Publish verified EMAGE metrics and Three.js comparison

.gitattributes

@@ -96,3 +96,18 @@ media/2_scott_0_65_65_language_of_motion.mp4 filter=lfs diff=lfs merge=lfs -text
 media/2_scott_0_95_95_audio.wav filter=lfs diff=lfs merge=lfs -text
 media/2_scott_0_95_95_gt.mp4 filter=lfs diff=lfs merge=lfs -text
 media/2_scott_0_95_95_language_of_motion.mp4 filter=lfs diff=lfs merge=lfs -text
+cases/assets/2_scott_0_103_103_emage.smpl filter=lfs diff=lfs merge=lfs -text
+cases/assets/2_scott_0_111_111_emage.smpl filter=lfs diff=lfs merge=lfs -text
+cases/assets/2_scott_0_1_1_emage.smpl filter=lfs diff=lfs merge=lfs -text
+cases/assets/2_scott_0_2_2_emage.smpl filter=lfs diff=lfs merge=lfs -text
+cases/assets/2_scott_0_3_3_emage.smpl filter=lfs diff=lfs merge=lfs -text
+cases/assets/2_scott_0_4_4_emage.smpl filter=lfs diff=lfs merge=lfs -text
+cases/assets/2_scott_0_5_5_emage.smpl filter=lfs diff=lfs merge=lfs -text
+cases/assets/2_scott_0_65_65_emage.smpl filter=lfs diff=lfs merge=lfs -text
+cases/assets/2_scott_0_6_6_emage.smpl filter=lfs diff=lfs merge=lfs -text
+cases/assets/2_scott_0_73_73_emage.smpl filter=lfs diff=lfs merge=lfs -text
+cases/assets/2_scott_0_7_7_emage.smpl filter=lfs diff=lfs merge=lfs -text
+cases/assets/2_scott_0_81_81_emage.smpl filter=lfs diff=lfs merge=lfs -text
+cases/assets/2_scott_0_87_87_emage.smpl filter=lfs diff=lfs merge=lfs -text
+cases/assets/2_scott_0_8_8_emage.smpl filter=lfs diff=lfs merge=lfs -text
+cases/assets/2_scott_0_95_95_emage.smpl filter=lfs diff=lfs merge=lfs -text

README.md

@@ -20,6 +20,10 @@ numbers are not inserted as submissions.
 The public LoM checkpoint is explicitly marked as visualization-only. Its
 measured metrics remain useful for adapter verification, but are not presented
 as a reproduction of the paper's private/full evaluation checkpoint.
+EMAGE uses the maintained official PantoMatrix audio checkpoint pinned by
+revision. On the same 15 paired clips it measures `FGD = 6.199`, `BC = 7.564`,
+and `Diversity = 12.476`; its paper values remain a separate reference because
+the paper model also consumed text.
 
 GT is displayed as a reference and is excluded from ranking.
 

cases/assets/2_scott_0_103_103_emage.smpl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bb3e2f4fac5495ef73d285ddba5dd067332260b7692e830a434d96c2e2f1028e
+size 459810

cases/assets/2_scott_0_111_111_emage.smpl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:858e373482aaf080269e23fcaa99689cbaf642eae25d296132106a280e55e104
+size 484380

cases/assets/2_scott_0_1_1_emage.smpl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5f6fa4d2d9bc6ed862e04116b2b7f9e5f148bee5e49c76677a2e3d409ba7ebc6
+size 514620

cases/assets/2_scott_0_2_2_emage.smpl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:75d1319bd55628549fede9849593183887e147cf010feae4cef35ce1f08683b7
+size 500580

cases/assets/2_scott_0_3_3_emage.smpl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:509d1d98cfa3eb3124145bc011b4cd0350cd89c2df54d13a0bd3588b381eac98
+size 549180

cases/assets/2_scott_0_4_4_emage.smpl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:727aef762f52581f2bc3b0e38932e68d2021c29cc1e566fe51cfb4f689c9283d
+size 541080

cases/assets/2_scott_0_5_5_emage.smpl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7e2ef7c6fc5239aaf6915c0b7490032130aa53d46c442e62e5b2b68c3096c5a7
+size 622080

cases/assets/2_scott_0_65_65_emage.smpl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c28f05c165c93145d5e44a25bd963e37b77fb46b2b11383fe7b6fb7e05e5a23d
+size 427680

