Codex commited on
Commit ·
b4232ba
1
Parent(s): 79f8770
Prune retired task and training artifacts; retain evaluated checkpoints and repair release references
Browse filesThis view is limited to 50 files because it contains too many changes. See raw diff
- README.md +4 -0
- archive/task-transfer/manifest.json +4 -34
- archive/task-transfer/openvla_bridge_cross_mixed_all_024_5env200ep_deadly1000ep_exp1/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt +0 -3
- archive/task-transfer/openvla_bridge_cross_mixed_all_024_5env200ep_deadly1000ep_exp1/checkpoints/steps_15000_state/optimizer.bin +0 -3
- archive/task-transfer/openvla_bridge_cross_mixed_all_024_5env200ep_deadly1000ep_exp1/checkpoints/steps_15000_state/random_states_0.pkl +0 -3
- archive/task-transfer/openvla_bridge_cross_mixed_all_024_5env200ep_deadly1000ep_exp1/wandb/wandb/run-20260923_064644-r8358sud/run-r8358sud.wandb +0 -0
- archive/task-transfer/openvla_bridge_cross_mixed_all_024_5env200ep_deadly1000ep_exp1/wandb/wandb/run-20260923_065340-67ykhdhp/run-67ykhdhp.wandb +0 -3
- archive/task-transfer/openvla_bridge_cross_mixed_all_024_5env200ep_deadly1000ep_exp1/wandb/wandb/run-20260923_105406-7ujo2tgu/run-7ujo2tgu.wandb +0 -3
- archive/task-transfer/openvla_bridge_demon_attack_latency_mixed_024_exp1/steps_5000_state/optimizer.bin +0 -3
- archive/task-transfer/openvla_bridge_demon_attack_latency_mixed_024_exp1/steps_5000_state/random_states_0.pkl +0 -3
- experiments/latency-transfer/archive/deadly-corridor/manifest.json +0 -15
- experiments/latency-transfer/archive/flappy/manifest.json +0 -50
- experiments/observation-stride/deadly-corridor/manifest.json +0 -63
- latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_0_200ep/filter_report.json +0 -19
- latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_0_200ep/metadata.json +0 -289
- latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_0_200ep/train.parquet +0 -3
- latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_0_200ep/val.parquet +0 -3
- latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_2_200ep/filter_report.json +0 -19
- latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_2_200ep/metadata.json +0 -289
- latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_2_200ep/train.parquet +0 -3
- latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_2_200ep/val.parquet +0 -3
- latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_4_200ep/filter_report.json +0 -19
- latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_4_200ep/metadata.json +0 -289
- latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_4_200ep/train.parquet +0 -3
- latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_4_200ep/val.parquet +0 -3
- latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_6_200ep/filter_report.json +0 -19
- latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_6_200ep/metadata.json +0 -289
- latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_6_200ep/train.parquet +0 -3
- latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_6_200ep/val.parquet +0 -3
- latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_8_200ep/filter_report.json +0 -19
- latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_8_200ep/metadata.json +0 -289
- latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_8_200ep/train.parquet +0 -3
- latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_8_200ep/val.parquet +0 -3
- latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/README.md +0 -1
- latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/lerobot/data/chunk-000/episode_000000.parquet +0 -3
- latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/lerobot/data/chunk-000/episode_000001.parquet +0 -3
- latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/lerobot/data/chunk-000/episode_000002.parquet +0 -3
- latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/lerobot/data/chunk-000/episode_000003.parquet +0 -3
- latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/lerobot/data/chunk-000/episode_000004.parquet +0 -3
- latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/lerobot/data/chunk-000/episode_000005.parquet +0 -3
- latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/lerobot/data/chunk-000/episode_000006.parquet +0 -3
- latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/lerobot/data/chunk-000/episode_000007.parquet +0 -3
- latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/lerobot/data/chunk-000/episode_000008.parquet +0 -3
- latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/lerobot/data/chunk-000/episode_000009.parquet +0 -3
- latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/lerobot/data/chunk-000/episode_000010.parquet +0 -3
- latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/lerobot/data/chunk-000/episode_000011.parquet +0 -3
- latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/lerobot/data/chunk-000/episode_000012.parquet +0 -3
- latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/lerobot/data/chunk-000/episode_000013.parquet +0 -3
- latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/lerobot/data/chunk-000/episode_000014.parquet +0 -3
- latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/lerobot/data/chunk-000/episode_000015.parquet +0 -3
README.md
CHANGED
|
@@ -3909,3 +3909,7 @@ Training data, episode-level evaluations and paper experiment manifests.
|
|
| 3909 |
| profiles | [latency-sensitive-bench/profiles](https://huggingface.co/datasets/latency-sensitive-bench/profiles) |
|
| 3910 |
|
| 3911 |
[Benchmark release inventory](reports/benchmark-release/README.md)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3909 |
| profiles | [latency-sensitive-bench/profiles](https://huggingface.co/datasets/latency-sensitive-bench/profiles) |
|
| 3910 |
|
| 3911 |
[Benchmark release inventory](reports/benchmark-release/README.md)
|
| 3912 |
+
|
| 3913 |
+
## Artifact retention
|
| 3914 |
+
|
| 3915 |
+
HAIC and Extreme Parkour releases are retired. Model bundles retain the published evaluation checkpoint, or the latest checkpoint when no evaluation selection exists. Optimizer and trainer recovery state are not release assets. Identical dataset copies use the canonical task paths. Original configurations and experiment evidence remain source records.
|
archive/task-transfer/manifest.json
CHANGED
|
@@ -275,10 +275,10 @@
|
|
| 275 |
"path": "archive/task-transfer/openvla_bridge_cross_mixed_all_024_5env200ep_deadly1000ep_exp1/.gitattributes"
|
| 276 |
},
|
| 277 |
{
|
| 278 |
-
"repo": "
|
| 279 |
-
"repo_type": "
|
| 280 |
-
"revision": "
|
| 281 |
-
"path": "
|
| 282 |
},
|
| 283 |
{
|
| 284 |
"repo": "latency-sensitive-bench/benchmark-models",
|
|
@@ -286,18 +286,6 @@
|
|
| 286 |
"revision": "671e2a7451fda7f7dba5606c6d11183ebd804b6d",
|
| 287 |
"path": "archive/task-transfer/openvla_bridge_cross_mixed_all_024_5env200ep_deadly1000ep_exp1/checkpoints/steps_15000_model.safetensors"
|
| 288 |
},
|
| 289 |
-
{
|
| 290 |
-
"repo": "latency-sensitive-bench/benchmark-datasets",
|
| 291 |
-
"repo_type": "dataset",
|
| 292 |
-
"revision": "d6f0ad0995f6e4b095f315920e8de18e04297048",
|
| 293 |
-
"path": "archive/task-transfer/openvla_bridge_cross_mixed_all_024_5env200ep_deadly1000ep_exp1/checkpoints/steps_15000_state/optimizer.bin"
|
| 294 |
-
},
|
| 295 |
-
{
|
| 296 |
-
"repo": "latency-sensitive-bench/benchmark-datasets",
|
| 297 |
-
"repo_type": "dataset",
|
| 298 |
-
"revision": "d6f0ad0995f6e4b095f315920e8de18e04297048",
|
| 299 |
-
"path": "archive/task-transfer/openvla_bridge_cross_mixed_all_024_5env200ep_deadly1000ep_exp1/checkpoints/steps_15000_state/random_states_0.pkl"
|
| 300 |
-
},
|
| 301 |
{
|
| 302 |
"repo": "latency-sensitive-bench/benchmark-models",
|
| 303 |
"repo_type": "model",
|
|
@@ -400,12 +388,6 @@
|
|
| 400 |
"revision": "d6f0ad0995f6e4b095f315920e8de18e04297048",
|
| 401 |
"path": "archive/task-transfer/openvla_bridge_cross_mixed_all_024_5env200ep_deadly1000ep_exp1/wandb/wandb/run-20260923_064644-r8358sud/logs/debug.log"
|
| 402 |
},
|
| 403 |
-
{
|
| 404 |
-
"repo": "latency-sensitive-bench/benchmark-datasets",
|
| 405 |
-
"repo_type": "dataset",
|
| 406 |
-
"revision": "d6f0ad0995f6e4b095f315920e8de18e04297048",
|
| 407 |
-
"path": "archive/task-transfer/openvla_bridge_cross_mixed_all_024_5env200ep_deadly1000ep_exp1/wandb/wandb/run-20260923_064644-r8358sud/run-r8358sud.wandb"
|
| 408 |
-
},
|
| 409 |
{
|
| 410 |
"repo": "latency-sensitive-bench/benchmark-datasets",
|
| 411 |
"repo_type": "dataset",
|
|
@@ -454,12 +436,6 @@
|
|
| 454 |
"revision": "d6f0ad0995f6e4b095f315920e8de18e04297048",
|
| 455 |
"path": "archive/task-transfer/openvla_bridge_cross_mixed_all_024_5env200ep_deadly1000ep_exp1/wandb/wandb/run-20260923_065340-67ykhdhp/logs/debug.log"
|
| 456 |
},
|
| 457 |
-
{
|
| 458 |
-
"repo": "latency-sensitive-bench/benchmark-datasets",
|
| 459 |
-
"repo_type": "dataset",
|
| 460 |
-
"revision": "d6f0ad0995f6e4b095f315920e8de18e04297048",
|
| 461 |
-
"path": "archive/task-transfer/openvla_bridge_cross_mixed_all_024_5env200ep_deadly1000ep_exp1/wandb/wandb/run-20260923_065340-67ykhdhp/run-67ykhdhp.wandb"
|
| 462 |
-
},
|
| 463 |
{
|
| 464 |
"repo": "latency-sensitive-bench/benchmark-datasets",
|
| 465 |
"repo_type": "dataset",
|
|
@@ -507,12 +483,6 @@
|
|
| 507 |
"repo_type": "dataset",
|
| 508 |
"revision": "d6f0ad0995f6e4b095f315920e8de18e04297048",
|
| 509 |
"path": "archive/task-transfer/openvla_bridge_cross_mixed_all_024_5env200ep_deadly1000ep_exp1/wandb/wandb/run-20260923_105406-7ujo2tgu/logs/debug.log"
|
| 510 |
-
},
|
| 511 |
-
{
|
| 512 |
-
"repo": "latency-sensitive-bench/benchmark-datasets",
|
| 513 |
-
"repo_type": "dataset",
|
| 514 |
-
"revision": "d6f0ad0995f6e4b095f315920e8de18e04297048",
|
| 515 |
-
"path": "archive/task-transfer/openvla_bridge_cross_mixed_all_024_5env200ep_deadly1000ep_exp1/wandb/wandb/run-20260923_105406-7ujo2tgu/run-7ujo2tgu.wandb"
|
| 516 |
}
|
| 517 |
],
|
| 518 |
"training_data": [
|
|
|
|
| 275 |
"path": "archive/task-transfer/openvla_bridge_cross_mixed_all_024_5env200ep_deadly1000ep_exp1/.gitattributes"
|
| 276 |
},
|
| 277 |
{
|
| 278 |
+
"repo": "StarVLA/Qwen3VL-OFT-Bridge-RT-1",
|
| 279 |
+
"repo_type": "model",
|
| 280 |
+
"revision": "1c814bc494cd95c844535fffed2a890838651dd6",
|
| 281 |
+
"path": "checkpoints/steps_5000_pytorch_model.pt"
|
| 282 |
},
|
| 283 |
{
|
| 284 |
"repo": "latency-sensitive-bench/benchmark-models",
|
|
|
|
| 286 |
"revision": "671e2a7451fda7f7dba5606c6d11183ebd804b6d",
|
| 287 |
"path": "archive/task-transfer/openvla_bridge_cross_mixed_all_024_5env200ep_deadly1000ep_exp1/checkpoints/steps_15000_model.safetensors"
|
| 288 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 289 |
{
|
| 290 |
"repo": "latency-sensitive-bench/benchmark-models",
|
| 291 |
"repo_type": "model",
|
|
|
|
| 388 |
"revision": "d6f0ad0995f6e4b095f315920e8de18e04297048",
|
| 389 |
"path": "archive/task-transfer/openvla_bridge_cross_mixed_all_024_5env200ep_deadly1000ep_exp1/wandb/wandb/run-20260923_064644-r8358sud/logs/debug.log"
|
| 390 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 391 |
{
|
| 392 |
"repo": "latency-sensitive-bench/benchmark-datasets",
|
| 393 |
"repo_type": "dataset",
|
|
|
|
| 436 |
"revision": "d6f0ad0995f6e4b095f315920e8de18e04297048",
|
| 437 |
"path": "archive/task-transfer/openvla_bridge_cross_mixed_all_024_5env200ep_deadly1000ep_exp1/wandb/wandb/run-20260923_065340-67ykhdhp/logs/debug.log"
|
| 438 |
},
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 439 |
{
|
| 440 |
"repo": "latency-sensitive-bench/benchmark-datasets",
|
| 441 |
"repo_type": "dataset",
|
|
|
|
| 483 |
"repo_type": "dataset",
|
| 484 |
"revision": "d6f0ad0995f6e4b095f315920e8de18e04297048",
|
| 485 |
"path": "archive/task-transfer/openvla_bridge_cross_mixed_all_024_5env200ep_deadly1000ep_exp1/wandb/wandb/run-20260923_105406-7ujo2tgu/logs/debug.log"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 486 |
}
|
| 487 |
],
|
| 488 |
"training_data": [
|
archive/task-transfer/openvla_bridge_cross_mixed_all_024_5env200ep_deadly1000ep_exp1/checkpoints/_initialization/StarVLA_Qwen3VL-OFT-Bridge-RT-1/checkpoints/steps_5000_pytorch_model.pt
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:371cb744227687bb99bcad7f9ff2250cf06da75631359ad3eba4c6bc52570607
|
| 3 |
-
size 9785060316
|
|
|
|
|
|
|
|
|
|
|
|
archive/task-transfer/openvla_bridge_cross_mixed_all_024_5env200ep_deadly1000ep_exp1/checkpoints/steps_15000_state/optimizer.bin
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:e8b2dc7f1a406b093db1b4a03d3ae0dc2d77415802548498921dd5b5668fe8ed
|
| 3 |
-
size 18276885098
|
|
|
|
|
|
|
|
|
|
|
|
archive/task-transfer/openvla_bridge_cross_mixed_all_024_5env200ep_deadly1000ep_exp1/checkpoints/steps_15000_state/random_states_0.pkl
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:8f041f3f444886eb524567d2ff1df968b6f86a749c7ad6ad18b8d1478843cfb4
|
| 3 |
-
size 14821
|
|
|
|
|
|
|
|
|
|
|
|
archive/task-transfer/openvla_bridge_cross_mixed_all_024_5env200ep_deadly1000ep_exp1/wandb/wandb/run-20260923_064644-r8358sud/run-r8358sud.wandb
DELETED
|
Binary file (616 Bytes)
|
|
|
archive/task-transfer/openvla_bridge_cross_mixed_all_024_5env200ep_deadly1000ep_exp1/wandb/wandb/run-20260923_065340-67ykhdhp/run-67ykhdhp.wandb
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:51c7b17f1b22d417c4f38fbee97d787f97cd9ac52981b0a09dfb0b58552c6819
|
| 3 |
-
size 5928658
|
|
|
|
|
|
|
|
|
|
|
|
archive/task-transfer/openvla_bridge_cross_mixed_all_024_5env200ep_deadly1000ep_exp1/wandb/wandb/run-20260923_105406-7ujo2tgu/run-7ujo2tgu.wandb
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:0c8e7caa217b97d88dfb6c0cc1138eae16c3183996e52dd464d3ed3df20439e5
|
| 3 |
-
size 102100192
|
|
|
|
|
|
|
|
|
|
|
|
archive/task-transfer/openvla_bridge_demon_attack_latency_mixed_024_exp1/steps_5000_state/optimizer.bin
