Index: RoboTwin-10 seven-model release

README.md

@@ -9,7 +9,7 @@ tags:
 
 # LoopWAM final checkpoints
 
-Final trained policy weights from the LoopWAM experiments. This repository contains **25 checkpoints**, separated by training data scope. Both original and repeated 4/4 runs are retained. All production checkpoints are from completed ten-epoch runs with global batch 128 and training seed 42.
+Final trained policy weights from the LoopWAM experiments. This repository contains **32 checkpoints**, separated by training data scope. Both original and repeated 4/4 runs are retained. All production checkpoints are from completed ten-epoch runs with global batch 128 and training seed 42.
 
 ## October 9 additions: all-loop KV
 
@@ -93,10 +93,33 @@ v0 (video 4 / action 4) retrained on all four suites with the same data, normali
 
 Aligned (late) 4/2 v0 trained on all four suites with the same data, normalization, initialization recipe, schedule, precision and 2-GPU microbatch-8 layout as `libero-all-suites/v0-video4-action1` (21,700 updates, global batch 128, seed 42) on two H100s. Uses the LoopWAM_NT branch (revision `56e3328`). Not evaluated.
 
+## November 2026 addition: RoboTwin-10 (LiLa-WAM 10-task subset)
+
+Seven LoopWAM architectures trained for five epochs on the 10-task RoboTwin-10 subset with the same data order, noise, initialization recipe and schedule; see [robotwin-10/README.md](robotwin-10/README.md).
+
+| Directory | Model | Video/action loops | Parameters | Training time | Final loss (last 100) | Evaluation |
+|---|---|---|---:|---:|---:|---|
+| [robotwin-10/v0-video4-action4](robotwin-10/v0-video4-action4) | v0 | 4/4 | 584,567,886 | 17.8 h | 0.0453 | Training only; evaluation not run |
+| [robotwin-10/dense-s12](robotwin-10/dense-s12) | dense_s12 | 1/1 | 584,567,886 | 8.0 h | 0.0485 | Training only; evaluation not run |
+| [robotwin-10/dense-s30](robotwin-10/dense-s30) | dense_s30 | 1/1 | 1,416,145,998 | 19.4 h | 0.0350 | Training only; evaluation not run |
+| [robotwin-10/v0-video4-action2](robotwin-10/v0-video4-action2) | v0 aligned | 4/2 | 584,567,886 | 16.7 h | 0.0452 | Training only; evaluation not run |
+| [robotwin-10/v0-video4-action1](robotwin-10/v0-video4-action1) | v0 aligned | 4/1 | 584,567,886 | 16.3 h | 0.0453 | Training only; evaluation not run |
+| [robotwin-10/v0-video2-action2](robotwin-10/v0-video2-action2) | v0 aligned | 2/2 | 584,567,886 | 11.1 h | 0.0469 | Training only; evaluation not run |
+| [robotwin-10/v0-video1-action4](robotwin-10/v0-video1-action4) | v0 aligned | 1/4 | 584,567,886 | 11.9 h | 0.0485 | Training only; evaluation not run |
+
+Training only; evaluation not run.
+
 ## Checkpoint index
 