cases/assets/2_scott_0_6_6_emage.smpl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4b98f92884c574cec69e1b0ec23e836683e372b815b4a668e793347eff6ee28d
+size 670680

cases/assets/2_scott_0_73_73_emage.smpl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:45b3e11f0848d45a8f3b8bbea1d9e9d356fb04f8dfc19303ef0d694aa75a69d1
+size 476280

cases/assets/2_scott_0_7_7_emage.smpl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:99195963f4d2826cbd9daa3f1e8199a3400f5377df33dd53372eb00cd2bc08b4
+size 565380

cases/assets/2_scott_0_81_81_emage.smpl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:52281932571fc50aab5e160f545cf2d0bafce6f4001544c202053005aec9f15d
+size 540810

cases/assets/2_scott_0_87_87_emage.smpl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:83ee5b750abbf65a6e9b71abe11e2a81f6e2642fc2765d1bc0474b0f4002d8b4
+size 630180

cases/assets/2_scott_0_8_8_emage.smpl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8dc5d16bcc5109dcfce986a3dd14acb2150aa8c104b1608febf7a07552d09154
+size 628290

cases/assets/2_scott_0_95_95_emage.smpl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6e01602162fc3d7e9eeda9836d50a8f3e0d828b540d6ba09a4a69f7ea99b3745
+size 395280