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:1f232fb50fdd4dfe48e1ae8eed8aa5f937fef45da60d646493e0bf06a64ee6f6
|
| 3 |
-
size 18276885098
|
|
|
|
|
|
|
|
|
|
|
|
archive/task-transfer/openvla_bridge_demon_attack_latency_mixed_024_exp1/steps_5000_state/random_states_0.pkl
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:608b90a865655c4bd9bf57be0147891841890389d1b5b4d14494f4e90353f43e
|
| 3 |
-
size 14821
|
|
|
|
|
|
|
|
|
|
|
|
experiments/latency-transfer/archive/deadly-corridor/manifest.json
CHANGED
|
@@ -10,11 +10,6 @@
|
|
| 10 |
"path": "historical/openvla_deadly_corridor_fix_latency_0"
|
| 11 |
},
|
| 12 |
"checkpoint": [
|
| 13 |
-
{
|
| 14 |
-
"repo": "latency-sensitive-bench/benchmark-models",
|
| 15 |
-
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
| 16 |
-
"path": "zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0/checkpoints/best_model.safetensors"
|
| 17 |
-
},
|
| 18 |
{
|
| 19 |
"repo": "latency-sensitive-bench/benchmark-models",
|
| 20 |
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
|
@@ -103,11 +98,6 @@
|
|
| 103 |
"path": "historical/openvla_deadly_corridor_fix_latency_0_full"
|
| 104 |
},
|
| 105 |
"checkpoint": [
|
| 106 |
-
{
|
| 107 |
-
"repo": "latency-sensitive-bench/benchmark-models",
|
| 108 |
-
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
| 109 |
-
"path": "zero-latency/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_0_full/checkpoints/best_model.safetensors"
|
| 110 |
-
},
|
| 111 |
{
|
| 112 |
"repo": "latency-sensitive-bench/benchmark-models",
|
| 113 |
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
|
@@ -195,11 +185,6 @@
|
|
| 195 |
"path": "historical/openvla_deadly_corridor_fix_latency_2"
|
| 196 |
},
|
| 197 |
"checkpoint": [
|
| 198 |
-
{
|
| 199 |
-
"repo": "latency-sensitive-bench/benchmark-models",
|
| 200 |
-
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
| 201 |
-
"path": "latency-aware/deadly-corridor/vla/starvla-qwenoft-h1/openvla_deadly_corridor_fix_latency_2/checkpoints/best_model.safetensors"
|
| 202 |
-
},
|
| 203 |
{
|
| 204 |
"repo": "latency-sensitive-bench/benchmark-models",
|
| 205 |
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
|
|
|
| 10 |
"path": "historical/openvla_deadly_corridor_fix_latency_0"
|
| 11 |
},
|
| 12 |
"checkpoint": [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 13 |
{
|
| 14 |
"repo": "latency-sensitive-bench/benchmark-models",
|
| 15 |
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
|
|
|
| 98 |
"path": "historical/openvla_deadly_corridor_fix_latency_0_full"
|
| 99 |
},
|
| 100 |
"checkpoint": [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 101 |
{
|
| 102 |
"repo": "latency-sensitive-bench/benchmark-models",
|
| 103 |
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
|
|
|
| 185 |
"path": "historical/openvla_deadly_corridor_fix_latency_2"
|
| 186 |
},
|
| 187 |
"checkpoint": [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 188 |
{
|
| 189 |
"repo": "latency-sensitive-bench/benchmark-models",
|
| 190 |
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
experiments/latency-transfer/archive/flappy/manifest.json
CHANGED
|
@@ -10,11 +10,6 @@
|
|
| 10 |
"path": "historical/openvla_flappy_fix_latency_0"
|
| 11 |
},
|
| 12 |
"checkpoint": [
|
| 13 |
-
{
|
| 14 |
-
"repo": "latency-sensitive-bench/benchmark-models",
|
| 15 |
-
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
| 16 |
-
"path": "zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0/checkpoints/steps_5000_model.safetensors"
|
| 17 |
-
},
|
| 18 |
{
|
| 19 |
"repo": "latency-sensitive-bench/benchmark-models",
|
| 20 |
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
|
@@ -92,11 +87,6 @@
|
|
| 92 |
"path": "historical/openvla_flappy_fix_latency_0_small"
|
| 93 |
},
|
| 94 |
"checkpoint": [
|
| 95 |
-
{
|
| 96 |
-
"repo": "latency-sensitive-bench/benchmark-models",
|
| 97 |
-
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
| 98 |
-
"path": "zero-latency/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_0_small/checkpoints/best_model.safetensors"
|
| 99 |
-
},
|
| 100 |
{
|
| 101 |
"repo": "latency-sensitive-bench/benchmark-models",
|
| 102 |
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
|
@@ -185,11 +175,6 @@
|
|
| 185 |
"path": "historical/openvla_flappy_fix_latency_1_small"
|
| 186 |
},
|
| 187 |
"checkpoint": [
|
| 188 |
-
{
|
| 189 |
-
"repo": "latency-sensitive-bench/benchmark-models",
|
| 190 |
-
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
| 191 |
-
"path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small/checkpoints/best_model.safetensors"
|
| 192 |
-
},
|
| 193 |
{
|
| 194 |
"repo": "latency-sensitive-bench/benchmark-models",
|
| 195 |
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
|
@@ -278,11 +263,6 @@
|
|
| 278 |
"path": "historical/openvla_flappy_fix_latency_1_small_tmp"
|
| 279 |
},
|
| 280 |
"checkpoint": [
|
| 281 |
-
{
|
| 282 |
-
"repo": "latency-sensitive-bench/benchmark-models",
|
| 283 |
-
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
| 284 |
-
"path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_1_small_tmp/checkpoints/best_model.safetensors"
|
| 285 |
-
},
|
| 286 |
{
|
| 287 |
"repo": "latency-sensitive-bench/benchmark-models",
|
| 288 |
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
|
@@ -361,11 +341,6 @@
|
|
| 361 |
"path": "historical/openvla_flappy_fix_latency_2"
|
| 362 |
},
|
| 363 |
"checkpoint": [
|
| 364 |
-
{
|
| 365 |
-
"repo": "latency-sensitive-bench/benchmark-models",
|
| 366 |
-
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
| 367 |
-
"path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2/checkpoints/steps_5000_model.safetensors"
|
| 368 |
-
},
|
| 369 |
{
|
| 370 |
"repo": "latency-sensitive-bench/benchmark-models",
|
| 371 |
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
|
@@ -443,11 +418,6 @@
|
|
| 443 |
"path": "historical/openvla_flappy_fix_latency_2_small"
|
| 444 |
},
|
| 445 |
"checkpoint": [
|
| 446 |
-
{
|
| 447 |
-
"repo": "latency-sensitive-bench/benchmark-models",
|
| 448 |
-
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
| 449 |
-
"path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_2_small/checkpoints/best_model.safetensors"
|
| 450 |
-
},
|
| 451 |
{
|
| 452 |
"repo": "latency-sensitive-bench/benchmark-models",
|
| 453 |
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
|
@@ -536,11 +506,6 @@
|
|
| 536 |
"path": "historical/openvla_flappy_fix_latency_3_small"
|
| 537 |
},
|
| 538 |
"checkpoint": [
|
| 539 |
-
{
|
| 540 |
-
"repo": "latency-sensitive-bench/benchmark-models",
|
| 541 |
-
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
| 542 |
-
"path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_3_small/checkpoints/best_model.safetensors"
|
| 543 |
-
},
|
| 544 |
{
|
| 545 |
"repo": "latency-sensitive-bench/benchmark-models",
|
| 546 |
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
|
@@ -629,11 +594,6 @@
|
|
| 629 |
"path": "historical/openvla_flappy_fix_latency_4_exp"
|
| 630 |
},
|
| 631 |
"checkpoint": [
|
| 632 |
-
{
|
| 633 |
-
"repo": "latency-sensitive-bench/benchmark-models",
|
| 634 |
-
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
| 635 |
-
"path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_4_exp/checkpoints/steps_2500_model.safetensors"
|
| 636 |
-
},
|
| 637 |
{
|
| 638 |
"repo": "latency-sensitive-bench/benchmark-models",
|
| 639 |
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
|
@@ -711,11 +671,6 @@
|
|
| 711 |
"path": "historical/openvla_flappy_fix_latency_4_small"
|
| 712 |
},
|
| 713 |
"checkpoint": [
|
| 714 |
-
{
|
| 715 |
-
"repo": "latency-sensitive-bench/benchmark-models",
|
| 716 |
-
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
| 717 |
-
"path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_4_small/checkpoints/best_model.safetensors"
|
| 718 |
-
},
|
| 719 |
{
|
| 720 |
"repo": "latency-sensitive-bench/benchmark-models",
|
| 721 |
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
|
@@ -804,11 +759,6 @@
|
|
| 804 |
"path": "historical/openvla_flappy_fix_latency_6"
|
| 805 |
},
|
| 806 |
"checkpoint": [
|
| 807 |
-
{
|
| 808 |
-
"repo": "latency-sensitive-bench/benchmark-models",
|
| 809 |
-
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
| 810 |
-
"path": "latency-aware/flappy/vla/starvla-qwenoft-h1/openvla_flappy_fix_latency_6/checkpoints/steps_5000_model.safetensors"
|
| 811 |
-
},
|
| 812 |
{
|
| 813 |
"repo": "latency-sensitive-bench/benchmark-models",
|
| 814 |
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
|
|
|
| 10 |
"path": "historical/openvla_flappy_fix_latency_0"
|
| 11 |
},
|
| 12 |
"checkpoint": [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 13 |
{
|
| 14 |
"repo": "latency-sensitive-bench/benchmark-models",
|
| 15 |
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
|
|
|
| 87 |
"path": "historical/openvla_flappy_fix_latency_0_small"
|
| 88 |
},
|
| 89 |
"checkpoint": [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 90 |
{
|
| 91 |
"repo": "latency-sensitive-bench/benchmark-models",
|
| 92 |
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
|
|
|
| 175 |
"path": "historical/openvla_flappy_fix_latency_1_small"
|
| 176 |
},
|
| 177 |
"checkpoint": [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 178 |
{
|
| 179 |
"repo": "latency-sensitive-bench/benchmark-models",
|
| 180 |
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
|
|
|
| 263 |
"path": "historical/openvla_flappy_fix_latency_1_small_tmp"
|
| 264 |
},
|
| 265 |
"checkpoint": [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 266 |
{
|
| 267 |
"repo": "latency-sensitive-bench/benchmark-models",
|
| 268 |
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
|
|
|
| 341 |
"path": "historical/openvla_flappy_fix_latency_2"
|
| 342 |
},
|
| 343 |
"checkpoint": [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 344 |
{
|
| 345 |
"repo": "latency-sensitive-bench/benchmark-models",
|
| 346 |
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
|
|
|
| 418 |
"path": "historical/openvla_flappy_fix_latency_2_small"
|
| 419 |
},
|
| 420 |
"checkpoint": [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 421 |
{
|
| 422 |
"repo": "latency-sensitive-bench/benchmark-models",
|
| 423 |
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
|
|
|
| 506 |
"path": "historical/openvla_flappy_fix_latency_3_small"
|
| 507 |
},
|
| 508 |
"checkpoint": [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 509 |
{
|
| 510 |
"repo": "latency-sensitive-bench/benchmark-models",
|
| 511 |
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
|
|
|
| 594 |
"path": "historical/openvla_flappy_fix_latency_4_exp"
|
| 595 |
},
|
| 596 |
"checkpoint": [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 597 |
{
|
| 598 |
"repo": "latency-sensitive-bench/benchmark-models",
|
| 599 |
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
|
|
|
| 671 |
"path": "historical/openvla_flappy_fix_latency_4_small"
|
| 672 |
},
|
| 673 |
"checkpoint": [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 674 |
{
|
| 675 |
"repo": "latency-sensitive-bench/benchmark-models",
|
| 676 |
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
|
|
|
| 759 |
"path": "historical/openvla_flappy_fix_latency_6"
|
| 760 |
},
|
| 761 |
"checkpoint": [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 762 |
{
|
| 763 |
"repo": "latency-sensitive-bench/benchmark-models",
|
| 764 |
"revision": "46266ba5a184ac622f7fa3386ac689d8d755d64c",
|
experiments/observation-stride/deadly-corridor/manifest.json
CHANGED
|
@@ -671,13 +671,6 @@
|
|
| 671 |
"path": "latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:observation_stride:fixed_l4:fs4:obs5p83333:stride6:seed12/config.json"
|
| 672 |
},
|
| 673 |
"retained_checkpoints": [
|
| 674 |
-
{
|
| 675 |
-
"repo_id": "latency-sensitive-bench/benchmark-models",
|
| 676 |
-
"repo_type": "model",
|
| 677 |
-
"revision": "365fe5402de3fb464d5da8da9f944b969c5801c4",
|
| 678 |
-
"path": "latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:observation_stride:fixed_l4:fs4:obs5p83333:stride6:seed12/checkpoint_p0/checkpoint_000048836_25004032.pth",
|
| 679 |
-
"sha256": "6e990229272411cf177d56e5c97cfe85341497828a3848cace8c6055498b1a72"
|
| 680 |
-
},
|
| 681 |
{
|
| 682 |
"repo_id": "latency-sensitive-bench/benchmark-models",
|
| 683 |
"repo_type": "model",
|
|
@@ -884,13 +877,6 @@
|
|
| 884 |
"path": "latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:observation_stride:fixed_l4:fs4:obs7:stride5:seed12/config.json"
|
| 885 |
},
|
| 886 |
"retained_checkpoints": [
|
| 887 |
-
{
|
| 888 |
-
"repo_id": "latency-sensitive-bench/benchmark-models",
|
| 889 |
-
"repo_type": "model",
|
| 890 |
-
"revision": "365fe5402de3fb464d5da8da9f944b969c5801c4",
|
| 891 |
-
"path": "latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:observation_stride:fixed_l4:fs4:obs7:stride5:seed12/checkpoint_p0/checkpoint_000048836_25004032.pth",
|
| 892 |
-
"sha256": "318304e07d049d916671413cc0f4d83581b8fcc99535392b3ebc63d64c79fda1"
|
| 893 |
-
},
|
| 894 |
{
|
| 895 |
"repo_id": "latency-sensitive-bench/benchmark-models",
|
| 896 |
"repo_type": "model",
|
|
@@ -1241,13 +1227,6 @@
|
|
| 1241 |
"path": "latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:observation_stride:uniform_u2_6:fs4:obs11p6667:stride3:seed11/config.json"
|
| 1242 |
},
|
| 1243 |
"retained_checkpoints": [
|
| 1244 |
-
{
|
| 1245 |
-
"repo_id": "latency-sensitive-bench/benchmark-models",
|
| 1246 |
-
"repo_type": "model",
|
| 1247 |
-
"revision": "365fe5402de3fb464d5da8da9f944b969c5801c4",
|
| 1248 |
-
"path": "latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:observation_stride:uniform_u2_6:fs4:obs11p6667:stride3:seed11/checkpoint_p0/checkpoint_000048836_25004032.pth",
|
| 1249 |
-
"sha256": "973d2b152476dde80b0ab1be9d66ed11a6e44a37529e8a8a8aa7d0b5bd2877b3"
|
| 1250 |
-
},
|
| 1251 |
{
|
| 1252 |
