
+
uses `simple-v0` + `medium-v0` + `expert-v0`; the SoccerTwos bundle remains
|
| 94 |
+
trained without expert trajectories.
|
| 95 |
|
| 96 |
Metrics differ by scene type and are not comparable across rows. DungeonEscape is
|
| 97 |
cooperative: **agent success** is the fraction of ego agents that finish with a
|
|
|
|
| 124 |
|
| 125 |
| Bundle | Runtime | 16 (0.5Γ) | 32 (1Γ) | 64 (2Γ) |
|
| 126 |
|---|---|---:|---:|---:|
|
| 127 |
+
| `dungeon-escape` (legacy checkpoint) | PyTorch (cached) | 57.41% | **63.89%** | 61.11% |
|
| 128 |
+
| `dungeon-escape` (legacy ONNX) | ONNX (windowed) | 56.48% | **60.19%** | not published β‘ |
|
| 129 |
| `soccer-twos` | PyTorch (cached) | β β | **41.25%** | β β |
|
| 130 |
| `soccer-twos` | ONNX (windowed) | 40.00% | **40.00%** | 32.50% |
|
| 131 |
| `pushblock` | PyTorch (cached) | 159/160 | 159/160 | 159/160 |
|
|
|
|
| 146 |
<details>
|
| 147 |
<summary><strong>Per-environment protocol and full results</strong></summary>
|
| 148 |
|
| 149 |
+
**DungeonEscape** β current PyTorch bundle, training step `70,800`,
|
| 150 |
+
matching release-23 model-removed build. This policy was trained with all three
|
| 151 |
+
data tiers: `simple-v0` + `medium-v0` + `expert-v0`.
|
|
|
|
| 152 |
|
| 153 |
+
Closed-loop cached-runtime evaluation used CUDA, 36 ego agents per launch
|
| 154 |
+
(12 cooperative groups of three), and Unity seeds `100..108`. Unity truncations
|
| 155 |
+
are reported as timeouts; episode length is not used to invalidate a seed.
|
| 156 |
+
|
| 157 |
+
| Runtime KV | Seeds | Agent success | Cooperative-group success | Natural termination |
|
| 158 |
+
|---:|---:|---:|---:|---:|
|
| 159 |
+
| 32 | `100..108` | 186/324 (57.41%) | **101/108 (93.52%)** | 322/324 |
|
| 160 |
+
| **64** | **`100..108`** | **187/324 (57.72%)** | **101/108 (93.52%)** | **324/324** |
|
| 161 |
+
| 128 (preliminary) | `100..102` | 63/108 (58.33%) | 35/36 (97.22%) | 107/108 |
|
| 162 |
+
|
| 163 |
+
KV64 is the recommended cached-runtime setting: it matches KV32 group success
|
| 164 |
+
while eliminating the two observed timeouts. The bundle's training context is
|
| 165 |
+
32; KV64 and KV128 are inference-time retention settings.
|
| 166 |
+
|
| 167 |
+
The ONNX files currently in this folder are retained legacy deployment artifacts
|
| 168 |
+
from the preceding DungeonEscape checkpoint. Their published validation below
|
| 169 |
+
does **not** describe the step-70,800 safetensors bundle.
|
| 170 |
+
|
| 171 |
+
| Legacy ONNX context | Agent success | Cooperative-group success |
|
| 172 |
+
|---:|---:|---:|
|
| 173 |
+
| 16 | 61/108 (56.48%) | 33/36 (91.67%) |
|
| 174 |
+
| 32 | 65/108 (60.19%) | 34/36 (94.44%) |
|
| 175 |
|
| 176 |
**SoccerTwos** β matching release-23 model-removed build, with the stock
|
| 177 |
`SoccerTwos.onnx` from the same release as the opponent. Every row covers
|
dungeon-escape/config.json
CHANGED
|
@@ -1,6 +1,7 @@
|
|
| 1 |
{
|
| 2 |
"bundle_format_version": 2,
|
| 3 |
-
"
|
|
|
|
| 4 |
"model_config": {
|
| 5 |
"network_name": "Llama",
|
| 6 |
"num_layers": 4,
|
|
@@ -9,7 +10,7 @@
|
|
| 9 |
"dropout": 0.05,
|
| 10 |
"use_eos": true,
|
| 11 |
"use_bos_action_gate": true,
|
| 12 |
-
"rope_theta":
|
| 13 |
"intermediate_size": 1024,
|
| 14 |
"max_position_embeddings": 256,
|
| 15 |
"context_length": 32
|
|
|
|
| 1 |
{
|
| 2 |
"bundle_format_version": 2,
|
| 3 |
+
"package": "causal-gpt-rl",
|
| 4 |
+
"package_version": "0.18.0",
|
| 5 |
"model_config": {
|
| 6 |
"network_name": "Llama",
|
| 7 |
"num_layers": 4,
|
|
|
|
| 10 |
"dropout": 0.05,
|
| 11 |
"use_eos": true,
|
| 12 |
"use_bos_action_gate": true,
|
| 13 |
+
"rope_theta": 10000.0,
|
| 14 |
"intermediate_size": 1024,
|
| 15 |
"max_position_embeddings": 256,
|
| 16 |
"context_length": 32
|
dungeon-escape/metrics.json
ADDED
|
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"step": 70800,
|
| 3 |
+
"reasons": [
|
| 4 |
+
"best"
|
| 5 |
+
],
|
| 6 |
+
"metrics": {
|
| 7 |
+
"eval_offline/action_nll": 1.7410476793746104,
|
| 8 |
+
"eval_offline/checkpoint_score": 0.46170878410339355
|
| 9 |
+
},
|
| 10 |
+
"metrics_step": 70800
|
| 11 |
+
}
|
dungeon-escape/model.safetensors
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
size 17194284
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d1e47d1139fd57f3af53d525586fe22a6924f99155e3825532814dae6b82ee12
|
| 3 |
size 17194284
|