Document verified download revision and matched robot checkpoint pairs

README.md

@@ -18,3 +18,7 @@ The MIT model terms retain upstream iVideoGPT and RLVR-World attribution. Datase
 ## Download
 
 Clone [ReCAP](https://github.com/Alexander-wu/ReCAP), install its assets extra, then run `recap download --dataset calvin --root .`. The code release manifest pins an immutable Hub revision and verifies SHA-256 values. Use `rt1`, `calvin`, `libero`, or `bridge` to select a matched pair.
+
+## Version
+
+The verified model tensors are pinned to `61f3c350349f58ba8619c2454b5fac5802756282` in the release manifest. Later model-card updates do not alter that immutable asset revision.

weights/bridge_tokenizer/README.md

@@ -20,7 +20,7 @@ Dataset-specific tokenizer for ReCAP visual open-loop experiments.
 - Initialization: thuml/rt1-compressive-tokenizer (codec) and thuml/rt1-world-model-multi-step-rlvr (world model).
 - Format: diffusers; full strict-load weights, no optimizer/RNG state in inference release.
 - Pair only with `bridge_tokenizer`, `bridge_world`, and `bridge_action_ranges` from the same manifest.
-- Download: https://huggingface.co/easylearning/recap-robot-models/tree/main/weights/bridge_tokenizer
+- Download: [matched model files](https://huggingface.co/easylearning/recap-robot-models/tree/61f3c350349f58ba8619c2454b5fac5802756282/weights/bridge_tokenizer).
 - License: MIT, retaining the starting model terms and attribution. This card does not grant rights over the training dataset. The Apache-2.0 code license is separate from checkpoint licensing.
 - Intended use: research video prediction with supplied actions. Not validated as a deployed robot controller.
 - Metrics: see artifacts/results; selected GIFs are not representative aggregate estimates.

weights/bridge_world/README.md

@@ -20,7 +20,7 @@ Dataset-specific world model for ReCAP visual open-loop experiments.
 - Initialization: thuml/rt1-compressive-tokenizer (codec) and thuml/rt1-world-model-multi-step-rlvr (world model).
 - Format: huggingface; full strict-load weights, no optimizer/RNG state in inference release.
 - Pair only with `bridge_tokenizer`, `bridge_world`, and `bridge_action_ranges` from the same manifest.
-- Download: https://huggingface.co/easylearning/recap-robot-models/tree/main/weights/bridge_world
+- Download: [matched model files](https://huggingface.co/easylearning/recap-robot-models/tree/61f3c350349f58ba8619c2454b5fac5802756282/weights/bridge_world).
 - License: MIT, retaining the starting model terms and attribution. This card does not grant rights over the training dataset. The Apache-2.0 code license is separate from checkpoint licensing.
 - Intended use: research video prediction with supplied actions. Not validated as a deployed robot controller.
 - Metrics: see artifacts/results; selected GIFs are not representative aggregate estimates.

weights/calvin_tokenizer/README.md

@@ -20,7 +20,7 @@ Dataset-specific tokenizer for ReCAP visual open-loop experiments.
 - Initialization: thuml/rt1-compressive-tokenizer (codec) and thuml/rt1-world-model-multi-step-rlvr (world model).
 - Format: diffusers; full strict-load weights, no optimizer/RNG state in inference release.
 - Pair only with `calvin_tokenizer`, `calvin_world`, and `calvin_action_ranges` from the same manifest.
-- Download: https://huggingface.co/easylearning/recap-robot-models/tree/main/weights/calvin_tokenizer
+- Download: [matched model files](https://huggingface.co/easylearning/recap-robot-models/tree/61f3c350349f58ba8619c2454b5fac5802756282/weights/calvin_tokenizer).
 - License: MIT, retaining the starting model terms and attribution. This card does not grant rights over the training dataset. The Apache-2.0 code license is separate from checkpoint licensing.
 - Intended use: research video prediction with supplied actions. Not validated as a deployed robot controller.
 - Metrics: see artifacts/results; selected GIFs are not representative aggregate estimates.

weights/calvin_world/README.md

@@ -20,7 +20,7 @@ Dataset-specific world model for ReCAP visual open-loop experiments.
 - Initialization: thuml/rt1-compressive-tokenizer (codec) and thuml/rt1-world-model-multi-step-rlvr (world model).
 - Format: huggingface; full strict-load weights, no optimizer/RNG state in inference release.
 - Pair only with `calvin_tokenizer`, `calvin_world`, and `calvin_action_ranges` from the same manifest.
-- Download: https://huggingface.co/easylearning/recap-robot-models/tree/main/weights/calvin_world
+- Download: [matched model files](https://huggingface.co/easylearning/recap-robot-models/tree/61f3c350349f58ba8619c2454b5fac5802756282/weights/calvin_world).
 - License: MIT, retaining the starting model terms and attribution. This card does not grant rights over the training dataset. The Apache-2.0 code license is separate from checkpoint licensing.
 - Intended use: research video prediction with supplied actions. Not validated as a deployed robot controller.
 - Metrics: see artifacts/results; selected GIFs are not representative aggregate estimates.