cases/manifest.json

@@ -2,7 +2,7 @@
   "schema_version": 2,
   "representation": "smpl_motion135",
   "task": "speech_to_gesture",
-  "title": "Language of Motion on BEAT2 speaker 2",
+  "title": "Speech-to-Gesture on BEAT2 speaker 2",
   "protocol": "BEAT2 official test clips, speaker 2",
   "population": 15,
   "asset_base_url": "./",
@@ -18,15 +18,20 @@
       "key": "language_of_motion",
       "label": "Language of Motion",
       "accent": "#087d72"
+    },
+    {
+      "key": "emage",
+      "label": "EMAGE",
+      "accent": "#315f9d"
     }
   ],
   "checkpoint_scope": "public_demo_visualization_only",
   "provenance": {
     "dataset": "BEAT2 English v2.0.0",
     "speaker": 2,
-    "artifact": "JuzeZhang/language_of_motion",
+    "artifact": "ZeyuLing/Motius-Language-of-Motion",
     "seed_policy": "reset_official_seed_before_each_sample",
-    "visualization": "Shared neutral SMPL-22 body projection of the native SMPL-H/SMPL-X poses; LoM fingers and expressions are retained in raw/language_of_motion but not rendered by this web mesh."
+    "visualization": "Shared neutral SMPL-22 body projection of the native SMPL-H/SMPL-X poses; LoM and EMAGE fingers and expressions are retained in raw/ but not rendered by this web mesh."
   },
   "cases": [
     {
@@ -78,6 +83,27 @@
           "translation_offset": 0,
           "rotation_offset": 10218,
           "fps": 30.0
+        },
+        "emage": {
+          "frames": 1703,
+          "display_frames": 1703,
+          "stride": 1,
+          "translation_count": 5109,
+          "rotation_count": 224796,
+          "translation_minimum": [
+            -0.09501680731773376,
+            1.2184923887252808,
+            -7.0816264152526855
+          ],
+          "translation_scale": [
+            1.1390543477318715e-05,
+            2.0440222670004005e-06,
+            0.0001087606287910603
+          ],
+          "asset": "assets/2_scott_0_103_103_emage.smpl",
+          "translation_offset": 0,
+          "rotation_offset": 10218,
+          "fps": 30.0
         }
       }
     },
@@ -130,6 +156,27 @@
           "translation_offset": 0,
           "rotation_offset": 10764,
           "fps": 30.0
+        },
+        "emage": {
+          "frames": 1794,
+          "display_frames": 1794,
+          "stride": 1,
+          "translation_count": 5382,
+          "rotation_count": 236808,
+          "translation_minimum": [
+            -0.0026984205469489098,
+            1.2143173217773438,
+            -3.9132707118988037
+          ],
+          "translation_scale": [
+            5.110363781568594e-05,
+            2.046983581749373e-06,
+            6.014310201862827e-05
+          ],
+          "asset": "assets/2_scott_0_111_111_emage.smpl",
+          "translation_offset": 0,
+          "rotation_offset": 10764,
+          "fps": 30.0
         }
       }
     },
@@ -182,6 +229,27 @@
           "translation_offset": 0,
           "rotation_offset": 11436,
           "fps": 30.0
+        },
+        "emage": {
+          "frames": 1906,
+          "display_frames": 1906,
+          "stride": 1,
+          "translation_count": 5718,
+          "rotation_count": 251592,
+          "translation_minimum": [
+            -0.28490129113197327,
+            1.2197376489639282,
+            -5.417672634124756
+          ],
+          "translation_scale": [
+            1.3098448107484728e-05,
+            1.8892112620960688e-06,
+            8.266838267445564e-05
+          ],
+          "asset": "assets/2_scott_0_1_1_emage.smpl",
+          "translation_offset": 0,
+          "rotation_offset": 11436,
+          "fps": 30.0
         }
       }
     },
@@ -234,6 +302,27 @@
           "translation_offset": 0,
           "rotation_offset": 11124,
           "fps": 30.0
+        },
+        "emage": {
+          "frames": 1854,
+          "display_frames": 1854,
+          "stride": 1,
+          "translation_count": 5562,
+          "rotation_count": 244728,
+          "translation_minimum": [
+            -0.006783625110983849,
+            1.2228302955627441,
+            -6.143184185028076
+          ],
+          "translation_scale": [
+            2.3385980966850184e-05,
+            1.8716631302595488e-06,
+            9.373898501507938e-05
+          ],
+          "asset": "assets/2_scott_0_2_2_emage.smpl",
+          "translation_offset": 0,
+          "rotation_offset": 11124,
+          "fps": 30.0
         }
       }
     },
@@ -286,6 +375,27 @@
           "translation_offset": 0,
           "rotation_offset": 12204,
           "fps": 30.0
+        },
+        "emage": {
+          "frames": 2034,
+          "display_frames": 2034,
+          "stride": 1,
+          "translation_count": 6102,
+          "rotation_count": 268488,
+          "translation_minimum": [
+            -0.03589795157313347,
+            1.146770715713501,
+            -7.534869194030762
+          ],
+          "translation_scale": [
+            4.922703010379337e-05,
+            2.9561283554357942e-06,
+            0.0001149747331510298
+          ],
+          "asset": "assets/2_scott_0_3_3_emage.smpl",
+          "translation_offset": 0,
+          "rotation_offset": 12204,
+          "fps": 30.0
         }
       }
     },
@@ -338,6 +448,27 @@
           "translation_offset": 0,
           "rotation_offset": 12024,
           "fps": 30.0
+        },
+        "emage": {
+          "frames": 2004,
+          "display_frames": 2004,
+          "stride": 1,
+          "translation_count": 6012,
+          "rotation_count": 264528,
+          "translation_minimum": [
+            -0.007614895701408386,
+            1.2340558767318726,
+            -6.051639080047607
+          ],
+          "translation_scale": [
+            2.5403154722880572e-05,
+            1.6880807152119814e-06,
+            9.234209574060515e-05
+          ],
+          "asset": "assets/2_scott_0_4_4_emage.smpl",
+          "translation_offset": 0,
+          "rotation_offset": 12024,
+          "fps": 30.0
         }
       }
     },
@@ -390,6 +521,27 @@
           "translation_offset": 0,
           "rotation_offset": 13824,
           "fps": 30.0
+        },
+        "emage": {
+          "frames": 2304,
+          "display_frames": 2304,
+          "stride": 1,
+          "translation_count": 6912,
+          "rotation_count": 304128,
+          "translation_minimum": [
+            -0.016887430101633072,
+            1.0714370012283325,
+            -5.762335777282715
+          ],
+          "translation_scale": [
+            9.422039875062183e-05,