"repo_id": "latency-sensitive-bench/benchmark-models",
|
| 1253 |
"repo_type": "model",
|
|
@@ -1343,13 +1322,6 @@
|
|
| 1343 |
"path": "latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:observation_stride:uniform_u2_6:fs4:obs11p6667:stride3:seed13/config.json"
|
| 1344 |
},
|
| 1345 |
"retained_checkpoints": [
|
| 1346 |
-
{
|
| 1347 |
-
"repo_id": "latency-sensitive-bench/benchmark-models",
|
| 1348 |
-
"repo_type": "model",
|
| 1349 |
-
"revision": "365fe5402de3fb464d5da8da9f944b969c5801c4",
|
| 1350 |
-
"path": "latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:observation_stride:uniform_u2_6:fs4:obs11p6667:stride3:seed13/checkpoint_p0/checkpoint_000048836_25004032.pth",
|
| 1351 |
-
"sha256": "380ee41d7dcb3407473049516e7e13514d3bddd981a164065767dc1ea31f660a"
|
| 1352 |
-
},
|
| 1353 |
{
|
| 1354 |
"repo_id": "latency-sensitive-bench/benchmark-models",
|
| 1355 |
"repo_type": "model",
|
|
@@ -1404,13 +1376,6 @@
|
|
| 1404 |
"path": "latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:observation_stride:uniform_u2_6:fs4:obs11p6667:stride3:seed14/config.json"
|
| 1405 |
},
|
| 1406 |
"retained_checkpoints": [
|
| 1407 |
-
{
|
| 1408 |
-
"repo_id": "latency-sensitive-bench/benchmark-models",
|
| 1409 |
-
"repo_type": "model",
|
| 1410 |
-
"revision": "365fe5402de3fb464d5da8da9f944b969c5801c4",
|
| 1411 |
-
"path": "latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:observation_stride:uniform_u2_6:fs4:obs11p6667:stride3:seed14/checkpoint_p0/checkpoint_000048836_25004032.pth",
|
| 1412 |
-
"sha256": "345c4b62af029741aa4815321434e353537482184c2cc4540fd17e343b07413a"
|
| 1413 |
-
},
|
| 1414 |
{
|
| 1415 |
"repo_id": "latency-sensitive-bench/benchmark-models",
|
| 1416 |
"repo_type": "model",
|
|
@@ -1465,13 +1430,6 @@
|
|
| 1465 |
"path": "latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:observation_stride:uniform_u2_6:fs4:obs17p5:stride2:seed10/config.json"
|
| 1466 |
},
|
| 1467 |
"retained_checkpoints": [
|
| 1468 |
-
{
|
| 1469 |
-
"repo_id": "latency-sensitive-bench/benchmark-models",
|
| 1470 |
-
"repo_type": "model",
|
| 1471 |
-
"revision": "365fe5402de3fb464d5da8da9f944b969c5801c4",
|
| 1472 |
-
"path": "latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:observation_stride:uniform_u2_6:fs4:obs17p5:stride2:seed10/checkpoint_p0/checkpoint_000048836_25004032.pth",
|
| 1473 |
-
"sha256": "c7c054e7d5a7f05fd6d4fe310eac21d1f7e4f137807af7eadc8cb7f2626425e5"
|
| 1474 |
-
},
|
| 1475 |
{
|
| 1476 |
"repo_id": "latency-sensitive-bench/benchmark-models",
|
| 1477 |
"repo_type": "model",
|
|
@@ -1928,13 +1886,6 @@
|
|
| 1928 |
"path": "latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:observation_stride:uniform_u2_6:fs4:obs5p83333:stride6:seed12/config.json"
|
| 1929 |
},
|
| 1930 |
"retained_checkpoints": [
|
| 1931 |
-
{
|
| 1932 |
-
"repo_id": "latency-sensitive-bench/benchmark-models",
|
| 1933 |
-
"repo_type": "model",
|
| 1934 |
-
"revision": "365fe5402de3fb464d5da8da9f944b969c5801c4",
|
| 1935 |
-
"path": "latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:observation_stride:uniform_u2_6:fs4:obs5p83333:stride6:seed12/checkpoint_p0/checkpoint_000048836_25004032.pth",
|
| 1936 |
-
"sha256": "d9ee24783c3136de50dcf7e259e05e51331de8692287f647acccb26d710ba651"
|
| 1937 |
-
},
|
| 1938 |
{
|
| 1939 |
"repo_id": "latency-sensitive-bench/benchmark-models",
|
| 1940 |
"repo_type": "model",
|
|
@@ -1989,13 +1940,6 @@
|
|
| 1989 |
"path": "latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:observation_stride:uniform_u2_6:fs4:obs5p83333:stride6:seed13/config.json"
|
| 1990 |
},
|
| 1991 |
"retained_checkpoints": [
|
| 1992 |
-
{
|
| 1993 |
-
"repo_id": "latency-sensitive-bench/benchmark-models",
|
| 1994 |
-
"repo_type": "model",
|
| 1995 |
-
"revision": "365fe5402de3fb464d5da8da9f944b969c5801c4",
|
| 1996 |
-
"path": "latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:observation_stride:uniform_u2_6:fs4:obs5p83333:stride6:seed13/checkpoint_p0/checkpoint_000048836_25004032.pth",
|
| 1997 |
-
"sha256": "0bc5c89dd97295a9c816c93db075e96b3ddd261bf340b3ada6369f366e472001"
|
| 1998 |
-
},
|
| 1999 |
{
|
| 2000 |
"repo_id": "latency-sensitive-bench/benchmark-models",
|
| 2001 |
"repo_type": "model",
|
|
@@ -2198,13 +2142,6 @@
|
|
| 2198 |
"path": "latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:observation_stride:uniform_u2_6:fs4:obs7:stride5:seed13/config.json"
|
| 2199 |
},
|
| 2200 |
"retained_checkpoints": [
|
| 2201 |
-
{
|
| 2202 |
-
"repo_id": "latency-sensitive-bench/benchmark-models",
|
| 2203 |
-
"repo_type": "model",
|
| 2204 |
-
"revision": "365fe5402de3fb464d5da8da9f944b969c5801c4",
|
| 2205 |
-
"path": "latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:observation_stride:uniform_u2_6:fs4:obs7:stride5:seed13/checkpoint_p0/checkpoint_000048836_25004032.pth",
|
| 2206 |
-
"sha256": "73be58dc658781d947739a61fea15385bcd078f71510c1d1302db69589f70cf1"
|
| 2207 |
-
},
|
| 2208 |
{
|
| 2209 |
"repo_id": "latency-sensitive-bench/benchmark-models",
|
| 2210 |
"repo_type": "model",
|
|
|
|
| 671 |
"path": "latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:observation_stride:fixed_l4:fs4:obs5p83333:stride6:seed12/config.json"
|
| 672 |
},
|
| 673 |
"retained_checkpoints": [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 674 |
{
|
| 675 |
"repo_id": "latency-sensitive-bench/benchmark-models",
|
| 676 |
"repo_type": "model",
|
|
|
|
| 877 |
"path": "latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:observation_stride:fixed_l4:fs4:obs7:stride5:seed12/config.json"
|
| 878 |
},
|
| 879 |
"retained_checkpoints": [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 880 |
{
|
| 881 |
"repo_id": "latency-sensitive-bench/benchmark-models",
|
| 882 |
"repo_type": "model",
|
|
|
|
| 1227 |
"path": "latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:observation_stride:uniform_u2_6:fs4:obs11p6667:stride3:seed11/config.json"
|
| 1228 |
},
|
| 1229 |
"retained_checkpoints": [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1230 |
{
|
| 1231 |
"repo_id": "latency-sensitive-bench/benchmark-models",
|
| 1232 |
"repo_type": "model",
|
|
|
|
| 1322 |
"path": "latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:observation_stride:uniform_u2_6:fs4:obs11p6667:stride3:seed13/config.json"
|
| 1323 |
},
|
| 1324 |
"retained_checkpoints": [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1325 |
{
|
| 1326 |
"repo_id": "latency-sensitive-bench/benchmark-models",
|
| 1327 |
"repo_type": "model",
|
|
|
|
| 1376 |
"path": "latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:observation_stride:uniform_u2_6:fs4:obs11p6667:stride3:seed14/config.json"
|
| 1377 |
},
|
| 1378 |
"retained_checkpoints": [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1379 |
{
|
| 1380 |
"repo_id": "latency-sensitive-bench/benchmark-models",
|
| 1381 |
"repo_type": "model",
|
|
|
|
| 1430 |
"path": "latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:observation_stride:uniform_u2_6:fs4:obs17p5:stride2:seed10/config.json"
|
| 1431 |
},
|
| 1432 |
"retained_checkpoints": [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1433 |
{
|
| 1434 |
"repo_id": "latency-sensitive-bench/benchmark-models",
|
| 1435 |
"repo_type": "model",
|
|
|
|
| 1886 |
"path": "latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:observation_stride:uniform_u2_6:fs4:obs5p83333:stride6:seed12/config.json"
|
| 1887 |
},
|
| 1888 |
"retained_checkpoints": [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1889 |
{
|
| 1890 |
"repo_id": "latency-sensitive-bench/benchmark-models",
|
| 1891 |
"repo_type": "model",
|
|
|
|
| 1940 |
"path": "latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:observation_stride:uniform_u2_6:fs4:obs5p83333:stride6:seed13/config.json"
|
| 1941 |
},
|
| 1942 |
"retained_checkpoints": [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1943 |
{
|
| 1944 |
"repo_id": "latency-sensitive-bench/benchmark-models",
|
| 1945 |
"repo_type": "model",
|
|
|
|
| 2142 |
"path": "latency-aware/deadly-corridor/small-policy/factor_sweep:deadly_corridor:observation_stride:uniform_u2_6:fs4:obs7:stride5:seed13/config.json"
|
| 2143 |
},
|
| 2144 |
"retained_checkpoints": [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2145 |
{
|
| 2146 |
"repo_id": "latency-sensitive-bench/benchmark-models",
|
| 2147 |
"repo_type": "model",
|
latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_0_200ep/filter_report.json
DELETED
|
@@ -1,19 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"accepted_episodes": 200,
|
| 3 |
-
"accepted_train_episodes": 180,
|
| 4 |
-
"accepted_val_episodes": 20,
|
| 5 |
-
"attempted_episodes": 282,
|
| 6 |
-
"preset": "clean_v1",
|
| 7 |
-
"raw_score_bins": {
|
| 8 |
-
"6460-6889": 0,
|
| 9 |
-
"6890-7309": 0,
|
| 10 |
-
"7310-7999": 91,
|
| 11 |
-
"<6460": 0,
|
| 12 |
-
">=8000": 109
|
| 13 |
-
},
|
| 14 |
-
"rejected_episodes": 82,
|
| 15 |
-
"train_rows_after_row_filter": 435000,
|
| 16 |
-
"train_rows_before_row_filter": 648000,
|
| 17 |
-
"val_rows_after_row_filter": 68014,
|
| 18 |
-
"val_rows_before_row_filter": 72000
|
| 19 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_0_200ep/metadata.json
DELETED
|
@@ -1,289 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"action_labels": [
|
| 3 |
-
"NOOP",
|
| 4 |
-
"FIRE",
|
| 5 |
-
"RIGHT",
|
| 6 |
-
"LEFT",
|
| 7 |
-
"RIGHTFIRE",
|
| 8 |
-
"LEFTFIRE"
|
| 9 |
-
],
|
| 10 |
-
"action_layout": "demon_attack_discrete",
|
| 11 |
-
"base_prompt": "You are playing Demon Attack from a single game image. Choose exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, LEFTFIRE. Survive, avoid enemy fire, and shoot enemies to maximize score.",
|
| 12 |
-
"checkpoint_config": {
|
| 13 |
-
"action_repeat": 4,
|
| 14 |
-
"actor_critic_share_weights": true,
|
| 15 |
-
"actor_worker_gpus": [],
|
| 16 |
-
"adam_beta1": 0.9,
|
| 17 |
-
"adam_beta2": 0.999,
|
| 18 |
-
"adam_eps": 1e-05,
|
| 19 |
-
"adaptive_stddev": true,
|
| 20 |
-
"add_latency_info": false,
|
| 21 |
-
"algo": "APPO",
|
| 22 |
-
"async_rl": true,
|
| 23 |
-
"batch_size": 1024,
|
| 24 |
-
"batched_sampling": true,
|
| 25 |
-
"benchmark": false,
|
| 26 |
-
"cli_args": {
|
| 27 |
-
"action_repeat": 4,
|
| 28 |
-
"add_latency_info": false,
|
| 29 |
-
"algo": "APPO",
|
| 30 |
-
"async_rl": true,
|
| 31 |
-
"batch_size": 1024,
|
| 32 |
-
"batched_sampling": true,
|
| 33 |
-
"clip_reward": true,
|
| 34 |
-
"device": "gpu",
|
| 35 |
-
"env": "latency_demon_attack",
|
| 36 |
-
"env_id": "ALE/DemonAttack-v5",
|
| 37 |
-
"episode_metrics_path": "results/checkpoints/demon_attack_sf_same_latency_l0_seed0/episode_metrics.jsonl",
|
| 38 |
-
"episodes_per_stage": 10,
|
| 39 |
-
"eval_deterministic": true,
|
| 40 |
-
"eval_episodes": 20,
|
| 41 |
-
"eval_max_steps": 3600,
|
| 42 |
-
"eval_queue_latency_values": "0,1,2,3,4,5,10,15",
|
| 43 |
-
"eval_raw_reward": true,
|
| 44 |
-
"experiment": "demon_attack_sf_same_latency_l0_seed0",
|
| 45 |
-
"experiment_summaries_interval": 1,
|
| 46 |
-
"frame_stack": 4,
|
| 47 |
-
"full_action_space": false,
|
| 48 |
-
"gae_lambda": 0.95,
|
| 49 |
-
"gamma": 0.99,
|
| 50 |
-
"keep_checkpoints": 5,
|
| 51 |
-
"latency_mode": "episode_fixed",
|
| 52 |
-
"latency_sampling_mode": "uniform_all",
|
| 53 |
-
"learning_rate": 0.000303,
|
| 54 |
-
"max_episode_steps": 25000,
|
| 55 |
-
"max_grad_norm": 0.0,
|
| 56 |
-
"mode": "train",
|
| 57 |
-
"noop_max": 30,
|
| 58 |
-
"normalize_input": true,
|
| 59 |
-
"normalize_returns": true,
|
| 60 |
-
"num_batches_per_epoch": 2,
|
| 61 |
-
"num_envs_per_worker": 4,
|
| 62 |
-
"num_epochs": 2,
|
| 63 |
-
"num_policies": 1,
|
| 64 |
-
"num_workers": 16,
|
| 65 |
-
"ppo_clip_ratio": 0.1,
|
| 66 |
-
"ppo_clip_value": 0.2,
|
| 67 |
-
"queue_latency_values": "0",
|
| 68 |
-
"randomize_within_stage": false,
|
| 69 |
-
"recurrence": 32,
|
| 70 |
-
"repeat_action_probability": 0.0,
|
| 71 |
-
"restart_behavior": "overwrite",
|
| 72 |
-
"rollout": 128,
|
| 73 |
-
"save_every_sec": 600,
|
| 74 |
-
"screen_size": 84,
|
| 75 |
-
"seed": 0,
|
| 76 |
-
"stats_avg": 100,
|
| 77 |
-
"terminal_on_life_loss": true,
|
| 78 |
-
"train_dir": "results/checkpoints",
|
| 79 |
-
"train_for_env_steps": 50000000,
|
| 80 |
-
"use_rnn": false,
|
| 81 |
-
"value_loss_coeff": 0.5,
|
| 82 |
-
"wandb_group": "demon_attack",
|
| 83 |
-
"wandb_job_type": "sample_factory",
|
| 84 |
-
"wandb_project": "latency-sensitive-bench",
|
| 85 |
-
"with_wandb": true,
|
| 86 |
-
"within_episode_change_interval": 1,
|
| 87 |
-
"within_episode_change_prob": 1.0,
|
| 88 |
-
"worker_num_splits": 2
|
| 89 |
-
},
|
| 90 |
-
"clip_reward": true,
|
| 91 |
-
"command_line": "--mode train --algo APPO --env latency_demon_attack --experiment demon_attack_sf_same_latency_l0_seed0 --train_dir results/checkpoints --restart_behavior overwrite --device gpu --seed 0 --episode_metrics_path results/checkpoints/demon_attack_sf_same_latency_l0_seed0/episode_metrics.jsonl --train_for_env_steps 50000000 --num_workers 16 --num_envs_per_worker 4 --num_policies 1 --batch_size 1024 --rollout 128 --recurrence 32 --num_epochs 2 --num_batches_per_epoch 2 --worker_num_splits 2 --learning_rate 0.000303 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.0 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --async_rl True --batched_sampling True --use_rnn False --normalize_returns True --normalize_input True --queue-latency-values 0 --latency-mode episode_fixed --latency-sampling-mode uniform_all --episodes-per-stage 10 --randomize-within-stage False --within-episode-change-interval 1 --within-episode-change-prob 1.0 --add-latency-info False --eval-episodes 20 --eval-queue-latency-values 0,1,2,3,4,5,10,15 --eval-max-steps 3600 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project latency-sensitive-bench --wandb_group demon_attack --wandb_job_type sample_factory --env_id ALE/DemonAttack-v5 --screen_size 84 --noop_max 30 --action_repeat 4 --frame_stack 4 --terminal_on_life_loss True --clip_reward True --repeat_action_probability 0.0 --full_action_space False --max_episode_steps 25000",