weights/libero_tokenizer/README.md

@@ -20,7 +20,7 @@ Dataset-specific tokenizer for ReCAP visual open-loop experiments.
 - Initialization: thuml/rt1-compressive-tokenizer (codec) and thuml/rt1-world-model-multi-step-rlvr (world model).
 - Format: diffusers; full strict-load weights, no optimizer/RNG state in inference release.
 - Pair only with `libero_tokenizer`, `libero_world`, and `libero_action_ranges` from the same manifest.
-- Download: https://huggingface.co/easylearning/recap-robot-models/tree/main/weights/libero_tokenizer
+- Download: [matched model files](https://huggingface.co/easylearning/recap-robot-models/tree/61f3c350349f58ba8619c2454b5fac5802756282/weights/libero_tokenizer).
 - License: MIT, retaining the starting model terms and attribution. This card does not grant rights over the training dataset. The Apache-2.0 code license is separate from checkpoint licensing.
 - Intended use: research video prediction with supplied actions. Not validated as a deployed robot controller.
 - Metrics: see artifacts/results; selected GIFs are not representative aggregate estimates.

weights/libero_world/README.md

@@ -20,7 +20,7 @@ Dataset-specific world model for ReCAP visual open-loop experiments.
 - Initialization: thuml/rt1-compressive-tokenizer (codec) and thuml/rt1-world-model-multi-step-rlvr (world model).
 - Format: huggingface; full strict-load weights, no optimizer/RNG state in inference release.
 - Pair only with `libero_tokenizer`, `libero_world`, and `libero_action_ranges` from the same manifest.
-- Download: https://huggingface.co/easylearning/recap-robot-models/tree/main/weights/libero_world
+- Download: [matched model files](https://huggingface.co/easylearning/recap-robot-models/tree/61f3c350349f58ba8619c2454b5fac5802756282/weights/libero_world).
 - License: MIT, retaining the starting model terms and attribution. This card does not grant rights over the training dataset. The Apache-2.0 code license is separate from checkpoint licensing.
 - Intended use: research video prediction with supplied actions. Not validated as a deployed robot controller.
 - Metrics: see artifacts/results; selected GIFs are not representative aggregate estimates.

weights/rt1_tokenizer/README.md

@@ -20,7 +20,7 @@ Dataset-specific tokenizer for ReCAP visual open-loop experiments.
 - Initialization: thuml/rt1-compressive-tokenizer (codec) and thuml/rt1-world-model-multi-step-rlvr (world model).
 - Format: diffusers; full strict-load weights, no optimizer/RNG state in inference release.
 - Pair only with `rt1_tokenizer`, `rt1_world`, and `rt1_action_ranges` from the same manifest.
-- Download: https://huggingface.co/easylearning/recap-robot-models/tree/main/weights/rt1_tokenizer
+- Download: [matched model files](https://huggingface.co/easylearning/recap-robot-models/tree/61f3c350349f58ba8619c2454b5fac5802756282/weights/rt1_tokenizer).
 - License: MIT, retaining the starting model terms and attribution. This card does not grant rights over the training dataset. The Apache-2.0 code license is separate from checkpoint licensing.
 - Intended use: research video prediction with supplied actions. Not validated as a deployed robot controller.
 - Metrics: see artifacts/results; selected GIFs are not representative aggregate estimates.

weights/rt1_world/README.md

@@ -20,7 +20,7 @@ Dataset-specific world model for ReCAP visual open-loop experiments.
 - Initialization: thuml/rt1-compressive-tokenizer (codec) and thuml/rt1-world-model-multi-step-rlvr (world model).
 - Format: huggingface; full strict-load weights, no optimizer/RNG state in inference release.
 - Pair only with `rt1_tokenizer`, `rt1_world`, and `rt1_action_ranges` from the same manifest.
-- Download: https://huggingface.co/easylearning/recap-robot-models/tree/main/weights/rt1_world
+- Download: [matched model files](https://huggingface.co/easylearning/recap-robot-models/tree/61f3c350349f58ba8619c2454b5fac5802756282/weights/rt1_world).
 - License: MIT, retaining the starting model terms and attribution. This card does not grant rights over the training dataset. The Apache-2.0 code license is separate from checkpoint licensing.
 - Intended use: research video prediction with supplied actions. Not validated as a deployed robot controller.
 - Metrics: see artifacts/results; selected GIFs are not representative aggregate estimates.