 | Directory | Model | Video/action loops | Parameters | Final evaluation |
 |---|---|---|---:|---|
+| [robotwin-10/v0-video4-action4](robotwin-10/v0-video4-action4) | v0 (RoboTwin-10) | 4/4 | 584,567,886 | Training only; evaluation not run |
+| [robotwin-10/dense-s12](robotwin-10/dense-s12) | dense_s12 (RoboTwin-10) | 1/1 | 584,567,886 | Training only; evaluation not run |
+| [robotwin-10/dense-s30](robotwin-10/dense-s30) | dense_s30 (RoboTwin-10) | 1/1 | 1,416,145,998 | Training only; evaluation not run |
+| [robotwin-10/v0-video4-action2](robotwin-10/v0-video4-action2) | v0 aligned (RoboTwin-10) | 4/2 | 584,567,886 | Training only; evaluation not run |
+| [robotwin-10/v0-video4-action1](robotwin-10/v0-video4-action1) | v0 aligned (RoboTwin-10) | 4/1 | 584,567,886 | Training only; evaluation not run |
+| [robotwin-10/v0-video2-action2](robotwin-10/v0-video2-action2) | v0 aligned (RoboTwin-10) | 2/2 | 584,567,886 | Training only; evaluation not run |
+| [robotwin-10/v0-video1-action4](robotwin-10/v0-video1-action4) | v0 aligned (RoboTwin-10) | 1/4 | 584,567,886 | Training only; evaluation not run |
 | [libero-all-suites/init_variants/core_mean](libero-all-suites/init_variants/core_mean) | v0 init `core_mean` | 4/4 | 584,536,135 | Training only; evaluation not run |
 | [libero-all-suites/v0-video4-action2](libero-all-suites/v0-video4-action2) | v0 aligned | 4/2 | 584,536,135 | Training only; evaluation not run |
 | [libero-all-suites/init_variants/core16_21](libero-all-suites/init_variants/core16_21) | v0 init `core16_21` | 4/4 | 584,536,135 | Training only; evaluation not run |

