Update LAGEN resource links after transfer to MLL-Lab
Browse files- README.md +15 -13
- experiments/sim2real-calibration/README.md +7 -7
README.md
CHANGED
|
@@ -7,26 +7,28 @@ tags:
|
|
| 7 |
- multiple-checkpoints
|
| 8 |
---
|
| 9 |
|
| 10 |
-
# LAGEN
|
|
|
|
|
|
|
| 11 |
|
| 12 |
Inference checkpoints, original configurations and model provenance.
|
| 13 |
|
| 14 |
| Experiment | Entry |
|
| 15 |
|---|---|
|
| 16 |
-
| Sim2Real calibration | [30-case held-out calibration](https://huggingface.co/datasets/
|
| 17 |
-
| Visual history | [Visual history](https://huggingface.co/datasets/
|
| 18 |
-
| Latency in prompt | [Latency in prompt](https://huggingface.co/datasets/
|
| 19 |
-
| Latency transfer | [Latency transfer](https://huggingface.co/datasets/
|
| 20 |
-
| Task transfer | [Task transfer](https://huggingface.co/datasets/
|
| 21 |
-
| VLA fine-tuning scope | [VLA fine-tuning scope](https://huggingface.co/datasets/
|
| 22 |
-
| Mean vs. profile training | [Mean vs. profile training](https://huggingface.co/datasets/
|
| 23 |
-
| Observation stride | [Observation stride](https://huggingface.co/datasets/
|
| 24 |
-
| Context window | [Context window](https://huggingface.co/datasets/
|
| 25 |
|
| 26 |
| Resource | Repository |
|
| 27 |
|---|---|
|
| 28 |
-
| benchmark-datasets | [
|
| 29 |
-
| profiles | [
|
| 30 |
|
| 31 |
[Benchmark release inventory](reports/benchmark-release/README.md)
|
| 32 |
|
|
@@ -40,4 +42,4 @@ HAIC and Extreme Parkour releases are retired. Model bundles retain the publishe
|
|
| 40 |
|
| 41 |
## Source consolidation
|
| 42 |
|
| 43 |
-
[Verified source index](https://huggingface.co/datasets/
|
|
|
|
| 7 |
- multiple-checkpoints
|
| 8 |
---
|
| 9 |
|
| 10 |
+
# LAGEN models
|
| 11 |
+
|
| 12 |
+
Project resources: [LAGEN collection](https://huggingface.co/collections/MLL-Lab/lagen-6ac750f21139008b8114fff5).
|
| 13 |
|
| 14 |
Inference checkpoints, original configurations and model provenance.
|
| 15 |
|
| 16 |
| Experiment | Entry |
|
| 17 |
|---|---|
|
| 18 |
+
| Sim2Real calibration | [30-case held-out calibration](https://huggingface.co/datasets/MLL-Lab/LAGEN-datasets/tree/50d8f0aef64749a849760fc580ce3efb3c5c5e76/experiments/sim2real-calibration) |
|
| 19 |
+
| Visual history | [Visual history](https://huggingface.co/datasets/MLL-Lab/LAGEN-datasets/tree/main/experiments/visual-history) |
|
| 20 |
+
| Latency in prompt | [Latency in prompt](https://huggingface.co/datasets/MLL-Lab/LAGEN-datasets/tree/main/experiments/latency-in-prompt) |
|
| 21 |
+
| Latency transfer | [Latency transfer](https://huggingface.co/datasets/MLL-Lab/LAGEN-datasets/tree/main/experiments/latency-transfer) |
|
| 22 |
+
| Task transfer | [Task transfer](https://huggingface.co/datasets/MLL-Lab/LAGEN-datasets/tree/main/experiments/task-transfer) |
|
| 23 |
+
| VLA fine-tuning scope | [VLA fine-tuning scope](https://huggingface.co/datasets/MLL-Lab/LAGEN-datasets/tree/main/experiments/vla-finetuning-scope) |
|
| 24 |
+
| Mean vs. profile training | [Mean vs. profile training](https://huggingface.co/datasets/MLL-Lab/LAGEN-datasets/tree/main/experiments/mean-vs-profile) |
|
| 25 |
+
| Observation stride | [Observation stride](https://huggingface.co/datasets/MLL-Lab/LAGEN-datasets/tree/main/experiments/observation-stride) |
|
| 26 |
+
| Context window | [Context window](https://huggingface.co/datasets/MLL-Lab/LAGEN-datasets/tree/main/experiments/context-window) |
|
| 27 |
|
| 28 |
| Resource | Repository |
|
| 29 |
|---|---|
|
| 30 |
+
| benchmark-datasets | [MLL-Lab/LAGEN-datasets](https://huggingface.co/datasets/MLL-Lab/LAGEN-datasets) |
|
| 31 |
+
| profiles | [MLL-Lab/LAGEN-profiles](https://huggingface.co/datasets/MLL-Lab/LAGEN-profiles) |
|
| 32 |
|
| 33 |
[Benchmark release inventory](reports/benchmark-release/README.md)
|
| 34 |
|
|
|
|
| 42 |
|
| 43 |
## Source consolidation
|
| 44 |
|
| 45 |
+
[Verified source index](https://huggingface.co/datasets/MLL-Lab/LAGEN-datasets/tree/a6e93d4a9a1277c5cd5a7145e0295e3e190374e1/reports/source-repository-retirement-20261008) records the former Standard-Pipeline, HumanoidBench, MIKASA and Figure 2 sources. The 65 selected Standard-Pipeline inference bundles reuse existing canonical model paths. All six confirmed Figure 2 game checkpoints are published with their original SHA256 and size. Available original configs and statistics remain with their canonical bundles. Unverified runtime capability and missing historical bindings remain explicit.