|
| 92 |
-
"continuous_tanh_scale": 0.0,
|
| 93 |
-
"decoder_mlp_layers": [],
|
| 94 |
-
"decorrelate_envs_on_one_worker": true,
|
| 95 |
-
"decorrelate_experience_max_seconds": 0,
|
| 96 |
-
"default_niceness": 0,
|
| 97 |
-
"device": "gpu",
|
| 98 |
-
"encoder_conv_architecture": "convnet_atari",
|
| 99 |
-
"encoder_conv_mlp_layers": [
|
| 100 |
-
512
|
| 101 |
-
],
|
| 102 |
-
"encoder_mlp_layers": [
|
| 103 |
-
512,
|
| 104 |
-
512
|
| 105 |
-
],
|
| 106 |
-
"env": "latency_demon_attack",
|
| 107 |
-
"env_frameskip": 1,
|
| 108 |
-
"env_framestack": 1,
|
| 109 |
-
"env_gpu_actions": false,
|
| 110 |
-
"env_gpu_observations": true,
|
| 111 |
-
"env_id": "ALE/DemonAttack-v5",
|
| 112 |
-
"episode_metrics_path": "results/checkpoints/demon_attack_sf_same_latency_l0_seed0/episode_metrics.jsonl",
|
| 113 |
-
"episodes_per_stage": 10,
|
| 114 |
-
"eval_deterministic": true,
|
| 115 |
-
"eval_episodes": 20,
|
| 116 |
-
"eval_max_steps": 3600,
|
| 117 |
-
"eval_queue_latency_values": "0,1,2,3,4,5,10,15",
|
| 118 |
-
"eval_raw_reward": true,
|
| 119 |
-
"experiment": "demon_attack_sf_same_latency_l0_seed0",
|
| 120 |
-
"experiment_summaries_interval": 1,
|
| 121 |
-
"exploration_loss": "entropy",
|
| 122 |
-
"exploration_loss_coeff": 0.003,
|
| 123 |
-
"flush_summaries_interval": 30,
|
| 124 |
-
"force_envs_single_thread": false,
|
| 125 |
-
"frame_stack": 4,
|
| 126 |
-
"full_action_space": false,
|
| 127 |
-
"gae_lambda": 0.95,
|
| 128 |
-
"gamma": 0.99,
|
| 129 |
-
"git_hash": "41bc95edbf670a1190d128da22e9cd3777b87989",
|
| 130 |
-
"git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
|
| 131 |
-
"heartbeat_interval": 20,
|
| 132 |
-
"heartbeat_reporting_interval": 180,
|
| 133 |
-
"help": false,
|
| 134 |
-
"initial_stddev": 1.0,
|
| 135 |
-
"keep_checkpoints": 5,
|
| 136 |
-
"kl_loss_coeff": 0.0,
|
| 137 |
-
"latency_mode": "episode_fixed",
|
| 138 |
-
"latency_sampling_mode": "uniform_all",
|
| 139 |
-
"learning_rate": 0.000303,
|
| 140 |
-
"load_checkpoint_kind": "latest",
|
| 141 |
-
"log_to_file": true,
|
| 142 |
-
"lr_adaptive_max": 0.01,
|
| 143 |
-
"lr_adaptive_min": 1e-06,
|
| 144 |
-
"lr_schedule": "linear_decay",
|
| 145 |
-
"lr_schedule_kl_threshold": 0.008,
|
| 146 |
-
"max_episode_steps": 25000,
|
| 147 |
-
"max_grad_norm": 0.0,
|
| 148 |
-
"max_policy_lag": 1000,
|
| 149 |
-
"mode": "train",
|
| 150 |
-
"nonlinearity": "elu",
|
| 151 |
-
"noop_max": 30,
|
| 152 |
-
"normalize_input": true,
|
| 153 |
-
"normalize_input_keys": null,
|
| 154 |
-
"normalize_returns": true,
|
| 155 |
-
"num_batches_per_epoch": 2,
|
| 156 |
-
"num_batches_to_accumulate": 2,
|
| 157 |
-
"num_envs_per_worker": 4,
|
| 158 |
-
"num_epochs": 2,
|
| 159 |
-
"num_policies": 1,
|
| 160 |
-
"num_workers": 16,
|
| 161 |
-
"obs_scale": 255.0,
|
| 162 |
-
"obs_subtract_mean": 0.0,
|
| 163 |
-
"optimizer": "adam",
|
| 164 |
-
"output_dir": "outputs/train/demon_attack/latency_0/seed_0",
|
| 165 |
-
"pbt_mix_policies_in_one_env": true,
|
| 166 |
-
"pbt_mutation_rate": 0.15,
|
| 167 |
-
"pbt_optimize_gamma": false,
|
| 168 |
-
"pbt_period_env_steps": 5000000,
|
| 169 |
-
"pbt_perturb_max": 1.5,
|
| 170 |
-
"pbt_perturb_min": 1.1,
|
| 171 |
-
"pbt_replace_fraction": 0.3,
|
| 172 |
-
"pbt_replace_reward_gap": 0.1,
|
| 173 |
-
"pbt_replace_reward_gap_absolute": 1e-06,
|
| 174 |
-
"pbt_start_mutation": 20000000,
|
| 175 |
-
"pbt_target_objective": "true_objective",
|
| 176 |
-
"pixel_format": "CHW",
|
| 177 |
-
"policy_init_gain": 1.0,
|
| 178 |
-
"policy_initialization": "orthogonal",
|
| 179 |
-
"policy_workers_per_policy": 1,
|
| 180 |
-
"ppo_clip_ratio": 0.1,
|
| 181 |
-
"ppo_clip_value": 0.2,
|
| 182 |
-
"queue_latency_values": "0",
|
| 183 |
-
"randomize_within_stage": false,
|
| 184 |
-
"recurrence": 32,
|
| 185 |
-
"repeat_action_probability": 0.0,
|
| 186 |
-
"restart_behavior": "overwrite",
|
| 187 |
-
"reward_clip": 1000.0,
|
| 188 |
-
"reward_scale": 1.0,
|
| 189 |
-
"rnn_num_layers": 1,
|
| 190 |
-
"rnn_size": 512,
|
| 191 |
-
"rnn_type": "gru",
|
| 192 |
-
"rollout": 128,
|
| 193 |
-
"save_best_after": 100000,
|
| 194 |
-
"save_best_every_sec": 5,
|
| 195 |
-
"save_best_metric": "reward",
|
| 196 |
-
"save_every_sec": 600,
|
| 197 |
-
"save_milestones_sec": -1,
|
| 198 |
-
"screen_size": 84,
|
| 199 |
-
"seed": 0,
|
| 200 |
-
"serial_mode": false,
|
| 201 |
-
"set_workers_cpu_affinity": true,
|
| 202 |
-
"shuffle_minibatches": false,
|
| 203 |
-
"stats_avg": 100,
|
| 204 |
-
"summaries_use_frameskip": true,
|
| 205 |
-
"terminal_on_life_loss": true,
|
| 206 |
-
"train_dir": "results/checkpoints",
|
| 207 |
-
"train_for_env_steps": 50000000,
|
| 208 |
-
"train_for_seconds": 10000000000,
|
| 209 |
-
"use_env_info_cache": false,
|
| 210 |
-
"use_record_episode_statistics": false,
|
| 211 |
-
"use_rnn": false,
|
| 212 |
-
"value_bootstrap": false,
|
| 213 |
-
"value_loss_coeff": 0.5,
|
| 214 |
-
"vtrace_c": 1.0,
|
| 215 |
-
"vtrace_rho": 1.0,
|
| 216 |
-
"wandb_group": "demon_attack",
|
| 217 |
-
"wandb_job_type": "sample_factory",
|
| 218 |
-
"wandb_project": "latency-sensitive-bench",
|
| 219 |
-
"wandb_tags": [],
|
| 220 |
-
"wandb_unique_id": "demon-attack-l0-s0",
|
| 221 |
-
"wandb_user": null,
|
| 222 |
-
"with_pbt": false,
|
| 223 |
-
"with_vtrace": false,
|
| 224 |
-
"with_wandb": true,
|
| 225 |
-
"within_episode_change_interval": 1,
|
| 226 |
-
"within_episode_change_prob": 1.0,
|
| 227 |
-
"worker_num_splits": 2
|
| 228 |
-
},
|
| 229 |
-
"checkpoint_experiment_dir": "/workspace/latency-sensitive-bench/checkpoints/demon_attack_sf_same_latency_l0_seed0",
|
| 230 |
-
"checkpoint_path": null,
|
| 231 |
-
"checkpoint_train_step": 97664,
|
| 232 |
-
"config_source": "/workspace/latency-sensitive-bench/checkpoints/demon_attack_sf_same_latency_l0_seed0/config.json",
|
| 233 |
-
"env_fps": 60.0,
|
| 234 |
-
"env_name": "demon_attack",
|
| 235 |
-
"episode_return_field": "episode_raw_return",
|
| 236 |
-
"filter_preset": "clean_v1",
|
| 237 |
-
"filter_report": {
|
| 238 |
-
"accepted_episodes": 200,
|
| 239 |
-
"accepted_train_episodes": 180,
|
| 240 |
-
"accepted_val_episodes": 20,
|
| 241 |
-
"attempted_episodes": 282,
|
| 242 |
-
"preset": "clean_v1",
|
| 243 |
-
"raw_score_bins": {
|
| 244 |
-
"6460-6889": 0,
|
| 245 |
-
"6890-7309": 0,
|
| 246 |
-
"7310-7999": 91,
|
| 247 |
-
"<6460": 0,
|
| 248 |
-
">=8000": 109
|
| 249 |
-
},
|
| 250 |
-
"rejected_episodes": 82,
|
| 251 |
-
"train_rows_after_row_filter": 435000,
|
| 252 |
-
"train_rows_before_row_filter": 648000,
|
| 253 |
-
"val_rows_after_row_filter": 68014,
|
| 254 |
-
"val_rows_before_row_filter": 72000
|
| 255 |
-
},
|
| 256 |
-
"frame_stack": 4,
|
| 257 |
-
"image_resolution": [
|
| 258 |
-
210,
|
| 259 |
-
160,
|
| 260 |
-
3
|
| 261 |
-
],
|
| 262 |
-
"image_source": "env_raw_rgb",
|
| 263 |
-
"latency_override": {
|
| 264 |
-
"add_latency_info": false,
|
| 265 |
-
"fixed_latency_ms": 0.0,
|
| 266 |
-
"type": "fixed"
|
| 267 |
-
},
|
| 268 |
-
"latency_source": "legacy_checkpoint_name",
|
| 269 |
-
"max_episodes": 200,
|
| 270 |
-
"max_steps_per_episode": 3600,
|
| 271 |
-
"num_envs": 512,
|
| 272 |
-
"obs_fps": 15.0,
|
| 273 |
-
"obs_stride_raw_frames": 4,
|
| 274 |
-
"prompt_source": "latency_bench.data.prompts",
|
| 275 |
-
"prompt_timing_suffix": " Current action latency is {latency_raw_frames} raw frames ({latency_ms:.2f} ms). The environment runs at {env_fps:g} FPS and observations are emitted at {obs_fps:g} FPS. Choose the best next action.",
|
| 276 |
-
"replay_num_envs": 16,
|
| 277 |
-
"reward_field": "raw_reward",
|
| 278 |
-
"reward_semantics": "raw_reward",
|
| 279 |
-
"rows_unit": "decision_step",
|
| 280 |
-
"schema_version": 2,
|
| 281 |
-
"split_seed": 0,
|
| 282 |
-
"total_episodes": 200,
|
| 283 |
-
"total_rows": 503014,
|
| 284 |
-
"train_episodes": 180,
|
| 285 |
-
"train_rows": 435000,
|
| 286 |
-
"val_episodes": 20,
|
| 287 |
-
"val_fraction": 0.1,
|
| 288 |
-
"val_rows": 68014
|
| 289 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_0_200ep/train.parquet
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:fcceb2dd4cfc7c2a088142612533e5bdbf65e61067f20e31928a0e30ab2d0958
|
| 3 |
-
size 256157082
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_0_200ep/val.parquet
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:900cebe2cd7ac9be11c1c80f1f7f0abf7243bb610d555567f006069b8b4a83c6
|
| 3 |
-
size 39825924
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_2_200ep/filter_report.json
DELETED
|
@@ -1,19 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"accepted_episodes": 200,
|
| 3 |
-
"accepted_train_episodes": 180,
|
| 4 |
-
"accepted_val_episodes": 20,
|
| 5 |
-
"attempted_episodes": 904,
|
| 6 |
-
"preset": "clean_v1",
|
| 7 |
-
"raw_score_bins": {
|
| 8 |
-
"6460-6889": 0,
|
| 9 |
-
"6890-7309": 0,
|
| 10 |
-
"7310-7999": 128,
|
| 11 |
-
"<6460": 0,
|
| 12 |
-
">=8000": 72
|
| 13 |
-
},
|
| 14 |
-
"rejected_episodes": 704,
|
| 15 |
-
"train_rows_after_row_filter": 349558,
|
| 16 |
-
"train_rows_before_row_filter": 611368,
|
| 17 |
-
"val_rows_after_row_filter": 52326,
|
| 18 |
-
"val_rows_before_row_filter": 69880
|
| 19 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_2_200ep/metadata.json
DELETED
|
@@ -1,289 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"action_labels": [
|
| 3 |
-
"NOOP",
|
| 4 |
-
"FIRE",
|
| 5 |
-
"RIGHT",
|
| 6 |
-
"LEFT",
|
| 7 |
-
"RIGHTFIRE",
|
| 8 |
-
"LEFTFIRE"
|
| 9 |
-
],
|
| 10 |
-
"action_layout": "demon_attack_discrete",
|
| 11 |
-
"base_prompt": "You are playing Demon Attack from a single game image. Choose exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, LEFTFIRE. Survive, avoid enemy fire, and shoot enemies to maximize score.",
|
| 12 |
-
"checkpoint_config": {
|
| 13 |
-
"action_repeat": 4,
|
| 14 |
-
"actor_critic_share_weights": true,
|
| 15 |
-
"actor_worker_gpus": [],
|
| 16 |
-
"adam_beta1": 0.9,
|
| 17 |
-
"adam_beta2": 0.999,
|
| 18 |
-
"adam_eps": 1e-05,
|
| 19 |
-
"adaptive_stddev": true,
|
| 20 |
-
"add_latency_info": false,
|
| 21 |
-
"algo": "APPO",
|
| 22 |
-
"async_rl": true,
|
| 23 |
-
"batch_size": 1024,
|
| 24 |
-
"batched_sampling": true,
|
| 25 |
-
"benchmark": false,
|
| 26 |
-
"cli_args": {
|
| 27 |
-
"action_repeat": 4,
|
| 28 |
-
"add_latency_info": false,
|
| 29 |
-
"algo": "APPO",
|
| 30 |
-
"async_rl": true,
|
| 31 |
-
"batch_size": 1024,
|
| 32 |
-
"batched_sampling": true,
|
| 33 |
-
"clip_reward": true,
|
| 34 |
-
"device": "gpu",
|
| 35 |
-
"env": "latency_demon_attack",
|
| 36 |
-
"env_id": "ALE/DemonAttack-v5",
|
| 37 |
-
"episode_metrics_path": "results/checkpoints/demon_attack_sf_same_latency_l2_seed0/episode_metrics.jsonl",
|
| 38 |
-
"episodes_per_stage": 10,
|
| 39 |
-
"eval_deterministic": true,
|
| 40 |
-
"eval_episodes": 20,
|
| 41 |
-
"eval_max_steps": 3600,
|
| 42 |
-
"eval_queue_latency_values": "0,1,2,3,4,5,10,15",
|
| 43 |
-
"eval_raw_reward": true,
|
| 44 |
-
"experiment": "demon_attack_sf_same_latency_l2_seed0",
|
| 45 |
-
"experiment_summaries_interval": 1,
|
| 46 |
-
"frame_stack": 4,
|
| 47 |
-
"full_action_space": false,
|
| 48 |
-
"gae_lambda": 0.95,
|
| 49 |
-
"gamma": 0.99,
|
| 50 |
-
"keep_checkpoints": 5,
|
| 51 |
-
"latency_mode": "episode_fixed",
|
| 52 |
-
"latency_sampling_mode": "uniform_all",
|
| 53 |
-
"learning_rate": 0.000303,
|
| 54 |
-
"max_episode_steps": 25000,
|
| 55 |
-
"max_grad_norm": 0.0,
|
| 56 |
-
"mode": "train",
|
| 57 |
-
"noop_max": 30,
|
| 58 |
-
"normalize_input": true,
|
| 59 |
-
"normalize_returns": true,
|
| 60 |
-
"num_batches_per_epoch": 2,
|
| 61 |
-
"num_envs_per_worker": 4,
|
| 62 |
-
"num_epochs": 2,
|
| 63 |
-
"num_policies": 1,
|
| 64 |
-
"num_workers": 16,
|
| 65 |
-
"ppo_clip_ratio": 0.1,
|
| 66 |
-
"ppo_clip_value": 0.2,
|
| 67 |
-
"queue_latency_values": "2",
|
| 68 |
-
"randomize_within_stage": false,
|
| 69 |
-
"recurrence": 32,
|
| 70 |
-
"repeat_action_probability": 0.0,
|
| 71 |
-
"restart_behavior": "overwrite",
|
| 72 |
-
"rollout": 128,
|
| 73 |
-
"save_every_sec": 600,
|
| 74 |
-
"screen_size": 84,
|
| 75 |
-
"seed": 0,
|
| 76 |
-
"stats_avg": 100,
|
| 77 |
-
"terminal_on_life_loss": true,
|
| 78 |
-
"train_dir": "results/checkpoints",
|
| 79 |
-
"train_for_env_steps": 50000000,
|
| 80 |
-
"use_rnn": false,
|
| 81 |
-
"value_loss_coeff": 0.5,
|
| 82 |
-
"wandb_group": "demon_attack",
|
| 83 |
-
"wandb_job_type": "sample_factory",
|
| 84 |
-
"wandb_project": "latency-sensitive-bench",
|
| 85 |
-
"with_wandb": true,
|
| 86 |
-
"within_episode_change_interval": 1,
|
| 87 |
-
"within_episode_change_prob": 1.0,
|
| 88 |
-
"worker_num_splits": 2
|
| 89 |
-
},
|
| 90 |
-
"clip_reward": true,
|
| 91 |
-
"command_line": "--mode train --algo APPO --env latency_demon_attack --experiment demon_attack_sf_same_latency_l2_seed0 --train_dir results/checkpoints --restart_behavior overwrite --device gpu --seed 0 --episode_metrics_path results/checkpoints/demon_attack_sf_same_latency_l2_seed0/episode_metrics.jsonl --train_for_env_steps 50000000 --num_workers 16 --num_envs_per_worker 4 --num_policies 1 --batch_size 1024 --rollout 128 --recurrence 32 --num_epochs 2 --num_batches_per_epoch 2 --worker_num_splits 2 --learning_rate 0.000303 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.0 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --async_rl True --batched_sampling True --use_rnn False --normalize_returns True --normalize_input True --queue-latency-values 2 --latency-mode episode_fixed --latency-sampling-mode uniform_all --episodes-per-stage 10 --randomize-within-stage False --within-episode-change-interval 1 --within-episode-change-prob 1.0 --add-latency-info False --eval-episodes 20 --eval-queue-latency-values 0,1,2,3,4,5,10,15 --eval-max-steps 3600 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project latency-sensitive-bench --wandb_group demon_attack --wandb_job_type sample_factory --env_id ALE/DemonAttack-v5 --screen_size 84 --noop_max 30 --action_repeat 4 --frame_stack 4 --terminal_on_life_loss True --clip_reward True --repeat_action_probability 0.0 --full_action_space False --max_episode_steps 25000",