+            4.203217486065114e-06,
+            8.792760490905493e-05
+          ],
+          "asset": "assets/2_scott_0_5_5_emage.smpl",
+          "translation_offset": 0,
+          "rotation_offset": 13824,
+          "fps": 30.0
         }
       }
     },
@@ -442,6 +594,27 @@
           "translation_offset": 0,
           "rotation_offset": 9504,
           "fps": 30.0
+        },
+        "emage": {
+          "frames": 1584,
+          "display_frames": 1584,
+          "stride": 1,
+          "translation_count": 4752,
+          "rotation_count": 209088,
+          "translation_minimum": [
+            -0.022363023832440376,
+            1.2545968294143677,
+            -4.660858631134033
+          ],
+          "translation_scale": [
+            4.774642002303153e-05,
+            1.3576707260654075e-06,
+            7.112014282029122e-05
+          ],
+          "asset": "assets/2_scott_0_65_65_emage.smpl",
+          "translation_offset": 0,
+          "rotation_offset": 9504,
+          "fps": 30.0
         }
       }
     },
@@ -494,6 +667,27 @@
           "translation_offset": 0,
           "rotation_offset": 14904,
           "fps": 30.0
+        },
+        "emage": {
+          "frames": 2484,
+          "display_frames": 2484,
+          "stride": 1,
+          "translation_count": 7452,
+          "rotation_count": 327888,
+          "translation_minimum": [
+            -0.5164570212364197,
+            1.232894778251648,
+            -7.773210048675537
+          ],
+          "translation_scale": [
+            1.4914938219590113e-05,
+            1.8072317971018492e-06,
+            0.00011861158418469131
+          ],
+          "asset": "assets/2_scott_0_6_6_emage.smpl",
+          "translation_offset": 0,
+          "rotation_offset": 14904,
+          "fps": 30.0
         }
       }
     },
@@ -546,6 +740,27 @@
           "translation_offset": 0,
           "rotation_offset": 10584,
           "fps": 30.0
+        },
+        "emage": {
+          "frames": 1764,
+          "display_frames": 1764,
+          "stride": 1,
+          "translation_count": 5292,
+          "rotation_count": 232848,
+          "translation_minimum": [
+            -0.37450623512268066,
+            1.1181925535202026,
+            -5.255575656890869
+          ],
+          "translation_scale": [
+            1.3074798516754527e-05,
+            3.3883759442687733e-06,
+            8.019494271138683e-05
+          ],
+          "asset": "assets/2_scott_0_73_73_emage.smpl",
+          "translation_offset": 0,
+          "rotation_offset": 10584,
+          "fps": 30.0
         }
       }
     },
@@ -598,6 +813,27 @@
           "translation_offset": 0,
           "rotation_offset": 12564,
           "fps": 30.0
+        },
+        "emage": {
+          "frames": 2094,
+          "display_frames": 2094,
+          "stride": 1,
+          "translation_count": 6282,
+          "rotation_count": 276408,
+          "translation_minimum": [
+            -0.1412588357925415,
+            1.2360010147094727,
+            -5.455987930297852
+          ],
+          "translation_scale": [
+            2.534701707190834e-05,
+            1.800605105017894e-06,
+            8.325304224854335e-05
+          ],
+          "asset": "assets/2_scott_0_7_7_emage.smpl",
+          "translation_offset": 0,
+          "rotation_offset": 12564,
+          "fps": 30.0
         }
       }
     },
@@ -650,6 +886,27 @@
           "translation_offset": 0,
           "rotation_offset": 12018,
           "fps": 30.0
+        },
+        "emage": {
+          "frames": 2003,
+          "display_frames": 2003,
+          "stride": 1,
+          "translation_count": 6009,
+          "rotation_count": 264396,
+          "translation_minimum": [
+            -0.12677711248397827,
+            1.2433542013168335,
+            -6.3567423820495605
+          ],
+          "translation_scale": [
+            7.399420155707048e-06,
+            1.679549427535676e-06,
+            9.699766815174371e-05
+          ],
+          "asset": "assets/2_scott_0_81_81_emage.smpl",
+          "translation_offset": 0,
+          "rotation_offset": 12018,
+          "fps": 30.0
         }
       }
     },
@@ -702,6 +959,27 @@
           "translation_offset": 0,
           "rotation_offset": 14004,
           "fps": 30.0
+        },
+        "emage": {
+          "frames": 2334,
+          "display_frames": 2334,
+          "stride": 1,
+          "translation_count": 7002,
+          "rotation_count": 308088,
+          "translation_minimum": [
+            -0.04473575949668884,
+            1.1370047330856323,
+            -5.367419242858887
+          ],
+          "translation_scale": [
+            8.168394560925663e-05,
+            3.1432944069820223e-06,
+            8.190156950149685e-05
+          ],
+          "asset": "assets/2_scott_0_87_87_emage.smpl",
+          "translation_offset": 0,
+          "rotation_offset": 14004,
+          "fps": 30.0
         }
       }
     },
@@ -754,6 +1032,27 @@
           "translation_offset": 0,
           "rotation_offset": 13962,
           "fps": 30.0
+        },
+        "emage": {
+          "frames": 2327,
+          "display_frames": 2327,
+          "stride": 1,
+          "translation_count": 6981,
+          "rotation_count": 307164,
+          "translation_minimum": [
+            -0.009003915823996067,
+            1.2368403673171997,
+            -6.321014881134033
+          ],
+          "translation_scale": [
+            2.471315565344412e-05,
+            1.810556909731531e-06,
+            9.645250247558579e-05
+          ],
+          "asset": "assets/2_scott_0_8_8_emage.smpl",
+          "translation_offset": 0,
+          "rotation_offset": 13962,
+          "fps": 30.0
         }
       }
     },
@@ -806,6 +1105,27 @@
           "translation_offset": 0,
           "rotation_offset": 8784,
           "fps": 30.0
+        },
+        "emage": {
+          "frames": 1464,
+          "display_frames": 1464,
+          "stride": 1,
+          "translation_count": 4392,
+          "rotation_count": 193248,
+          "translation_minimum": [
+            -0.09548475593328476,
+            1.1936589479446411,
+            -3.5384674072265625
+          ],
+          "translation_scale": [
+            6.725939601892605e-05,
+            2.283377625644789e-06,
+            5.399354995461181e-05
+          ],
+          "asset": "assets/2_scott_0_95_95_emage.smpl",
+          "translation_offset": 0,
+          "rotation_offset": 8784,
+          "fps": 30.0
         }
       }
     }