|
experiments/sim2real-calibration/README.md
CHANGED
|
@@ -4,14 +4,14 @@ Six existing policies are shared across five inference GPU conditions.
|
|
| 4 |
|
| 5 |
| Task | Model | Existing bundle |
|
| 6 |
|---|---|---|
|
| 7 |
-
| flappy | qwenoft | [latency-aware/flappy/vla/starvla-qwenoft-h1/flappy-qwenoft-hist8-scratch-s42-v2](/
|
| 8 |
-
| flappy | qwengr00t | [latency-aware/flappy/vla/starvla-qwengr00t-h1/flappy-qwengr00t-hist8-scratch-s42-v2](/
|
| 9 |
-
| inverted_pendulum | qwenoft | [latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/inverted_pendulum_rgb_state_ip10_20260911T010114Z_openvla_native_sft_2k](/
|
| 10 |
-
| inverted_pendulum | qwengr00t | [latency-aware/inverted-pendulum/vla/starvla-qwengr00t-h1/ip_profile_gr00t_h1_5fps_1h200_20260917](/
|
| 11 |
-
| mikasa_intercept_grab_fast | qwenoft | [latency-aware/mikasa-intercept-grab-fast/vla/starvla-qwenoft-h1/mikasa_qwenoft_latency_3090_2xh200_20260922_evalfix_20260923](/
|
| 12 |
-
| mikasa_intercept_grab_fast | qwengr00t | [latency-aware/mikasa-intercept-grab-fast/vla/starvla-qwengr00t-h1/mikasa_qwengr00t_latency_3090_2xh200_20260922_evalfix_20260923](/
|
| 13 |
|
| 14 |
-
[Experiment settings, checkpoints, 30 cases and results](https://huggingface.co/datasets/
|
| 15 |
|
| 16 |
Checkpoint, configuration, dataset statistics and policy metadata identities
|
| 17 |
match the experiment snapshot `180da0effdaaaeacef218cf5fbaa4db41af67495`.
|
|
|
|
| 4 |
|
| 5 |
| Task | Model | Existing bundle |
|
| 6 |
|---|---|---|
|
| 7 |
+
| flappy | qwenoft | [latency-aware/flappy/vla/starvla-qwenoft-h1/flappy-qwenoft-hist8-scratch-s42-v2](/MLL-Lab/LAGEN-models/tree/180da0effdaaaeacef218cf5fbaa4db41af67495/latency-aware/flappy/vla/starvla-qwenoft-h1/flappy-qwenoft-hist8-scratch-s42-v2) |
|
| 8 |
+
| flappy | qwengr00t | [latency-aware/flappy/vla/starvla-qwengr00t-h1/flappy-qwengr00t-hist8-scratch-s42-v2](/MLL-Lab/LAGEN-models/tree/180da0effdaaaeacef218cf5fbaa4db41af67495/latency-aware/flappy/vla/starvla-qwengr00t-h1/flappy-qwengr00t-hist8-scratch-s42-v2) |
|
| 9 |
+
| inverted_pendulum | qwenoft | [latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/inverted_pendulum_rgb_state_ip10_20260911T010114Z_openvla_native_sft_2k](/MLL-Lab/LAGEN-models/tree/180da0effdaaaeacef218cf5fbaa4db41af67495/latency-aware/inverted-pendulum/vla/starvla-qwenoft-h1/inverted_pendulum_rgb_state_ip10_20260911T010114Z_openvla_native_sft_2k) |
|
| 10 |
+
| inverted_pendulum | qwengr00t | [latency-aware/inverted-pendulum/vla/starvla-qwengr00t-h1/ip_profile_gr00t_h1_5fps_1h200_20260917](/MLL-Lab/LAGEN-models/tree/180da0effdaaaeacef218cf5fbaa4db41af67495/latency-aware/inverted-pendulum/vla/starvla-qwengr00t-h1/ip_profile_gr00t_h1_5fps_1h200_20260917) |
|
| 11 |
+
| mikasa_intercept_grab_fast | qwenoft | [latency-aware/mikasa-intercept-grab-fast/vla/starvla-qwenoft-h1/mikasa_qwenoft_latency_3090_2xh200_20260922_evalfix_20260923](/MLL-Lab/LAGEN-models/tree/180da0effdaaaeacef218cf5fbaa4db41af67495/latency-aware/mikasa-intercept-grab-fast/vla/starvla-qwenoft-h1/mikasa_qwenoft_latency_3090_2xh200_20260922_evalfix_20260923) |
|
| 12 |
+
| mikasa_intercept_grab_fast | qwengr00t | [latency-aware/mikasa-intercept-grab-fast/vla/starvla-qwengr00t-h1/mikasa_qwengr00t_latency_3090_2xh200_20260922_evalfix_20260923](/MLL-Lab/LAGEN-models/tree/180da0effdaaaeacef218cf5fbaa4db41af67495/latency-aware/mikasa-intercept-grab-fast/vla/starvla-qwengr00t-h1/mikasa_qwengr00t_latency_3090_2xh200_20260922_evalfix_20260923) |
|
| 13 |
|
| 14 |
+
[Experiment settings, checkpoints, 30 cases and results](https://huggingface.co/datasets/MLL-Lab/LAGEN-datasets/tree/50d8f0aef64749a849760fc580ce3efb3c5c5e76/experiments/sim2real-calibration)
|
| 15 |
|
| 16 |
Checkpoint, configuration, dataset statistics and policy metadata identities
|
| 17 |
match the experiment snapshot `180da0effdaaaeacef218cf5fbaa4db41af67495`.
|