|
| 92 |
-
"continuous_tanh_scale": 0.0,
|
| 93 |
-
"decoder_mlp_layers": [],
|
| 94 |
-
"decorrelate_envs_on_one_worker": true,
|
| 95 |
-
"decorrelate_experience_max_seconds": 0,
|
| 96 |
-
"default_niceness": 0,
|
| 97 |
-
"device": "gpu",
|
| 98 |
-
"encoder_conv_architecture": "convnet_atari",
|
| 99 |
-
"encoder_conv_mlp_layers": [
|
| 100 |
-
512
|
| 101 |
-
],
|
| 102 |
-
"encoder_mlp_layers": [
|
| 103 |
-
512,
|
| 104 |
-
512
|
| 105 |
-
],
|
| 106 |
-
"env": "latency_demon_attack",
|
| 107 |
-
"env_frameskip": 1,
|
| 108 |
-
"env_framestack": 1,
|
| 109 |
-
"env_gpu_actions": false,
|
| 110 |
-
"env_gpu_observations": true,
|
| 111 |
-
"env_id": "ALE/DemonAttack-v5",
|
| 112 |
-
"episode_metrics_path": "results/checkpoints/demon_attack_sf_same_latency_l2_seed0/episode_metrics.jsonl",
|
| 113 |
-
"episodes_per_stage": 10,
|
| 114 |
-
"eval_deterministic": true,
|
| 115 |
-
"eval_episodes": 20,
|
| 116 |
-
"eval_max_steps": 3600,
|
| 117 |
-
"eval_queue_latency_values": "0,1,2,3,4,5,10,15",
|
| 118 |
-
"eval_raw_reward": true,
|
| 119 |
-
"experiment": "demon_attack_sf_same_latency_l2_seed0",
|
| 120 |
-
"experiment_summaries_interval": 1,
|
| 121 |
-
"exploration_loss": "entropy",
|
| 122 |
-
"exploration_loss_coeff": 0.003,
|
| 123 |
-
"flush_summaries_interval": 30,
|
| 124 |
-
"force_envs_single_thread": false,
|
| 125 |
-
"frame_stack": 4,
|
| 126 |
-
"full_action_space": false,
|
| 127 |
-
"gae_lambda": 0.95,
|
| 128 |
-
"gamma": 0.99,
|
| 129 |
-
"git_hash": "41bc95edbf670a1190d128da22e9cd3777b87989",
|
| 130 |
-
"git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
|
| 131 |
-
"heartbeat_interval": 20,
|
| 132 |
-
"heartbeat_reporting_interval": 180,
|
| 133 |
-
"help": false,
|
| 134 |
-
"initial_stddev": 1.0,
|
| 135 |
-
"keep_checkpoints": 5,
|
| 136 |
-
"kl_loss_coeff": 0.0,
|
| 137 |
-
"latency_mode": "episode_fixed",
|
| 138 |
-
"latency_sampling_mode": "uniform_all",
|
| 139 |
-
"learning_rate": 0.000303,
|
| 140 |
-
"load_checkpoint_kind": "latest",
|
| 141 |
-
"log_to_file": true,
|
| 142 |
-
"lr_adaptive_max": 0.01,
|
| 143 |
-
"lr_adaptive_min": 1e-06,
|
| 144 |
-
"lr_schedule": "linear_decay",
|
| 145 |
-
"lr_schedule_kl_threshold": 0.008,
|
| 146 |
-
"max_episode_steps": 25000,
|
| 147 |
-
"max_grad_norm": 0.0,
|
| 148 |
-
"max_policy_lag": 1000,
|
| 149 |
-
"mode": "train",
|
| 150 |
-
"nonlinearity": "elu",
|
| 151 |
-
"noop_max": 30,
|
| 152 |
-
"normalize_input": true,
|
| 153 |
-
"normalize_input_keys": null,
|
| 154 |
-
"normalize_returns": true,
|
| 155 |
-
"num_batches_per_epoch": 2,
|
| 156 |
-
"num_batches_to_accumulate": 2,
|
| 157 |
-
"num_envs_per_worker": 4,
|
| 158 |
-
"num_epochs": 2,
|
| 159 |
-
"num_policies": 1,
|
| 160 |
-
"num_workers": 16,
|
| 161 |
-
"obs_scale": 255.0,
|
| 162 |
-
"obs_subtract_mean": 0.0,
|
| 163 |
-
"optimizer": "adam",
|
| 164 |
-
"output_dir": "outputs/train/demon_attack/latency_2/seed_0",
|
| 165 |
-
"pbt_mix_policies_in_one_env": true,
|
| 166 |
-
"pbt_mutation_rate": 0.15,
|
| 167 |
-
"pbt_optimize_gamma": false,
|
| 168 |
-
"pbt_period_env_steps": 5000000,
|
| 169 |
-
"pbt_perturb_max": 1.5,
|
| 170 |
-
"pbt_perturb_min": 1.1,
|
| 171 |
-
"pbt_replace_fraction": 0.3,
|
| 172 |
-
"pbt_replace_reward_gap": 0.1,
|
| 173 |
-
"pbt_replace_reward_gap_absolute": 1e-06,
|
| 174 |
-
"pbt_start_mutation": 20000000,
|
| 175 |
-
"pbt_target_objective": "true_objective",
|
| 176 |
-
"pixel_format": "CHW",
|
| 177 |
-
"policy_init_gain": 1.0,
|
| 178 |
-
"policy_initialization": "orthogonal",
|
| 179 |
-
"policy_workers_per_policy": 1,
|
| 180 |
-
"ppo_clip_ratio": 0.1,
|
| 181 |
-
"ppo_clip_value": 0.2,
|
| 182 |
-
"queue_latency_values": "2",
|
| 183 |
-
"randomize_within_stage": false,
|
| 184 |
-
"recurrence": 32,
|
| 185 |
-
"repeat_action_probability": 0.0,
|
| 186 |
-
"restart_behavior": "overwrite",
|
| 187 |
-
"reward_clip": 1000.0,
|
| 188 |
-
"reward_scale": 1.0,
|
| 189 |
-
"rnn_num_layers": 1,
|
| 190 |
-
"rnn_size": 512,
|
| 191 |
-
"rnn_type": "gru",
|
| 192 |
-
"rollout": 128,
|
| 193 |
-
"save_best_after": 100000,
|
| 194 |
-
"save_best_every_sec": 5,
|
| 195 |
-
"save_best_metric": "reward",
|
| 196 |
-
"save_every_sec": 600,
|
| 197 |
-
"save_milestones_sec": -1,
|
| 198 |
-
"screen_size": 84,
|
| 199 |
-
"seed": 0,
|
| 200 |
-
"serial_mode": false,
|
| 201 |
-
"set_workers_cpu_affinity": true,
|
| 202 |
-
"shuffle_minibatches": false,
|
| 203 |
-
"stats_avg": 100,
|
| 204 |
-
"summaries_use_frameskip": true,
|
| 205 |
-
"terminal_on_life_loss": true,
|
| 206 |
-
"train_dir": "results/checkpoints",
|
| 207 |
-
"train_for_env_steps": 50000000,
|
| 208 |
-
"train_for_seconds": 10000000000,
|
| 209 |
-
"use_env_info_cache": false,
|
| 210 |
-
"use_record_episode_statistics": false,
|
| 211 |
-
"use_rnn": false,
|
| 212 |
-
"value_bootstrap": false,
|
| 213 |
-
"value_loss_coeff": 0.5,
|
| 214 |
-
"vtrace_c": 1.0,
|
| 215 |
-
"vtrace_rho": 1.0,
|
| 216 |
-
"wandb_group": "demon_attack",
|
| 217 |
-
"wandb_job_type": "sample_factory",
|
| 218 |
-
"wandb_project": "latency-sensitive-bench",
|
| 219 |
-
"wandb_tags": [],
|
| 220 |
-
"wandb_unique_id": "demon-attack-l2-s0",
|
| 221 |
-
"wandb_user": null,
|
| 222 |
-
"with_pbt": false,
|
| 223 |
-
"with_vtrace": false,
|
| 224 |
-
"with_wandb": true,
|
| 225 |
-
"within_episode_change_interval": 1,
|
| 226 |
-
"within_episode_change_prob": 1.0,
|
| 227 |
-
"worker_num_splits": 2
|
| 228 |
-
},
|
| 229 |
-
"checkpoint_experiment_dir": "/inspire/hdd/global_user/liumingyu-253208120284/lzj/latency-sensitive-bench/checkpoints/demon_attack_sf_same_latency_l2_seed0",
|
| 230 |
-
"checkpoint_path": null,
|
| 231 |
-
"checkpoint_train_step": 97664,
|
| 232 |
-
"config_source": "/inspire/hdd/global_user/liumingyu-253208120284/lzj/latency-sensitive-bench/checkpoints/demon_attack_sf_same_latency_l2_seed0/config.json",
|
| 233 |
-
"env_fps": 60.0,
|
| 234 |
-
"env_name": "demon_attack",
|
| 235 |
-
"episode_return_field": "episode_raw_return",
|
| 236 |
-
"filter_preset": "clean_v1",
|
| 237 |
-
"filter_report": {
|
| 238 |
-
"accepted_episodes": 200,
|
| 239 |
-
"accepted_train_episodes": 180,
|
| 240 |
-
"accepted_val_episodes": 20,
|
| 241 |
-
"attempted_episodes": 904,
|
| 242 |
-
"preset": "clean_v1",
|
| 243 |
-
"raw_score_bins": {
|
| 244 |
-
"6460-6889": 0,
|
| 245 |
-
"6890-7309": 0,
|
| 246 |
-
"7310-7999": 128,
|
| 247 |
-
"<6460": 0,
|
| 248 |
-
">=8000": 72
|
| 249 |
-
},
|
| 250 |
-
"rejected_episodes": 704,
|
| 251 |
-
"train_rows_after_row_filter": 349558,
|
| 252 |
-
"train_rows_before_row_filter": 611368,
|
| 253 |
-
"val_rows_after_row_filter": 52326,
|
| 254 |
-
"val_rows_before_row_filter": 69880
|
| 255 |
-
},
|
| 256 |
-
"frame_stack": 4,
|
| 257 |
-
"image_resolution": [
|
| 258 |
-
210,
|
| 259 |
-
160,
|
| 260 |
-
3
|
| 261 |
-
],
|
| 262 |
-
"image_source": "env_raw_rgb",
|
| 263 |
-
"latency_override": {
|
| 264 |
-
"add_latency_info": false,
|
| 265 |
-
"fixed_latency_ms": 133.33333333333334,
|
| 266 |
-
"type": "fixed"
|
| 267 |
-
},
|
| 268 |
-
"latency_source": "legacy_checkpoint_name",
|
| 269 |
-
"max_episodes": 200,
|
| 270 |
-
"max_steps_per_episode": 3600,
|
| 271 |
-
"num_envs": 64,
|
| 272 |
-
"obs_fps": 15.0,
|
| 273 |
-
"obs_stride_raw_frames": 4,
|
| 274 |
-
"prompt_source": "latency_bench.data.prompts",
|
| 275 |
-
"prompt_timing_suffix": " Current action latency is {latency_raw_frames} raw frames ({latency_ms:.2f} ms). The environment runs at {env_fps:g} FPS and observations are emitted at {obs_fps:g} FPS. Choose the best next action.",
|
| 276 |
-
"replay_num_envs": 16,
|
| 277 |
-
"reward_field": "raw_reward",
|
| 278 |
-
"reward_semantics": "raw_reward",
|
| 279 |
-
"rows_unit": "decision_step",
|
| 280 |
-
"schema_version": 2,
|
| 281 |
-
"split_seed": 0,
|
| 282 |
-
"total_episodes": 200,
|
| 283 |
-
"total_rows": 401884,
|
| 284 |
-
"train_episodes": 180,
|
| 285 |
-
"train_rows": 349558,
|
| 286 |
-
"val_episodes": 20,
|
| 287 |
-
"val_fraction": 0.1,
|
| 288 |
-
"val_rows": 52326
|
| 289 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_2_200ep/train.parquet
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:ed5ffa693f9dbbabc448992bd5b9cafcfc1c94fa7e173b0b3653e27d3699c552
|
| 3 |
-
size 204947383
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_2_200ep/val.parquet
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:78a1a4c69f601d9333ae092f69254233d66c48f39b804d0fbed619926f45771f
|
| 3 |
-
size 30312545
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_4_200ep/filter_report.json
DELETED
|
@@ -1,19 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"accepted_episodes": 200,
|
| 3 |
-
"accepted_train_episodes": 180,
|
| 4 |
-
"accepted_val_episodes": 20,
|
| 5 |
-
"attempted_episodes": 1676,
|
| 6 |
-
"preset": "clean_v1",
|
| 7 |
-
"raw_score_bins": {
|
| 8 |
-
"6460-6889": 0,
|
| 9 |
-
"6890-7309": 0,
|
| 10 |
-
"7310-7999": 0,
|
| 11 |
-
"<6460": 200,
|
| 12 |
-
">=8000": 0
|
| 13 |
-
},
|
| 14 |
-
"rejected_episodes": 1476,
|
| 15 |
-
"train_rows_after_row_filter": 332733,
|
| 16 |
-
"train_rows_before_row_filter": 644374,
|
| 17 |
-
"val_rows_after_row_filter": 60809,
|
| 18 |
-
"val_rows_before_row_filter": 71689
|
| 19 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_4_200ep/metadata.json
DELETED
|
@@ -1,289 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"action_labels": [
|
| 3 |
-
"NOOP",
|
| 4 |
-
"FIRE",
|
| 5 |
-
"RIGHT",
|
| 6 |
-
"LEFT",
|
| 7 |
-
"RIGHTFIRE",
|
| 8 |
-
"LEFTFIRE"
|
| 9 |
-
],
|
| 10 |
-
"action_layout": "demon_attack_discrete",
|
| 11 |
-
"base_prompt": "You are playing Demon Attack from a single game image. Choose exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, LEFTFIRE. Survive, avoid enemy fire, and shoot enemies to maximize score.",
|
| 12 |
-
"checkpoint_config": {
|
| 13 |
-
"action_repeat": 4,
|
| 14 |
-
"actor_critic_share_weights": true,
|
| 15 |
-
"actor_worker_gpus": [],
|
| 16 |
-
"adam_beta1": 0.9,
|
| 17 |
-
"adam_beta2": 0.999,
|
| 18 |
-
"adam_eps": 1e-05,
|
| 19 |
-
"adaptive_stddev": true,
|
| 20 |
-
"add_latency_info": false,
|
| 21 |
-
"algo": "APPO",
|
| 22 |
-
"async_rl": true,
|
| 23 |
-
"batch_size": 1024,
|
| 24 |
-
"batched_sampling": true,
|
| 25 |
-
"benchmark": false,
|
| 26 |
-
"cli_args": {
|
| 27 |
-
"action_repeat": 4,
|
| 28 |
-
"add_latency_info": false,
|
| 29 |
-
"algo": "APPO",
|
| 30 |
-
"async_rl": true,
|
| 31 |
-
"batch_size": 1024,
|
| 32 |
-
"batched_sampling": true,
|
| 33 |
-
"clip_reward": true,
|
| 34 |
-
"device": "gpu",
|
| 35 |
-
"env": "latency_demon_attack",
|
| 36 |
-
"env_id": "ALE/DemonAttack-v5",
|
| 37 |
-
"episode_metrics_path": "results/checkpoints/demon_attack_sf_same_latency_l4_seed0/episode_metrics.jsonl",
|
| 38 |
-
"episodes_per_stage": 10,
|
| 39 |
-
"eval_deterministic": true,
|
| 40 |
-
"eval_episodes": 20,
|
| 41 |
-
"eval_max_steps": 3600,
|
| 42 |
-
"eval_queue_latency_values": "0,1,2,3,4,5,10,15",
|
| 43 |
-
"eval_raw_reward": true,
|
| 44 |
-
"experiment": "demon_attack_sf_same_latency_l4_seed0",
|
| 45 |
-
"experiment_summaries_interval": 1,
|
| 46 |
-
"frame_stack": 4,
|
| 47 |
-
"full_action_space": false,
|
| 48 |
-
"gae_lambda": 0.95,
|
| 49 |
-
"gamma": 0.99,
|
| 50 |
-
"keep_checkpoints": 5,
|
| 51 |
-
"latency_mode": "episode_fixed",
|
| 52 |
-
"latency_sampling_mode": "uniform_all",
|
| 53 |
-
"learning_rate": 0.000303,
|
| 54 |
-
"max_episode_steps": 25000,
|
| 55 |
-
"max_grad_norm": 0.0,
|
| 56 |
-
"mode": "train",
|
| 57 |
-
"noop_max": 30,
|
| 58 |
-
"normalize_input": true,
|
| 59 |
-
"normalize_returns": true,
|
| 60 |
-
"num_batches_per_epoch": 2,
|
| 61 |
-
"num_envs_per_worker": 4,
|
| 62 |
-
"num_epochs": 2,
|
| 63 |
-
"num_policies": 1,
|
| 64 |
-
"num_workers": 16,
|
| 65 |
-
"ppo_clip_ratio": 0.1,
|
| 66 |
-
"ppo_clip_value": 0.2,
|
| 67 |
-
"queue_latency_values": "4",
|
| 68 |
-
"randomize_within_stage": false,
|
| 69 |
-
"recurrence": 32,
|
| 70 |
-
"repeat_action_probability": 0.0,
|
| 71 |
-
"restart_behavior": "overwrite",
|
| 72 |
-
"rollout": 128,
|
| 73 |
-
"save_every_sec": 600,
|
| 74 |
-
"screen_size": 84,
|
| 75 |
-
"seed": 0,
|
| 76 |
-
"stats_avg": 100,
|
| 77 |
-
"terminal_on_life_loss": true,
|
| 78 |
-
"train_dir": "results/checkpoints",
|
| 79 |
-
"train_for_env_steps": 50000000,
|
| 80 |
-
"use_rnn": false,
|
| 81 |
-
"value_loss_coeff": 0.5,
|
| 82 |
-
"wandb_group": "demon_attack",
|
| 83 |
-
"wandb_job_type": "sample_factory",
|
| 84 |
-
"wandb_project": "latency-sensitive-bench",
|
| 85 |
-
"with_wandb": true,
|
| 86 |
-
"within_episode_change_interval": 1,
|
| 87 |
-
"within_episode_change_prob": 1.0,
|
| 88 |
-
"worker_num_splits": 2
|
| 89 |
-
},
|
| 90 |
-
"clip_reward": true,
|
| 91 |
-
"command_line": "--mode train --algo APPO --env latency_demon_attack --experiment demon_attack_sf_same_latency_l4_seed0 --train_dir results/checkpoints --restart_behavior overwrite --device gpu --seed 0 --episode_metrics_path results/checkpoints/demon_attack_sf_same_latency_l4_seed0/episode_metrics.jsonl --train_for_env_steps 50000000 --num_workers 16 --num_envs_per_worker 4 --num_policies 1 --batch_size 1024 --rollout 128 --recurrence 32 --num_epochs 2 --num_batches_per_epoch 2 --worker_num_splits 2 --learning_rate 0.000303 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.0 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --async_rl True --batched_sampling True --use_rnn False --normalize_returns True --normalize_input True --queue-latency-values 4 --latency-mode episode_fixed --latency-sampling-mode uniform_all --episodes-per-stage 10 --randomize-within-stage False --within-episode-change-interval 1 --within-episode-change-prob 1.0 --add-latency-info False --eval-episodes 20 --eval-queue-latency-values 0,1,2,3,4,5,10,15 --eval-max-steps 3600 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project latency-sensitive-bench --wandb_group demon_attack --wandb_job_type sample_factory --env_id ALE/DemonAttack-v5 --screen_size 84 --noop_max 30 --action_repeat 4 --frame_stack 4 --terminal_on_life_loss True --clip_reward True --repeat_action_probability 0.0 --full_action_space False --max_episode_steps 25000",