checkpoint_index.json

@@ -788,7 +788,273 @@
       "microbatch": 8,
       "gradient_accumulation": 8,
       "created_utc": "2026-10-28T19:14:54.782633+00:00"
+    },
+    {
+      "id": "robotwin10_v0_video4_action4",
+      "repo_path": "robotwin-10/v0-video4-action4",
+      "training_path": "/groups/yshang/an221229/checkpoints/LoopWAM_NT/robotwin10/v0-video4-action4/train",
+      "export_sha256": "bb1563f16bcd102b3f4ad4a1a14377ee49de7e461d862b7bbc5dba497f036127",
+      "export_bytes": 2338751845,
+      "step": 45445,
+      "format": "loopwam-s-robot-v1",
+      "optimizer_included": false,
+      "version": "v0",
+      "video_loops": 4,
+      "action_loops": 4,
+      "loop_alignment": "late",
+      "parameters": 584567886,
+      "dataset_scope": "robotwin10",
+      "tasks": [
+        "adjust_bottle",
+        "grab_roller",
+        "hanging_mug",
+        "move_stapler_pad",
+        "open_microwave",
+        "press_stapler",
+        "scan_object",
+        "stack_blocks_two",
+        "stamp_seal",
+        "turn_switch"
+      ],
+      "epochs": 5,
+      "global_batch": 128,
+      "training_seed": 42,
+      "initialization_mode": "fresh_wan_donors_no_embodied_pretraining",
+      "evaluation_status": "not_evaluated",
+      "training_seconds": 64081.80207467079,
+      "final_loss_last100": 0.045258548930287364,
+      "tensor_equality_verified": true,
+      "source_revision": "bea472670ea52bb375ae75ba5380cd88087d2c30",
+      "hf_revision": "df3d575d3f3fde3cf1b8d5f609fab93dc6a95649"
+    },
+    {
+      "id": "robotwin10_dense_s12",
+      "repo_path": "robotwin-10/dense-s12",
+      "training_path": "/groups/yshang/an221229/checkpoints/LoopWAM_NT/robotwin10/dense-s12/train",
+      "export_sha256": "c0c0fbb7b4cc607c00f7ee783c1f0e43815daa9fa09b1141c0b456dec255d064",
+      "export_bytes": 2338751909,
+      "step": 45445,
+      "format": "loopwam-s-robot-v1",
+      "optimizer_included": false,
+      "version": "dense_s12",
+      "video_loops": 1,
+      "action_loops": 1,
+      "loop_alignment": "late",
+      "parameters": 584567886,
+      "dataset_scope": "robotwin10",
+      "tasks": [
+        "adjust_bottle",
+        "grab_roller",
+        "hanging_mug",
+        "move_stapler_pad",
+        "open_microwave",
+        "press_stapler",
+        "scan_object",
+        "stack_blocks_two",
+        "stamp_seal",
+        "turn_switch"
+      ],
+      "epochs": 5,
+      "global_batch": 128,
+      "training_seed": 42,
+      "initialization_mode": "fresh_wan_donors_no_embodied_pretraining",
+      "evaluation_status": "not_evaluated",
+      "training_seconds": 28743.092683792114,
+      "final_loss_last100": 0.0485357765853405,
+      "tensor_equality_verified": true,
+      "source_revision": "bea472670ea52bb375ae75ba5380cd88087d2c30",
+      "hf_revision": "828d6802c684d44b834a656367faeb1858abea74"
+    },
+    {
+      "id": "robotwin10_dense_s30",
+      "repo_path": "robotwin-10/dense-s30",
+      "training_path": "/groups/yshang/an221229/checkpoints/LoopWAM_NT/robotwin10/dense-s30/train",
+      "export_sha256": "cc549b65bb849198ab1d11fd78e2a14bd0d1682927665b24c5f82b3ee8b78985",
+      "export_bytes": 5665462576,
+      "step": 45445,
+      "format": "loopwam-s-robot-v1",
+      "optimizer_included": false,
+      "version": "dense_s30",
+      "video_loops": 1,
+      "action_loops": 1,
+      "loop_alignment": "late",
+      "parameters": 1416145998,
+      "dataset_scope": "robotwin10",
+      "tasks": [
+        "adjust_bottle",
+        "grab_roller",
+        "hanging_mug",
+        "move_stapler_pad",
+        "open_microwave",
+        "press_stapler",
+        "scan_object",
+        "stack_blocks_two",
+        "stamp_seal",
+        "turn_switch"
+      ],
+      "epochs": 5,
+      "global_batch": 128,
+      "training_seed": 42,
+      "initialization_mode": "fresh_wan_donors_no_embodied_pretraining",
+      "evaluation_status": "not_evaluated",
+      "training_seconds": 69898.21527791023,
+      "final_loss_last100": 0.03495093105360866,
+      "tensor_equality_verified": true,
+      "source_revision": "bea472670ea52bb375ae75ba5380cd88087d2c30",
+      "hf_revision": "cdf71976208ade41a93465f6b8166b5e6b7170ca"
+    },
+    {
+      "id": "robotwin10_v0_video4_action2",
+      "repo_path": "robotwin-10/v0-video4-action2",
+      "training_path": "/groups/yshang/an221229/checkpoints/LoopWAM_NT/robotwin10/v0-video4-action2/train",
+      "export_sha256": "fea2ce45c09197fe68df9345d1f60e520dab8193dd7ee9b5147001e64efc32c9",
+      "export_bytes": 2338751845,
+      "step": 45445,
+      "format": "loopwam-s-robot-v1",
+      "optimizer_included": false,
+      "version": "v0",
+      "video_loops": 4,
+      "action_loops": 2,
+      "loop_alignment": "late",
+      "parameters": 584567886,
+      "dataset_scope": "robotwin10",
+      "tasks": [
+        "adjust_bottle",
+        "grab_roller",