weights_manifest.json

@@ -15,7 +15,8 @@
         }
       ],
       "format": "diffusers",
-      "published": true
+      "published": true,
+      "url": "https://huggingface.co/easylearning/recap-robot-models/tree/61f3c350349f58ba8619c2454b5fac5802756282/weights/rt1_tokenizer"
     },
     "rt1_world": {
       "files": [
@@ -46,7 +47,8 @@
         }
       ],
       "format": "huggingface",
-      "published": true
+      "published": true,
+      "url": "https://huggingface.co/easylearning/recap-robot-models/tree/61f3c350349f58ba8619c2454b5fac5802756282/weights/rt1_world"
     },
     "calvin_tokenizer": {
       "files": [
@@ -62,7 +64,8 @@
         }
       ],
       "format": "diffusers",
-      "published": true
+      "published": true,
+      "url": "https://huggingface.co/easylearning/recap-robot-models/tree/61f3c350349f58ba8619c2454b5fac5802756282/weights/calvin_tokenizer"
     },
     "calvin_world": {
       "files": [
@@ -93,7 +96,8 @@
         }
       ],
       "format": "huggingface",
-      "published": true
+      "published": true,
+      "url": "https://huggingface.co/easylearning/recap-robot-models/tree/61f3c350349f58ba8619c2454b5fac5802756282/weights/calvin_world"
     },
     "libero_tokenizer": {
       "files": [
@@ -109,7 +113,8 @@
         }
       ],
       "format": "diffusers",
-      "published": true
+      "published": true,
+      "url": "https://huggingface.co/easylearning/recap-robot-models/tree/61f3c350349f58ba8619c2454b5fac5802756282/weights/libero_tokenizer"
     },
     "libero_world": {
       "files": [
@@ -140,7 +145,8 @@
         }
       ],
       "format": "huggingface",
-      "published": true
+      "published": true,
+      "url": "https://huggingface.co/easylearning/recap-robot-models/tree/61f3c350349f58ba8619c2454b5fac5802756282/weights/libero_world"
     },
     "bridge_tokenizer": {
       "files": [
@@ -156,7 +162,8 @@
         }
       ],
       "format": "diffusers",
-      "published": true
+      "published": true,
+      "url": "https://huggingface.co/easylearning/recap-robot-models/tree/61f3c350349f58ba8619c2454b5fac5802756282/weights/bridge_tokenizer"
     },
     "bridge_world": {
       "files": [
@@ -187,7 +194,8 @@
         }
       ],
       "format": "huggingface",
-      "published": true
+      "published": true,
+      "url": "https://huggingface.co/easylearning/recap-robot-models/tree/61f3c350349f58ba8619c2454b5fac5802756282/weights/bridge_world"
     },
     "calvin_action_ranges": {
       "files": [
@@ -202,7 +210,8 @@
           "bytes": 379
         }
       ],
-      "published": true
+      "published": true,
+      "url": "https://huggingface.co/easylearning/recap-robot-models/tree/61f3c350349f58ba8619c2454b5fac5802756282/weights/calvin"
     },
     "libero_action_ranges": {
       "files": [
@@ -217,7 +226,8 @@
           "bytes": 384
         }
       ],
-      "published": true
+      "published": true,
+      "url": "https://huggingface.co/easylearning/recap-robot-models/tree/61f3c350349f58ba8619c2454b5fac5802756282/weights/libero"
     },
     "bridge_action_ranges": {
       "files": [
@@ -232,7 +242,8 @@
           "bytes": 417
         }
       ],
-      "published": true
+      "published": true,
+      "url": "https://huggingface.co/easylearning/recap-robot-models/tree/61f3c350349f58ba8619c2454b5fac5802756282/weights/bridge"
     },
     "rt1_action_ranges": {
       "files": [
@@ -247,13 +258,14 @@
           "sha256": "a55e18454996743ff27b4680ca21ba6bb8ff1731e7ab63935e5783215c492172"
         }
       ],
-      "published": true
+      "published": true,
+      "url": "https://huggingface.co/easylearning/recap-robot-models/tree/61f3c350349f58ba8619c2454b5fac5802756282/weights/rt1"
     }
   },
   "distribution": {
     "repo_id": "easylearning/recap-robot-models",
+    "revision": "61f3c350349f58ba8619c2454b5fac5802756282",
     "repo_type": "model",
-    "status": "published",
-    "note": "Use the immutable revision recorded by the GitHub release manifest."
+    "url": "https://huggingface.co/easylearning/recap-robot-models"
   }
 }