|
| 92 |
-
"continuous_tanh_scale": 0.0,
|
| 93 |
-
"decoder_mlp_layers": [],
|
| 94 |
-
"decorrelate_envs_on_one_worker": true,
|
| 95 |
-
"decorrelate_experience_max_seconds": 0,
|
| 96 |
-
"default_niceness": 0,
|
| 97 |
-
"device": "gpu",
|
| 98 |
-
"encoder_conv_architecture": "convnet_atari",
|
| 99 |
-
"encoder_conv_mlp_layers": [
|
| 100 |
-
512
|
| 101 |
-
],
|
| 102 |
-
"encoder_mlp_layers": [
|
| 103 |
-
512,
|
| 104 |
-
512
|
| 105 |
-
],
|
| 106 |
-
"env": "latency_demon_attack",
|
| 107 |
-
"env_frameskip": 1,
|
| 108 |
-
"env_framestack": 1,
|
| 109 |
-
"env_gpu_actions": false,
|
| 110 |
-
"env_gpu_observations": true,
|
| 111 |
-
"env_id": "ALE/DemonAttack-v5",
|
| 112 |
-
"episode_metrics_path": "results/checkpoints/demon_attack_sf_same_latency_l4_seed0/episode_metrics.jsonl",
|
| 113 |
-
"episodes_per_stage": 10,
|
| 114 |
-
"eval_deterministic": true,
|
| 115 |
-
"eval_episodes": 20,
|
| 116 |
-
"eval_max_steps": 3600,
|
| 117 |
-
"eval_queue_latency_values": "0,1,2,3,4,5,10,15",
|
| 118 |
-
"eval_raw_reward": true,
|
| 119 |
-
"experiment": "demon_attack_sf_same_latency_l4_seed0",
|
| 120 |
-
"experiment_summaries_interval": 1,
|
| 121 |
-
"exploration_loss": "entropy",
|
| 122 |
-
"exploration_loss_coeff": 0.003,
|
| 123 |
-
"flush_summaries_interval": 30,
|
| 124 |
-
"force_envs_single_thread": false,
|
| 125 |
-
"frame_stack": 4,
|
| 126 |
-
"full_action_space": false,
|
| 127 |
-
"gae_lambda": 0.95,
|
| 128 |
-
"gamma": 0.99,
|
| 129 |
-
"git_hash": "41bc95edbf670a1190d128da22e9cd3777b87989",
|
| 130 |
-
"git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
|
| 131 |
-
"heartbeat_interval": 20,
|
| 132 |
-
"heartbeat_reporting_interval": 180,
|
| 133 |
-
"help": false,
|
| 134 |
-
"initial_stddev": 1.0,
|
| 135 |
-
"keep_checkpoints": 5,
|
| 136 |
-
"kl_loss_coeff": 0.0,
|
| 137 |
-
"latency_mode": "episode_fixed",
|
| 138 |
-
"latency_sampling_mode": "uniform_all",
|
| 139 |
-
"learning_rate": 0.000303,
|
| 140 |
-
"load_checkpoint_kind": "latest",
|
| 141 |
-
"log_to_file": true,
|
| 142 |
-
"lr_adaptive_max": 0.01,
|
| 143 |
-
"lr_adaptive_min": 1e-06,
|
| 144 |
-
"lr_schedule": "linear_decay",
|
| 145 |
-
"lr_schedule_kl_threshold": 0.008,
|
| 146 |
-
"max_episode_steps": 25000,
|
| 147 |
-
"max_grad_norm": 0.0,
|
| 148 |
-
"max_policy_lag": 1000,
|
| 149 |
-
"mode": "train",
|
| 150 |
-
"nonlinearity": "elu",
|
| 151 |
-
"noop_max": 30,
|
| 152 |
-
"normalize_input": true,
|
| 153 |
-
"normalize_input_keys": null,
|
| 154 |
-
"normalize_returns": true,
|
| 155 |
-
"num_batches_per_epoch": 2,
|
| 156 |
-
"num_batches_to_accumulate": 2,
|
| 157 |
-
"num_envs_per_worker": 4,
|
| 158 |
-
"num_epochs": 2,
|
| 159 |
-
"num_policies": 1,
|
| 160 |
-
"num_workers": 16,
|
| 161 |
-
"obs_scale": 255.0,
|
| 162 |
-
"obs_subtract_mean": 0.0,
|
| 163 |
-
"optimizer": "adam",
|
| 164 |
-
"output_dir": "outputs/train/demon_attack/latency_4/seed_0",
|
| 165 |
-
"pbt_mix_policies_in_one_env": true,
|
| 166 |
-
"pbt_mutation_rate": 0.15,
|
| 167 |
-
"pbt_optimize_gamma": false,
|
| 168 |
-
"pbt_period_env_steps": 5000000,
|
| 169 |
-
"pbt_perturb_max": 1.5,
|
| 170 |
-
"pbt_perturb_min": 1.1,
|
| 171 |
-
"pbt_replace_fraction": 0.3,
|
| 172 |
-
"pbt_replace_reward_gap": 0.1,
|
| 173 |
-
"pbt_replace_reward_gap_absolute": 1e-06,
|
| 174 |
-
"pbt_start_mutation": 20000000,
|
| 175 |
-
"pbt_target_objective": "true_objective",
|
| 176 |
-
"pixel_format": "CHW",
|
| 177 |
-
"policy_init_gain": 1.0,
|
| 178 |
-
"policy_initialization": "orthogonal",
|
| 179 |
-
"policy_workers_per_policy": 1,
|
| 180 |
-
"ppo_clip_ratio": 0.1,
|
| 181 |
-
"ppo_clip_value": 0.2,
|
| 182 |
-
"queue_latency_values": "4",
|
| 183 |
-
"randomize_within_stage": false,
|
| 184 |
-
"recurrence": 32,
|
| 185 |
-
"repeat_action_probability": 0.0,
|
| 186 |
-
"restart_behavior": "overwrite",
|
| 187 |
-
"reward_clip": 1000.0,
|
| 188 |
-
"reward_scale": 1.0,
|
| 189 |
-
"rnn_num_layers": 1,
|
| 190 |
-
"rnn_size": 512,
|
| 191 |
-
"rnn_type": "gru",
|
| 192 |
-
"rollout": 128,
|
| 193 |
-
"save_best_after": 100000,
|
| 194 |
-
"save_best_every_sec": 5,
|
| 195 |
-
"save_best_metric": "reward",
|
| 196 |
-
"save_every_sec": 600,
|
| 197 |
-
"save_milestones_sec": -1,
|
| 198 |
-
"screen_size": 84,
|
| 199 |
-
"seed": 0,
|
| 200 |
-
"serial_mode": false,
|
| 201 |
-
"set_workers_cpu_affinity": true,
|
| 202 |
-
"shuffle_minibatches": false,
|
| 203 |
-
"stats_avg": 100,
|
| 204 |
-
"summaries_use_frameskip": true,
|
| 205 |
-
"terminal_on_life_loss": true,
|
| 206 |
-
"train_dir": "results/checkpoints",
|
| 207 |
-
"train_for_env_steps": 50000000,
|
| 208 |
-
"train_for_seconds": 10000000000,
|
| 209 |
-
"use_env_info_cache": false,
|
| 210 |
-
"use_record_episode_statistics": false,
|
| 211 |
-
"use_rnn": false,
|
| 212 |
-
"value_bootstrap": false,
|
| 213 |
-
"value_loss_coeff": 0.5,
|
| 214 |
-
"vtrace_c": 1.0,
|
| 215 |
-
"vtrace_rho": 1.0,
|
| 216 |
-
"wandb_group": "demon_attack",
|
| 217 |
-
"wandb_job_type": "sample_factory",
|
| 218 |
-
"wandb_project": "latency-sensitive-bench",
|
| 219 |
-
"wandb_tags": [],
|
| 220 |
-
"wandb_unique_id": "demon-attack-l4-s0",
|
| 221 |
-
"wandb_user": null,
|
| 222 |
-
"with_pbt": false,
|
| 223 |
-
"with_vtrace": false,
|
| 224 |
-
"with_wandb": true,
|
| 225 |
-
"within_episode_change_interval": 1,
|
| 226 |
-
"within_episode_change_prob": 1.0,
|
| 227 |
-
"worker_num_splits": 2
|
| 228 |
-
},
|
| 229 |
-
"checkpoint_experiment_dir": "/inspire/hdd/global_user/liumingyu-253208120284/lzj/latency-sensitive-bench/checkpoints/demon_attack_sf_same_latency_l4_seed0",
|
| 230 |
-
"checkpoint_path": null,
|
| 231 |
-
"checkpoint_train_step": 97664,
|
| 232 |
-
"config_source": "/inspire/hdd/global_user/liumingyu-253208120284/lzj/latency-sensitive-bench/checkpoints/demon_attack_sf_same_latency_l4_seed0/config.json",
|
| 233 |
-
"env_fps": 60.0,
|
| 234 |
-
"env_name": "demon_attack",
|
| 235 |
-
"episode_return_field": "episode_raw_return",
|
| 236 |
-
"filter_preset": "clean_v1",
|
| 237 |
-
"filter_report": {
|
| 238 |
-
"accepted_episodes": 200,
|
| 239 |
-
"accepted_train_episodes": 180,
|
| 240 |
-
"accepted_val_episodes": 20,
|
| 241 |
-
"attempted_episodes": 1676,
|
| 242 |
-
"preset": "clean_v1",
|
| 243 |
-
"raw_score_bins": {
|
| 244 |
-
"6460-6889": 0,
|
| 245 |
-
"6890-7309": 0,
|
| 246 |
-
"7310-7999": 0,
|
| 247 |
-
"<6460": 200,
|
| 248 |
-
">=8000": 0
|
| 249 |
-
},
|
| 250 |
-
"rejected_episodes": 1476,
|
| 251 |
-
"train_rows_after_row_filter": 332733,
|
| 252 |
-
"train_rows_before_row_filter": 644374,
|
| 253 |
-
"val_rows_after_row_filter": 60809,
|
| 254 |
-
"val_rows_before_row_filter": 71689
|
| 255 |
-
},
|
| 256 |
-
"frame_stack": 4,
|
| 257 |
-
"image_resolution": [
|
| 258 |
-
210,
|
| 259 |
-
160,
|
| 260 |
-
3
|
| 261 |
-
],
|
| 262 |
-
"image_source": "env_raw_rgb",
|
| 263 |
-
"latency_override": {
|
| 264 |
-
"add_latency_info": false,
|
| 265 |
-
"fixed_latency_ms": 266.6666666666667,
|
| 266 |
-
"type": "fixed"
|
| 267 |
-
},
|
| 268 |
-
"latency_source": "legacy_checkpoint_name",
|
| 269 |
-
"max_episodes": 200,
|
| 270 |
-
"max_steps_per_episode": 3600,
|
| 271 |
-
"num_envs": 128,
|
| 272 |
-
"obs_fps": 15.0,
|
| 273 |
-
"obs_stride_raw_frames": 4,
|
| 274 |
-
"prompt_source": "latency_bench.data.prompts",
|
| 275 |
-
"prompt_timing_suffix": " Current action latency is {latency_raw_frames} raw frames ({latency_ms:.2f} ms). The environment runs at {env_fps:g} FPS and observations are emitted at {obs_fps:g} FPS. Choose the best next action.",
|
| 276 |
-
"replay_num_envs": 16,
|
| 277 |
-
"reward_field": "raw_reward",
|
| 278 |
-
"reward_semantics": "raw_reward",
|
| 279 |
-
"rows_unit": "decision_step",
|
| 280 |
-
"schema_version": 2,
|
| 281 |
-
"split_seed": 0,
|
| 282 |
-
"total_episodes": 200,
|
| 283 |
-
"total_rows": 393542,
|
| 284 |
-
"train_episodes": 180,
|
| 285 |
-
"train_rows": 332733,
|
| 286 |
-
"val_episodes": 20,
|
| 287 |
-
"val_fraction": 0.1,
|
| 288 |
-
"val_rows": 60809
|
| 289 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_4_200ep/train.parquet
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:cbc9ea8ecc830e2f15042ef12ac52577c8db04d593047aec1e30e7f110f933a1
|
| 3 |
-
size 194112105
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_4_200ep/val.parquet
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:8ea8cd47267494ef55f4d5e10595b5b4b2e73a51992f2fdcc910687c50fb6507
|
| 3 |
-
size 35026048
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_6_200ep/filter_report.json
DELETED
|
@@ -1,19 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"accepted_episodes": 200,
|
| 3 |
-
"accepted_train_episodes": 180,
|
| 4 |
-
"accepted_val_episodes": 20,
|
| 5 |
-
"attempted_episodes": 1358,
|
| 6 |
-
"preset": "clean_v1",
|
| 7 |
-
"raw_score_bins": {
|
| 8 |
-
"6460-6889": 0,
|
| 9 |
-
"6890-7309": 0,
|
| 10 |
-
"7310-7999": 0,
|
| 11 |
-
"<6460": 200,
|
| 12 |
-
">=8000": 0
|
| 13 |
-
},
|
| 14 |
-
"rejected_episodes": 1158,
|
| 15 |
-
"train_rows_after_row_filter": 310953,
|
| 16 |
-
"train_rows_before_row_filter": 619555,
|
| 17 |
-
"val_rows_after_row_filter": 54672,
|
| 18 |
-
"val_rows_before_row_filter": 68190
|
| 19 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_6_200ep/metadata.json
DELETED
|
@@ -1,289 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"action_labels": [
|
| 3 |
-
"NOOP",
|
| 4 |
-
"FIRE",
|
| 5 |
-
"RIGHT",
|
| 6 |
-
"LEFT",
|
| 7 |
-
"RIGHTFIRE",
|
| 8 |
-
"LEFTFIRE"
|
| 9 |
-
],
|
| 10 |
-
"action_layout": "demon_attack_discrete",
|
| 11 |
-
"base_prompt": "You are playing Demon Attack from a single game image. Choose exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, LEFTFIRE. Survive, avoid enemy fire, and shoot enemies to maximize score.",
|
| 12 |
-
"checkpoint_config": {
|
| 13 |
-
"action_repeat": 4,
|
| 14 |
-
"actor_critic_share_weights": true,
|
| 15 |
-
"actor_worker_gpus": [],
|
| 16 |
-
"adam_beta1": 0.9,
|
| 17 |
-
"adam_beta2": 0.999,
|
| 18 |
-
"adam_eps": 1e-05,
|
| 19 |
-
"adaptive_stddev": true,
|
| 20 |
-
"add_latency_info": false,
|
| 21 |
-
"algo": "APPO",
|
| 22 |
-
"async_rl": true,
|
| 23 |
-
"batch_size": 1024,
|
| 24 |
-
"batched_sampling": true,
|
| 25 |
-
"benchmark": false,
|
| 26 |
-
"cli_args": {
|
| 27 |
-
"action_repeat": 4,
|
| 28 |
-
"add_latency_info": false,
|
| 29 |
-
"algo": "APPO",
|
| 30 |
-
"async_rl": true,
|
| 31 |
-
"batch_size": 1024,
|
| 32 |
-
"batched_sampling": true,
|
| 33 |
-
"clip_reward": true,
|
| 34 |
-
"device": "gpu",
|
| 35 |
-
"env": "latency_demon_attack",
|
| 36 |
-
"env_id": "ALE/DemonAttack-v5",
|
| 37 |
-
"episode_metrics_path": "results/checkpoints/demon_attack_sf_same_latency_l6_seed0/episode_metrics.jsonl",
|
| 38 |
-
"episodes_per_stage": 10,
|
| 39 |
-
"eval_deterministic": true,
|
| 40 |
-
"eval_episodes": 20,
|
| 41 |
-
"eval_max_steps": 3600,
|
| 42 |
-
"eval_queue_latency_values": "0,1,2,3,4,5,10,15",
|
| 43 |
-
"eval_raw_reward": true,
|
| 44 |
-
"experiment": "demon_attack_sf_same_latency_l6_seed0",
|
| 45 |
-
"experiment_summaries_interval": 1,
|
| 46 |
-
"frame_stack": 4,
|
| 47 |
-
"full_action_space": false,
|
| 48 |
-
"gae_lambda": 0.95,
|
| 49 |
-
"gamma": 0.99,
|
| 50 |
-
"keep_checkpoints": 5,
|
| 51 |
-
"latency_mode": "episode_fixed",
|
| 52 |
-
"latency_sampling_mode": "uniform_all",
|
| 53 |
-
"learning_rate": 0.000303,
|
| 54 |
-
"max_episode_steps": 25000,
|
| 55 |
-
"max_grad_norm": 0.0,
|
| 56 |
-
"mode": "train",
|
| 57 |
-
"noop_max": 30,
|
| 58 |
-
"normalize_input": true,
|
| 59 |
-
"normalize_returns": true,
|
| 60 |
-
"num_batches_per_epoch": 2,
|
| 61 |
-
"num_envs_per_worker": 4,
|
| 62 |
-
"num_epochs": 2,
|
| 63 |
-
"num_policies": 1,
|
| 64 |
-
"num_workers": 16,
|
| 65 |
-
"ppo_clip_ratio": 0.1,
|
| 66 |
-
"ppo_clip_value": 0.2,
|
| 67 |
-
"queue_latency_values": "6",
|
| 68 |
-
"randomize_within_stage": false,
|
| 69 |
-
"recurrence": 32,
|
| 70 |
-
"repeat_action_probability": 0.0,
|
| 71 |
-
"restart_behavior": "overwrite",
|
| 72 |
-
"rollout": 128,
|
| 73 |
-
"save_every_sec": 600,
|
| 74 |
-
"screen_size": 84,
|
| 75 |
-
"seed": 0,
|
| 76 |
-
"stats_avg": 100,
|
| 77 |
-
"terminal_on_life_loss": true,
|
| 78 |
-
"train_dir": "results/checkpoints",
|
| 79 |
-
"train_for_env_steps": 50000000,
|
| 80 |
-
"use_rnn": false,
|
| 81 |
-
"value_loss_coeff": 0.5,
|
| 82 |
-
"wandb_group": "demon_attack",
|
| 83 |
-
"wandb_job_type": "sample_factory",
|
| 84 |
-
"wandb_project": "latency-sensitive-bench",
|
| 85 |
-
"with_wandb": true,
|
| 86 |
-
"within_episode_change_interval": 1,
|
| 87 |
-
"within_episode_change_prob": 1.0,
|
| 88 |
-
"worker_num_splits": 2
|
| 89 |
-
},
|
| 90 |
-
"clip_reward": true,
|
| 91 |
-