+        "hanging_mug",
+        "move_stapler_pad",
+        "open_microwave",
+        "press_stapler",
+        "scan_object",
+        "stack_blocks_two",
+        "stamp_seal",
+        "turn_switch"
+      ],
+      "epochs": 5,
+      "global_batch": 128,
+      "training_seed": 42,
+      "initialization_mode": "fresh_wan_donors_no_embodied_pretraining",
+      "evaluation_status": "not_evaluated",
+      "training_seconds": 60219.636355638504,
+      "final_loss_last100": 0.045222468487918374,
+      "tensor_equality_verified": true,
+      "source_revision": "bea472670ea52bb375ae75ba5380cd88087d2c30",
+      "hf_revision": "b24d5ed114c5d812284472c674bee31470a8d8ff"
+    },
+    {
+      "id": "robotwin10_v0_video4_action1",
+      "repo_path": "robotwin-10/v0-video4-action1",
+      "training_path": "/groups/yshang/an221229/checkpoints/LoopWAM_NT/robotwin10/v0-video4-action1/train",
+      "export_sha256": "b0fd9da51d9ff5ab512d5729de6565e85d240a10f078ddf9f803733a77aa757d",
+      "export_bytes": 2338751909,
+      "step": 45445,
+      "format": "loopwam-s-robot-v1",
+      "optimizer_included": false,
+      "version": "v0",
+      "video_loops": 4,
+      "action_loops": 1,
+      "loop_alignment": "late",
+      "parameters": 584567886,
+      "dataset_scope": "robotwin10",
+      "tasks": [
+        "adjust_bottle",
+        "grab_roller",
+        "hanging_mug",
+        "move_stapler_pad",
+        "open_microwave",
+        "press_stapler",
+        "scan_object",
+        "stack_blocks_two",
+        "stamp_seal",
+        "turn_switch"
+      ],
+      "epochs": 5,
+      "global_batch": 128,
+      "training_seed": 42,
+      "initialization_mode": "fresh_wan_donors_no_embodied_pretraining",
+      "evaluation_status": "not_evaluated",
+      "training_seconds": 58643.347306489944,
+      "final_loss_last100": 0.045315117612481116,
+      "tensor_equality_verified": true,
+      "source_revision": "bea472670ea52bb375ae75ba5380cd88087d2c30",
+      "hf_revision": "ba2d4a5e921e2e2e298ea55d811f421932f607b8"
+    },
+    {
+      "id": "robotwin10_v0_video2_action2",
+      "repo_path": "robotwin-10/v0-video2-action2",
+      "training_path": "/groups/yshang/an221229/checkpoints/LoopWAM_NT/robotwin10/v0-video2-action2/train",
+      "export_sha256": "aa835dce6898f92d49275bf6e6b4e8f9202d4a8025a1e810ae3158867369be59",
+      "export_bytes": 2338751845,
+      "step": 45445,
+      "format": "loopwam-s-robot-v1",
+      "optimizer_included": false,
+      "version": "v0",
+      "video_loops": 2,
+      "action_loops": 2,
+      "loop_alignment": "late",
+      "parameters": 584567886,
+      "dataset_scope": "robotwin10",
+      "tasks": [
+        "adjust_bottle",
+        "grab_roller",
+        "hanging_mug",
+        "move_stapler_pad",
+        "open_microwave",
+        "press_stapler",
+        "scan_object",
+        "stack_blocks_two",
+        "stamp_seal",
+        "turn_switch"
+      ],
+      "epochs": 5,
+      "global_batch": 128,
+      "training_seed": 42,
+      "initialization_mode": "fresh_wan_donors_no_embodied_pretraining",
+      "evaluation_status": "not_evaluated",
+      "training_seconds": 40010.85231876373,
+      "final_loss_last100": 0.04690537687391043,
+      "tensor_equality_verified": true,
+      "source_revision": "bea472670ea52bb375ae75ba5380cd88087d2c30",
+      "hf_revision": "34af15d8b3fa65c1938c48e806599df29da848d2"
+    },
+    {
+      "id": "robotwin10_v0_video1_action4",
+      "repo_path": "robotwin-10/v0-video1-action4",
+      "training_path": "/groups/yshang/an221229/checkpoints/LoopWAM_NT/robotwin10/v0-video1-action4/train",
+      "export_sha256": "f9b097164b1fb076a0440f7c42e6439ee87885fbcf21d88975ae3f2d44fd1daf",
+      "export_bytes": 2338751909,
+      "step": 45445,
+      "format": "loopwam-s-robot-v1",
+      "optimizer_included": false,
+      "version": "v0",
+      "video_loops": 1,
+      "action_loops": 4,
+      "loop_alignment": "late",
+      "parameters": 584567886,
+      "dataset_scope": "robotwin10",
+      "tasks": [
+        "adjust_bottle",
+        "grab_roller",
+        "hanging_mug",
+        "move_stapler_pad",
+        "open_microwave",
+        "press_stapler",
+        "scan_object",
+        "stack_blocks_two",
+        "stamp_seal",
+        "turn_switch"
+      ],
+      "epochs": 5,
+      "global_batch": 128,
+      "training_seed": 42,
+      "initialization_mode": "fresh_wan_donors_no_embodied_pretraining",
+      "evaluation_status": "not_evaluated",
+      "training_seconds": 42788.281959056854,
+      "final_loss_last100": 0.0485189451277256,
+      "tensor_equality_verified": true,
+      "source_revision": "bea472670ea52bb375ae75ba5380cd88087d2c30",
+      "hf_revision": "e4715a1148c88570090c5d470243c4e60701c2bb"
     }
   ],
-  "updated_utc": "2026-10-28T19:18:44.471588+00:00"
+  "updated_utc": "2026-11-05T11:35:03.567352+00:00"
 }