"command_line": "--mode train --algo APPO --env latency_demon_attack --experiment demon_attack_sf_same_latency_l6_seed0 --train_dir results/checkpoints --restart_behavior overwrite --device gpu --seed 0 --episode_metrics_path results/checkpoints/demon_attack_sf_same_latency_l6_seed0/episode_metrics.jsonl --train_for_env_steps 50000000 --num_workers 16 --num_envs_per_worker 4 --num_policies 1 --batch_size 1024 --rollout 128 --recurrence 32 --num_epochs 2 --num_batches_per_epoch 2 --worker_num_splits 2 --learning_rate 0.000303 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.0 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --async_rl True --batched_sampling True --use_rnn False --normalize_returns True --normalize_input True --queue-latency-values 6 --latency-mode episode_fixed --latency-sampling-mode uniform_all --episodes-per-stage 10 --randomize-within-stage False --within-episode-change-interval 1 --within-episode-change-prob 1.0 --add-latency-info False --eval-episodes 20 --eval-queue-latency-values 0,1,2,3,4,5,10,15 --eval-max-steps 3600 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project latency-sensitive-bench --wandb_group demon_attack --wandb_job_type sample_factory --env_id ALE/DemonAttack-v5 --screen_size 84 --noop_max 30 --action_repeat 4 --frame_stack 4 --terminal_on_life_loss True --clip_reward True --repeat_action_probability 0.0 --full_action_space False --max_episode_steps 25000",
|
| 92 |
-
"continuous_tanh_scale": 0.0,
|
| 93 |
-
"decoder_mlp_layers": [],
|
| 94 |
-
"decorrelate_envs_on_one_worker": true,
|
| 95 |
-
"decorrelate_experience_max_seconds": 0,
|
| 96 |
-
"default_niceness": 0,
|
| 97 |
-
"device": "gpu",
|
| 98 |
-
"encoder_conv_architecture": "convnet_atari",
|
| 99 |
-
"encoder_conv_mlp_layers": [
|
| 100 |
-
512
|
| 101 |
-
],
|
| 102 |
-
"encoder_mlp_layers": [
|
| 103 |
-
512,
|
| 104 |
-
512
|
| 105 |
-
],
|
| 106 |
-
"env": "latency_demon_attack",
|
| 107 |
-
"env_frameskip": 1,
|
| 108 |
-
"env_framestack": 1,
|
| 109 |
-
"env_gpu_actions": false,
|
| 110 |
-
"env_gpu_observations": true,
|
| 111 |
-
"env_id": "ALE/DemonAttack-v5",
|
| 112 |
-
"episode_metrics_path": "results/checkpoints/demon_attack_sf_same_latency_l6_seed0/episode_metrics.jsonl",
|
| 113 |
-
"episodes_per_stage": 10,
|
| 114 |
-
"eval_deterministic": true,
|
| 115 |
-
"eval_episodes": 20,
|
| 116 |
-
"eval_max_steps": 3600,
|
| 117 |
-
"eval_queue_latency_values": "0,1,2,3,4,5,10,15",
|
| 118 |
-
"eval_raw_reward": true,
|
| 119 |
-
"experiment": "demon_attack_sf_same_latency_l6_seed0",
|
| 120 |
-
"experiment_summaries_interval": 1,
|
| 121 |
-
"exploration_loss": "entropy",
|
| 122 |
-
"exploration_loss_coeff": 0.003,
|
| 123 |
-
"flush_summaries_interval": 30,
|
| 124 |
-
"force_envs_single_thread": false,
|
| 125 |
-
"frame_stack": 4,
|
| 126 |
-
"full_action_space": false,
|
| 127 |
-
"gae_lambda": 0.95,
|
| 128 |
-
"gamma": 0.99,
|
| 129 |
-
"git_hash": "41bc95edbf670a1190d128da22e9cd3777b87989",
|
| 130 |
-
"git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
|
| 131 |
-
"heartbeat_interval": 20,
|
| 132 |
-
"heartbeat_reporting_interval": 180,
|
| 133 |
-
"help": false,
|
| 134 |
-
"initial_stddev": 1.0,
|
| 135 |
-
"keep_checkpoints": 5,
|
| 136 |
-
"kl_loss_coeff": 0.0,
|
| 137 |
-
"latency_mode": "episode_fixed",
|
| 138 |
-
"latency_sampling_mode": "uniform_all",
|
| 139 |
-
"learning_rate": 0.000303,
|
| 140 |
-
"load_checkpoint_kind": "latest",
|
| 141 |
-
"log_to_file": true,
|
| 142 |
-
"lr_adaptive_max": 0.01,
|
| 143 |
-
"lr_adaptive_min": 1e-06,
|
| 144 |
-
"lr_schedule": "linear_decay",
|
| 145 |
-
"lr_schedule_kl_threshold": 0.008,
|
| 146 |
-
"max_episode_steps": 25000,
|
| 147 |
-
"max_grad_norm": 0.0,
|
| 148 |
-
"max_policy_lag": 1000,
|
| 149 |
-
"mode": "train",
|
| 150 |
-
"nonlinearity": "elu",
|
| 151 |
-
"noop_max": 30,
|
| 152 |
-
"normalize_input": true,
|
| 153 |
-
"normalize_input_keys": null,
|
| 154 |
-
"normalize_returns": true,
|
| 155 |
-
"num_batches_per_epoch": 2,
|
| 156 |
-
"num_batches_to_accumulate": 2,
|
| 157 |
-
"num_envs_per_worker": 4,
|
| 158 |
-
"num_epochs": 2,
|
| 159 |
-
"num_policies": 1,
|
| 160 |
-
"num_workers": 16,
|
| 161 |
-
"obs_scale": 255.0,
|
| 162 |
-
"obs_subtract_mean": 0.0,
|
| 163 |
-
"optimizer": "adam",
|
| 164 |
-
"output_dir": "outputs/train/demon_attack/latency_6/seed_0",
|
| 165 |
-
"pbt_mix_policies_in_one_env": true,
|
| 166 |
-
"pbt_mutation_rate": 0.15,
|
| 167 |
-
"pbt_optimize_gamma": false,
|
| 168 |
-
"pbt_period_env_steps": 5000000,
|
| 169 |
-
"pbt_perturb_max": 1.5,
|
| 170 |
-
"pbt_perturb_min": 1.1,
|
| 171 |
-
"pbt_replace_fraction": 0.3,
|
| 172 |
-
"pbt_replace_reward_gap": 0.1,
|
| 173 |
-
"pbt_replace_reward_gap_absolute": 1e-06,
|
| 174 |
-
"pbt_start_mutation": 20000000,
|
| 175 |
-
"pbt_target_objective": "true_objective",
|
| 176 |
-
"pixel_format": "CHW",
|
| 177 |
-
"policy_init_gain": 1.0,
|
| 178 |
-
"policy_initialization": "orthogonal",
|
| 179 |
-
"policy_workers_per_policy": 1,
|
| 180 |
-
"ppo_clip_ratio": 0.1,
|
| 181 |
-
"ppo_clip_value": 0.2,
|
| 182 |
-
"queue_latency_values": "6",
|
| 183 |
-
"randomize_within_stage": false,
|
| 184 |
-
"recurrence": 32,
|
| 185 |
-
"repeat_action_probability": 0.0,
|
| 186 |
-
"restart_behavior": "overwrite",
|
| 187 |
-
"reward_clip": 1000.0,
|
| 188 |
-
"reward_scale": 1.0,
|
| 189 |
-
"rnn_num_layers": 1,
|
| 190 |
-
"rnn_size": 512,
|
| 191 |
-
"rnn_type": "gru",
|
| 192 |
-
"rollout": 128,
|
| 193 |
-
"save_best_after": 100000,
|
| 194 |
-
"save_best_every_sec": 5,
|
| 195 |
-
"save_best_metric": "reward",
|
| 196 |
-
"save_every_sec": 600,
|
| 197 |
-
"save_milestones_sec": -1,
|
| 198 |
-
"screen_size": 84,
|
| 199 |
-
"seed": 0,
|
| 200 |
-
"serial_mode": false,
|
| 201 |
-
"set_workers_cpu_affinity": true,
|
| 202 |
-
"shuffle_minibatches": false,
|
| 203 |
-
"stats_avg": 100,
|
| 204 |
-
"summaries_use_frameskip": true,
|
| 205 |
-
"terminal_on_life_loss": true,
|
| 206 |
-
"train_dir": "results/checkpoints",
|
| 207 |
-
"train_for_env_steps": 50000000,
|
| 208 |
-
"train_for_seconds": 10000000000,
|
| 209 |
-
"use_env_info_cache": false,
|
| 210 |
-
"use_record_episode_statistics": false,
|
| 211 |
-
"use_rnn": false,
|
| 212 |
-
"value_bootstrap": false,
|
| 213 |
-
"value_loss_coeff": 0.5,
|
| 214 |
-
"vtrace_c": 1.0,
|
| 215 |
-
"vtrace_rho": 1.0,
|
| 216 |
-
"wandb_group": "demon_attack",
|
| 217 |
-
"wandb_job_type": "sample_factory",
|
| 218 |
-
"wandb_project": "latency-sensitive-bench",
|
| 219 |
-
"wandb_tags": [],
|
| 220 |
-
"wandb_unique_id": "demon-attack-l6-s0",
|
| 221 |
-
"wandb_user": null,
|
| 222 |
-
"with_pbt": false,
|
| 223 |
-
"with_vtrace": false,
|
| 224 |
-
"with_wandb": true,
|
| 225 |
-
"within_episode_change_interval": 1,
|
| 226 |
-
"within_episode_change_prob": 1.0,
|
| 227 |
-
"worker_num_splits": 2
|
| 228 |
-
},
|
| 229 |
-
"checkpoint_experiment_dir": "/inspire/hdd/global_user/liumingyu-253208120284/lzj/latency-sensitive-bench/checkpoints/demon_attack_sf_same_latency_l6_seed0",
|
| 230 |
-
"checkpoint_path": null,
|
| 231 |
-
"checkpoint_train_step": 97664,
|
| 232 |
-
"config_source": "/inspire/hdd/global_user/liumingyu-253208120284/lzj/latency-sensitive-bench/checkpoints/demon_attack_sf_same_latency_l6_seed0/config.json",
|
| 233 |
-
"env_fps": 60.0,
|
| 234 |
-
"env_name": "demon_attack",
|
| 235 |
-
"episode_return_field": "episode_raw_return",
|
| 236 |
-
"filter_preset": "clean_v1",
|
| 237 |
-
"filter_report": {
|
| 238 |
-
"accepted_episodes": 200,
|
| 239 |
-
"accepted_train_episodes": 180,
|
| 240 |
-
"accepted_val_episodes": 20,
|
| 241 |
-
"attempted_episodes": 1358,
|
| 242 |
-
"preset": "clean_v1",
|
| 243 |
-
"raw_score_bins": {
|
| 244 |
-
"6460-6889": 0,
|
| 245 |
-
"6890-7309": 0,
|
| 246 |
-
"7310-7999": 0,
|
| 247 |
-
"<6460": 200,
|
| 248 |
-
">=8000": 0
|
| 249 |
-
},
|
| 250 |
-
"rejected_episodes": 1158,
|
| 251 |
-
"train_rows_after_row_filter": 310953,
|
| 252 |
-
"train_rows_before_row_filter": 619555,
|
| 253 |
-
"val_rows_after_row_filter": 54672,
|
| 254 |
-
"val_rows_before_row_filter": 68190
|
| 255 |
-
},
|
| 256 |
-
"frame_stack": 4,
|
| 257 |
-
"image_resolution": [
|
| 258 |
-
210,
|
| 259 |
-
160,
|
| 260 |
-
3
|
| 261 |
-
],
|
| 262 |
-
"image_source": "env_raw_rgb",
|
| 263 |
-
"latency_override": {
|
| 264 |
-
"add_latency_info": false,
|
| 265 |
-
"fixed_latency_ms": 400.0,
|
| 266 |
-
"type": "fixed"
|
| 267 |
-
},
|
| 268 |
-
"latency_source": "legacy_checkpoint_name",
|
| 269 |
-
"max_episodes": 200,
|
| 270 |
-
"max_steps_per_episode": 3600,
|
| 271 |
-
"num_envs": 128,
|
| 272 |
-
"obs_fps": 15.0,
|
| 273 |
-
"obs_stride_raw_frames": 4,
|
| 274 |
-
"prompt_source": "latency_bench.data.prompts",
|
| 275 |
-
"prompt_timing_suffix": " Current action latency is {latency_raw_frames} raw frames ({latency_ms:.2f} ms). The environment runs at {env_fps:g} FPS and observations are emitted at {obs_fps:g} FPS. Choose the best next action.",
|
| 276 |
-
"replay_num_envs": 16,
|
| 277 |
-
"reward_field": "raw_reward",
|
| 278 |
-
"reward_semantics": "raw_reward",
|
| 279 |
-
"rows_unit": "decision_step",
|
| 280 |
-
"schema_version": 2,
|
| 281 |
-
"split_seed": 0,
|
| 282 |
-
"total_episodes": 200,
|
| 283 |
-
"total_rows": 365625,
|
| 284 |
-
"train_episodes": 180,
|
| 285 |
-
"train_rows": 310953,
|
| 286 |
-
"val_episodes": 20,
|
| 287 |
-
"val_fraction": 0.1,
|
| 288 |
-
"val_rows": 54672
|
| 289 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_6_200ep/train.parquet
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:191ed66ae8a15c7767a507e474311176d34cd78fbd14dc4edd5307296b7a005a
|
| 3 |
-
size 176346244
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_6_200ep/val.parquet
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:1becbfeca092153c722db7d810ae5663e916761e4d143d22494acf3ad000cc5f
|
| 3 |
-
size 30637048
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_8_200ep/filter_report.json
DELETED
|
@@ -1,19 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"accepted_episodes": 200,
|
| 3 |
-
"accepted_train_episodes": 180,
|
| 4 |
-
"accepted_val_episodes": 20,
|
| 5 |
-
"attempted_episodes": 2304,
|
| 6 |
-
"preset": "clean_v1",
|
| 7 |
-
"raw_score_bins": {
|
| 8 |
-
"6460-6889": 0,
|
| 9 |
-
"6890-7309": 0,
|
| 10 |
-
"7310-7999": 0,
|
| 11 |
-
"<6460": 200,
|
| 12 |
-
">=8000": 0
|
| 13 |
-
},
|
| 14 |
-
"rejected_episodes": 2104,
|
| 15 |
-
"train_rows_after_row_filter": 282142,
|
| 16 |
-
"train_rows_before_row_filter": 646284,
|
| 17 |
-
"val_rows_after_row_filter": 40599,
|
| 18 |
-
"val_rows_before_row_filter": 71901
|
| 19 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_8_200ep/metadata.json
DELETED
|
@@ -1,289 +0,0 @@
|
|
| 1 |
-
{
|
| 2 |
-
"action_labels": [
|
| 3 |
-
"NOOP",
|
| 4 |
-
"FIRE",
|
| 5 |
-
"RIGHT",
|
| 6 |
-
"LEFT",
|
| 7 |
-
"RIGHTFIRE",
|
| 8 |
-
"LEFTFIRE"
|
| 9 |
-
],
|
| 10 |
-
"action_layout": "demon_attack_discrete",
|
| 11 |
-
"base_prompt": "You are playing Demon Attack from a single game image. Choose exactly one action from: NOOP, FIRE, RIGHT, LEFT, RIGHTFIRE, LEFTFIRE. Survive, avoid enemy fire, and shoot enemies to maximize score.",
|
| 12 |
-
"checkpoint_config": {
|
| 13 |
-
"action_repeat": 4,
|
| 14 |
-
"actor_critic_share_weights": true,
|
| 15 |
-
"actor_worker_gpus": [],
|
| 16 |
-
"adam_beta1": 0.9,
|
| 17 |
-
"adam_beta2": 0.999,
|
| 18 |
-
"adam_eps": 1e-05,
|
| 19 |
-
"adaptive_stddev": true,
|
| 20 |
-
"add_latency_info": false,
|
| 21 |
-
"algo": "APPO",
|
| 22 |
-
"async_rl": true,
|
| 23 |
-
"batch_size": 1024,
|
| 24 |
-
"batched_sampling": true,
|
| 25 |
-
"benchmark": false,
|
| 26 |
-
"cli_args": {
|
| 27 |
-
"action_repeat": 4,
|
| 28 |
-
"add_latency_info": false,
|
| 29 |
-
"algo": "APPO",
|
| 30 |
-
"async_rl": true,
|
| 31 |
-
"batch_size": 1024,
|
| 32 |
-
"batched_sampling": true,
|
| 33 |
-
"clip_reward": true,
|
| 34 |
-
"device": "gpu",
|
| 35 |
-
"env": "latency_demon_attack",
|
| 36 |
-
"env_id": "ALE/DemonAttack-v5",
|
| 37 |
-
"episode_metrics_path": "results/checkpoints/demon_attack_sf_same_latency_l8_seed0/episode_metrics.jsonl",
|
| 38 |
-
"episodes_per_stage": 10,
|
| 39 |
-
"eval_deterministic": true,
|
| 40 |
-
"eval_episodes": 20,
|
| 41 |
-
"eval_max_steps": 3600,
|
| 42 |
-
"eval_queue_latency_values": "0,1,2,3,4,5,10,15",
|
| 43 |
-
"eval_raw_reward": true,
|
| 44 |
-
"experiment": "demon_attack_sf_same_latency_l8_seed0",
|
| 45 |
-
"experiment_summaries_interval": 1,
|
| 46 |
-
"frame_stack": 4,
|
| 47 |
-
"full_action_space": false,
|
| 48 |
-
"gae_lambda": 0.95,
|
| 49 |
-
"gamma": 0.99,
|
| 50 |
-
"keep_checkpoints": 5,
|
| 51 |
-
"latency_mode": "episode_fixed",
|
| 52 |
-
"latency_sampling_mode": "uniform_all",
|
| 53 |
-
"learning_rate": 0.000303,
|
| 54 |
-
"max_episode_steps": 25000,
|
| 55 |
-
"max_grad_norm": 0.0,
|
| 56 |
-
"mode": "train",
|
| 57 |
-
"noop_max": 30,
|
| 58 |
-
"normalize_input": true,
|
| 59 |
-
"normalize_returns": true,
|
| 60 |
-
"num_batches_per_epoch": 2,
|
| 61 |
-
"num_envs_per_worker": 4,
|
| 62 |
-
"num_epochs": 2,
|
| 63 |
-
"num_policies": 1,
|
| 64 |
-
"num_workers": 16,
|
| 65 |
-
"ppo_clip_ratio": 0.1,
|
| 66 |
-
"ppo_clip_value": 0.2,
|
| 67 |
-
"queue_latency_values": "8",
|
| 68 |
-
"randomize_within_stage": false,
|
| 69 |
-
"recurrence": 32,
|
| 70 |
-
"repeat_action_probability": 0.0,
|
| 71 |
-
"restart_behavior": "overwrite",
|
| 72 |
-
"rollout": 128,
|
| 73 |
-
"save_every_sec": 600,
|
| 74 |
-
"screen_size": 84,
|
| 75 |
-
"seed": 0,
|
| 76 |
-
"stats_avg": 100,
|
| 77 |
-
"terminal_on_life_loss": true,
|
| 78 |
-
"train_dir": "results/checkpoints",
|
| 79 |
-
"train_for_env_steps": 50000000,
|
| 80 |
-
"use_rnn": false,
|
| 81 |
-
"value_loss_coeff": 0.5,
|
| 82 |
-
"wandb_group": "demon_attack",
|
| 83 |
-
"wandb_job_type": "sample_factory",
|
| 84 |
-
"wandb_project": "latency-sensitive-bench",
|
| 85 |
-
"with_wandb": true,
|
| 86 |
-
"within_episode_change_interval": 1,
|
| 87 |
-
"within_episode_change_prob": 1.0,
|
| 88 |
-
"worker_num_splits": 2
|
| 89 |
-
},
|
| 90 |
-
"clip_reward": true,
|
| 91 |
-