robotwin-10/README.md

@@ -0,0 +1,20 @@
+# RoboTwin-10 checkpoints
+
+Seven LoopWAM models trained on **RoboTwin-10**, the 10-task LiLa-WAM ablation subset of the FastWAM RoboTwin 2.0
+release (adjust_bottle, grab_roller, hanging_mug, move_stapler_pad, open_microwave, press_stapler, scan_object, stack_blocks_two, stamp_seal, turn_switch; 50 clean + 500 randomized demonstrations per task, 5,500 episodes,
+1,163,351 windows per epoch). Every model uses the same recipe as `robotwin/v0-video4-action4` restricted to these
+tasks: fresh canonical Wan2.1 donors plus seeded 14-D robot interfaces, exactly 128 windows per update, five epochs
+(45,445 updates), seed 42, identical data order and per-window noise, AdamW 1e-4 with 5% warmup and cosine to
+1e-6. Only the architecture differs. Each folder has its own README, config, provenance, metrics and SHA256SUMS.
+
+| Directory | Model | Video/action loops | Parameters | Training time | Final loss (last 100) | Evaluation |
+|---|---|---|---:|---:|---:|---|
+| [robotwin-10/v0-video4-action4](robotwin-10/v0-video4-action4) | v0 | 4/4 | 584,567,886 | 17.8 h | 0.0453 | Training only; evaluation not run |
+| [robotwin-10/dense-s12](robotwin-10/dense-s12) | dense_s12 | 1/1 | 584,567,886 | 8.0 h | 0.0485 | Training only; evaluation not run |
+| [robotwin-10/dense-s30](robotwin-10/dense-s30) | dense_s30 | 1/1 | 1,416,145,998 | 19.4 h | 0.0350 | Training only; evaluation not run |
+| [robotwin-10/v0-video4-action2](robotwin-10/v0-video4-action2) | v0 aligned | 4/2 | 584,567,886 | 16.7 h | 0.0452 | Training only; evaluation not run |
+| [robotwin-10/v0-video4-action1](robotwin-10/v0-video4-action1) | v0 aligned | 4/1 | 584,567,886 | 16.3 h | 0.0453 | Training only; evaluation not run |
+| [robotwin-10/v0-video2-action2](robotwin-10/v0-video2-action2) | v0 aligned | 2/2 | 584,567,886 | 11.1 h | 0.0469 | Training only; evaluation not run |
+| [robotwin-10/v0-video1-action4](robotwin-10/v0-video1-action4) | v0 aligned | 1/4 | 584,567,886 | 11.9 h | 0.0485 | Training only; evaluation not run |
+
+**Not evaluated:** training-only releases; no RoboTwin success rate is claimed.