"command_line": "--mode train --algo APPO --env latency_demon_attack --experiment demon_attack_sf_same_latency_l8_seed0 --train_dir results/checkpoints --restart_behavior overwrite --device gpu --seed 0 --episode_metrics_path results/checkpoints/demon_attack_sf_same_latency_l8_seed0/episode_metrics.jsonl --train_for_env_steps 50000000 --num_workers 16 --num_envs_per_worker 4 --num_policies 1 --batch_size 1024 --rollout 128 --recurrence 32 --num_epochs 2 --num_batches_per_epoch 2 --worker_num_splits 2 --learning_rate 0.000303 --gamma 0.99 --gae_lambda 0.95 --ppo_clip_ratio 0.1 --ppo_clip_value 0.2 --value_loss_coeff 0.5 --max_grad_norm 0.0 --save_every_sec 600 --keep_checkpoints 5 --stats_avg 100 --experiment_summaries_interval 1 --async_rl True --batched_sampling True --use_rnn False --normalize_returns True --normalize_input True --queue-latency-values 8 --latency-mode episode_fixed --latency-sampling-mode uniform_all --episodes-per-stage 10 --randomize-within-stage False --within-episode-change-interval 1 --within-episode-change-prob 1.0 --add-latency-info False --eval-episodes 20 --eval-queue-latency-values 0,1,2,3,4,5,10,15 --eval-max-steps 3600 --eval-deterministic True --eval-raw-reward --with_wandb True --wandb_project latency-sensitive-bench --wandb_group demon_attack --wandb_job_type sample_factory --env_id ALE/DemonAttack-v5 --screen_size 84 --noop_max 30 --action_repeat 4 --frame_stack 4 --terminal_on_life_loss True --clip_reward True --repeat_action_probability 0.0 --full_action_space False --max_episode_steps 25000",
|
| 92 |
-
"continuous_tanh_scale": 0.0,
|
| 93 |
-
"decoder_mlp_layers": [],
|
| 94 |
-
"decorrelate_envs_on_one_worker": true,
|
| 95 |
-
"decorrelate_experience_max_seconds": 0,
|
| 96 |
-
"default_niceness": 0,
|
| 97 |
-
"device": "gpu",
|
| 98 |
-
"encoder_conv_architecture": "convnet_atari",
|
| 99 |
-
"encoder_conv_mlp_layers": [
|
| 100 |
-
512
|
| 101 |
-
],
|
| 102 |
-
"encoder_mlp_layers": [
|
| 103 |
-
512,
|
| 104 |
-
512
|
| 105 |
-
],
|
| 106 |
-
"env": "latency_demon_attack",
|
| 107 |
-
"env_frameskip": 1,
|
| 108 |
-
"env_framestack": 1,
|
| 109 |
-
"env_gpu_actions": false,
|
| 110 |
-
"env_gpu_observations": true,
|
| 111 |
-
"env_id": "ALE/DemonAttack-v5",
|
| 112 |
-
"episode_metrics_path": "results/checkpoints/demon_attack_sf_same_latency_l8_seed0/episode_metrics.jsonl",
|
| 113 |
-
"episodes_per_stage": 10,
|
| 114 |
-
"eval_deterministic": true,
|
| 115 |
-
"eval_episodes": 20,
|
| 116 |
-
"eval_max_steps": 3600,
|
| 117 |
-
"eval_queue_latency_values": "0,1,2,3,4,5,10,15",
|
| 118 |
-
"eval_raw_reward": true,
|
| 119 |
-
"experiment": "demon_attack_sf_same_latency_l8_seed0",
|
| 120 |
-
"experiment_summaries_interval": 1,
|
| 121 |
-
"exploration_loss": "entropy",
|
| 122 |
-
"exploration_loss_coeff": 0.003,
|
| 123 |
-
"flush_summaries_interval": 30,
|
| 124 |
-
"force_envs_single_thread": false,
|
| 125 |
-
"frame_stack": 4,
|
| 126 |
-
"full_action_space": false,
|
| 127 |
-
"gae_lambda": 0.95,
|
| 128 |
-
"gamma": 0.99,
|
| 129 |
-
"git_hash": "41bc95edbf670a1190d128da22e9cd3777b87989",
|
| 130 |
-
"git_repo_name": "git@github.com:ZihanWang314/latency-sensitive-bench.git",
|
| 131 |
-
"heartbeat_interval": 20,
|
| 132 |
-
"heartbeat_reporting_interval": 180,
|
| 133 |
-
"help": false,
|
| 134 |
-
"initial_stddev": 1.0,
|
| 135 |
-
"keep_checkpoints": 5,
|
| 136 |
-
"kl_loss_coeff": 0.0,
|
| 137 |
-
"latency_mode": "episode_fixed",
|
| 138 |
-
"latency_sampling_mode": "uniform_all",
|
| 139 |
-
"learning_rate": 0.000303,
|
| 140 |
-
"load_checkpoint_kind": "latest",
|
| 141 |
-
"log_to_file": true,
|
| 142 |
-
"lr_adaptive_max": 0.01,
|
| 143 |
-
"lr_adaptive_min": 1e-06,
|
| 144 |
-
"lr_schedule": "linear_decay",
|
| 145 |
-
"lr_schedule_kl_threshold": 0.008,
|
| 146 |
-
"max_episode_steps": 25000,
|
| 147 |
-
"max_grad_norm": 0.0,
|
| 148 |
-
"max_policy_lag": 1000,
|
| 149 |
-
"mode": "train",
|
| 150 |
-
"nonlinearity": "elu",
|
| 151 |
-
"noop_max": 30,
|
| 152 |
-
"normalize_input": true,
|
| 153 |
-
"normalize_input_keys": null,
|
| 154 |
-
"normalize_returns": true,
|
| 155 |
-
"num_batches_per_epoch": 2,
|
| 156 |
-
"num_batches_to_accumulate": 2,
|
| 157 |
-
"num_envs_per_worker": 4,
|
| 158 |
-
"num_epochs": 2,
|
| 159 |
-
"num_policies": 1,
|
| 160 |
-
"num_workers": 16,
|
| 161 |
-
"obs_scale": 255.0,
|
| 162 |
-
"obs_subtract_mean": 0.0,
|
| 163 |
-
"optimizer": "adam",
|
| 164 |
-
"output_dir": "outputs/train/demon_attack/latency_8/seed_0",
|
| 165 |
-
"pbt_mix_policies_in_one_env": true,
|
| 166 |
-
"pbt_mutation_rate": 0.15,
|
| 167 |
-
"pbt_optimize_gamma": false,
|
| 168 |
-
"pbt_period_env_steps": 5000000,
|
| 169 |
-
"pbt_perturb_max": 1.5,
|
| 170 |
-
"pbt_perturb_min": 1.1,
|
| 171 |
-
"pbt_replace_fraction": 0.3,
|
| 172 |
-
"pbt_replace_reward_gap": 0.1,
|
| 173 |
-
"pbt_replace_reward_gap_absolute": 1e-06,
|
| 174 |
-
"pbt_start_mutation": 20000000,
|
| 175 |
-
"pbt_target_objective": "true_objective",
|
| 176 |
-
"pixel_format": "CHW",
|
| 177 |
-
"policy_init_gain": 1.0,
|
| 178 |
-
"policy_initialization": "orthogonal",
|
| 179 |
-
"policy_workers_per_policy": 1,
|
| 180 |
-
"ppo_clip_ratio": 0.1,
|
| 181 |
-
"ppo_clip_value": 0.2,
|
| 182 |
-
"queue_latency_values": "8",
|
| 183 |
-
"randomize_within_stage": false,
|
| 184 |
-
"recurrence": 32,
|
| 185 |
-
"repeat_action_probability": 0.0,
|
| 186 |
-
"restart_behavior": "overwrite",
|
| 187 |
-
"reward_clip": 1000.0,
|
| 188 |
-
"reward_scale": 1.0,
|
| 189 |
-
"rnn_num_layers": 1,
|
| 190 |
-
"rnn_size": 512,
|
| 191 |
-
"rnn_type": "gru",
|
| 192 |
-
"rollout": 128,
|
| 193 |
-
"save_best_after": 100000,
|
| 194 |
-
"save_best_every_sec": 5,
|
| 195 |
-
"save_best_metric": "reward",
|
| 196 |
-
"save_every_sec": 600,
|
| 197 |
-
"save_milestones_sec": -1,
|
| 198 |
-
"screen_size": 84,
|
| 199 |
-
"seed": 0,
|
| 200 |
-
"serial_mode": false,
|
| 201 |
-
"set_workers_cpu_affinity": true,
|
| 202 |
-
"shuffle_minibatches": false,
|
| 203 |
-
"stats_avg": 100,
|
| 204 |
-
"summaries_use_frameskip": true,
|
| 205 |
-
"terminal_on_life_loss": true,
|
| 206 |
-
"train_dir": "results/checkpoints",
|
| 207 |
-
"train_for_env_steps": 50000000,
|
| 208 |
-
"train_for_seconds": 10000000000,
|
| 209 |
-
"use_env_info_cache": false,
|
| 210 |
-
"use_record_episode_statistics": false,
|
| 211 |
-
"use_rnn": false,
|
| 212 |
-
"value_bootstrap": false,
|
| 213 |
-
"value_loss_coeff": 0.5,
|
| 214 |
-
"vtrace_c": 1.0,
|
| 215 |
-
"vtrace_rho": 1.0,
|
| 216 |
-
"wandb_group": "demon_attack",
|
| 217 |
-
"wandb_job_type": "sample_factory",
|
| 218 |
-
"wandb_project": "latency-sensitive-bench",
|
| 219 |
-
"wandb_tags": [],
|
| 220 |
-
"wandb_unique_id": "demon-attack-l8-s0",
|
| 221 |
-
"wandb_user": null,
|
| 222 |
-
"with_pbt": false,
|
| 223 |
-
"with_vtrace": false,
|
| 224 |
-
"with_wandb": true,
|
| 225 |
-
"within_episode_change_interval": 1,
|
| 226 |
-
"within_episode_change_prob": 1.0,
|
| 227 |
-
"worker_num_splits": 2
|
| 228 |
-
},
|
| 229 |
-
"checkpoint_experiment_dir": "/inspire/hdd/global_user/liumingyu-253208120284/lzj/latency-sensitive-bench/checkpoints/demon_attack_sf_same_latency_l8_seed0",
|
| 230 |
-
"checkpoint_path": null,
|
| 231 |
-
"checkpoint_train_step": 97664,
|
| 232 |
-
"config_source": "/inspire/hdd/global_user/liumingyu-253208120284/lzj/latency-sensitive-bench/checkpoints/demon_attack_sf_same_latency_l8_seed0/config.json",
|
| 233 |
-
"env_fps": 60.0,
|
| 234 |
-
"env_name": "demon_attack",
|
| 235 |
-
"episode_return_field": "episode_raw_return",
|
| 236 |
-
"filter_preset": "clean_v1",
|
| 237 |
-
"filter_report": {
|
| 238 |
-
"accepted_episodes": 200,
|
| 239 |
-
"accepted_train_episodes": 180,
|
| 240 |
-
"accepted_val_episodes": 20,
|
| 241 |
-
"attempted_episodes": 2304,
|
| 242 |
-
"preset": "clean_v1",
|
| 243 |
-
"raw_score_bins": {
|
| 244 |
-
"6460-6889": 0,
|
| 245 |
-
"6890-7309": 0,
|
| 246 |
-
"7310-7999": 0,
|
| 247 |
-
"<6460": 200,
|
| 248 |
-
">=8000": 0
|
| 249 |
-
},
|
| 250 |
-
"rejected_episodes": 2104,
|
| 251 |
-
"train_rows_after_row_filter": 282142,
|
| 252 |
-
"train_rows_before_row_filter": 646284,
|
| 253 |
-
"val_rows_after_row_filter": 40599,
|
| 254 |
-
"val_rows_before_row_filter": 71901
|
| 255 |
-
},
|
| 256 |
-
"frame_stack": 4,
|
| 257 |
-
"image_resolution": [
|
| 258 |
-
210,
|
| 259 |
-
160,
|
| 260 |
-
3
|
| 261 |
-
],
|
| 262 |
-
"image_source": "env_raw_rgb",
|
| 263 |
-
"latency_override": {
|
| 264 |
-
"add_latency_info": false,
|
| 265 |
-
"fixed_latency_ms": 533.3333333333334,
|
| 266 |
-
"type": "fixed"
|
| 267 |
-
},
|
| 268 |
-
"latency_source": "legacy_checkpoint_name",
|
| 269 |
-
"max_episodes": 200,
|
| 270 |
-
"max_steps_per_episode": 3600,
|
| 271 |
-
"num_envs": 128,
|
| 272 |
-
"obs_fps": 15.0,
|
| 273 |
-
"obs_stride_raw_frames": 4,
|
| 274 |
-
"prompt_source": "latency_bench.data.prompts",
|
| 275 |
-
"prompt_timing_suffix": " Current action latency is {latency_raw_frames} raw frames ({latency_ms:.2f} ms). The environment runs at {env_fps:g} FPS and observations are emitted at {obs_fps:g} FPS. Choose the best next action.",
|
| 276 |
-
"replay_num_envs": 16,
|
| 277 |
-
"reward_field": "raw_reward",
|
| 278 |
-
"reward_semantics": "raw_reward",
|
| 279 |
-
"rows_unit": "decision_step",
|
| 280 |
-
"schema_version": 2,
|
| 281 |
-
"split_seed": 0,
|
| 282 |
-
"total_episodes": 200,
|
| 283 |
-
"total_rows": 322741,
|
| 284 |
-
"train_episodes": 180,
|
| 285 |
-
"train_rows": 282142,
|
| 286 |
-
"val_episodes": 20,
|
| 287 |
-
"val_fraction": 0.1,
|
| 288 |
-
"val_rows": 40599
|
| 289 |
-
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_8_200ep/train.parquet
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:04efa8cc7c30b52b8820a057825194edc793fbb20e0e2c214f09d11eb175fb2c
|
| 3 |
-
size 159016885
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/demon-attack/training_versions/demon_attack_200ep/0481972d94f6409a392c5b035f9a7f5c2681a96d/demon_attack_fix_latency_8_200ep/val.parquet
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:59080949006a7db9d6fd11affe01ca21ccd857b79ca9fd48d073e10015a1c334
|
| 3 |
-
size 22559655
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/README.md
DELETED
|
@@ -1 +0,0 @@
|
|
| 1 |
-
# parkour qwengr00t DAgger round 0
|
|
|
|
|
|
latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/lerobot/data/chunk-000/episode_000000.parquet
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:2f7ceaf9180ff4fcf932ea377beadcc6a75ef3779837a4b1a69bdb5c219fe006
|
| 3 |
-
size 20455
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/lerobot/data/chunk-000/episode_000001.parquet
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:e0d78fcb2230b6619f68059089d03ae4dfa8de5f23c9d3da42740a3237ed86f3
|
| 3 |
-
size 20489
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/lerobot/data/chunk-000/episode_000002.parquet
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:95948b6fd4d2a1f4447d94eae9d57aeb3f2097ff3de43cfa704c1a5f6079bd78
|
| 3 |
-
size 20465
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/lerobot/data/chunk-000/episode_000003.parquet
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:23ec33ce639445c8b1b86228671844bc6053da3b6c15ceee44091bcceaa1a3c3
|
| 3 |
-
size 24614
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/lerobot/data/chunk-000/episode_000004.parquet
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:05cc2bcdb985044c6a737a64bd2d88ce8b15e8ab4938ad589acd5b073860da39
|
| 3 |
-
size 24989
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/lerobot/data/chunk-000/episode_000005.parquet
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:cb6a39dc9045a575af7d74ca2691c9713681f8e064f3930457801ede9237aa99
|
| 3 |
-
size 24954
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/lerobot/data/chunk-000/episode_000006.parquet
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:1c26f5286332958ac13b25fb87f2b407bfbd366fedf39aa6ca7a6e9ba4049d7f
|
| 3 |
-
size 27782
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/lerobot/data/chunk-000/episode_000007.parquet
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:873105557ef245a70713b9e7514d6b4fc3dc7f0bc4e9488ec39f601a97893bba
|
| 3 |
-
size 28917
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/lerobot/data/chunk-000/episode_000008.parquet
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:0fdf0c9f7c43edec8a0be21735ade9456d29dbb050837bfe8bcf33db7beed965
|
| 3 |
-
size 29146
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/lerobot/data/chunk-000/episode_000009.parquet
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:ebffdf665ec20a874c7c4c3755613bfbafac6ecc705331bf08c23d2d25de108b
|
| 3 |
-
size 29476
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/lerobot/data/chunk-000/episode_000010.parquet
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:37b427dbd2f89cda5fc55f0ed4667bb8fa21812a713fc41f3b798432af30ef18
|
| 3 |
-
size 29479
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/lerobot/data/chunk-000/episode_000011.parquet
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:e8ea464a9ef107945cc47d0040bf95dbd4f348765f3766a2e81512e2af94d00b
|
| 3 |
-
size 29487
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/lerobot/data/chunk-000/episode_000012.parquet
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:f2bf699515d1c6675b40c7f5d102216f559e475570669b3e7b39330f63617b94
|
| 3 |
-
size 29523
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/lerobot/data/chunk-000/episode_000013.parquet
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:edd3aebb01b30aef347f7a19a0b50b53440ab25b76bb1bedcb51fdd87fceacf6
|
| 3 |
-
size 32414
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/lerobot/data/chunk-000/episode_000014.parquet
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:a5a9277c212fbaa4162328aa91537fb6c446f5407e1043e846c3abc764108e35
|
| 3 |
-
size 33055
|
|
|
|
|
|
|
|
|
|
|
|
latency-aware/extreme-parkour-go2/dagger/starvla-qwengr00t-h1/parkour-qwengr00t-latency-2xrtx6000pro-20260919/round-00/lerobot/data/chunk-000/episode_000015.parquet
DELETED
|
@@ -1,3 +0,0 @@
|
|
| 1 |
-
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:bb4480f70719e4d25442beb22ab14b5025c5eba1aae0594bf96245d529ba0af9
|
| 3 |
-
size 34077
|
|
|
|
|
|
|
|
|
|
|
|