cases/raw/emage/2_scott_0_103_103.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:449f761d00cdda11de7547036fcb0dd4ffeea2d7a78b1fdf7c69a147a3e7dfb5
+size 1595381

cases/raw/emage/2_scott_0_111_111.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b031a197d55b9b83bc8707cabda07745a0159ee9c58a23f2c05193dc85952ee9
+size 1665735

cases/raw/emage/2_scott_0_1_1.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:267762337a6180337c78302cff70fa59280c28618e72291243d40ad7ddd7f006
+size 1790928

cases/raw/emage/2_scott_0_2_2.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2068968353e2de44fea60e8fbf7b769f4e4622221bbf951efefdaf98ff7cdf3e
+size 1712274

cases/raw/emage/2_scott_0_3_3.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fc74bbcba2269dd26f96b68631324456bcc5bbe2c53f15ed23a1f297b5817e53
+size 1881315

cases/raw/emage/2_scott_0_4_4.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d00c751b2f8cbf8f107e5da98ef63b73e0c1a2bb4fe8058de60ce9937c7c5484
+size 1879385

cases/raw/emage/2_scott_0_5_5.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:11381de07a9585c964709454d719bbf3b4ed1189bcc4a1096f650dd0341e7133
+size 2149296

cases/raw/emage/2_scott_0_65_65.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:78e50a5457aa36ec0a8a432a0e17d929cb843b8a0ac5aaf6bf734c90ba2a8053
+size 1461526

cases/raw/emage/2_scott_0_6_6.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:144dea0e12dfa9753a34348f945c04c078d814b6b2781bc40d4b58369a0c7e96
+size 2284586

cases/raw/emage/2_scott_0_73_73.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0ebd7733c38a67ea5ea8ff5e70bc75e82769852e3abc718786762edead250968
+size 1646111

cases/raw/emage/2_scott_0_7_7.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d04cc6feb22068b6738463878345f5290fb3f0f84e39a0f57ba24c44ef58d887
+size 1917566

cases/raw/emage/2_scott_0_81_81.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1093c1b721f8da417333e838bd0c589e7cc876939b684089fcdcf4f13bbe1435
+size 1813299

cases/raw/emage/2_scott_0_87_87.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8b6b78d146e9b32b965385604823999f87b66aa0565339d7c690b25c072f3317
+size 2171154

cases/raw/emage/2_scott_0_8_8.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6a95a29c61fb9306158c6e83c0647ebc883118628eb334802e6d90d7371315d2
+size 2141649

cases/raw/emage/2_scott_0_95_95.npz

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:adcf4b2656c322718105cb95b79e90076dd16854e225cd88d4cefafd5ba5ca1e
+size 1377173

index.html

@@ -56,7 +56,7 @@
     </section>
 
     <section class="panel" id="case-explorer">
-      <div class="section-heading"><div><h2>Interactive Three.js Motion Comparison</h2><p>High-resolution, camera-controllable GT and Language of Motion previews across all 15 clips. Full native SMPL-X predictions remain downloadable as NPZ.</p></div><a class="section-link" href="cases/index.html" target="_blank" rel="noopener noreferrer">Open full screen</a></div>
+      <div class="section-heading"><div><h2>Interactive Three.js Motion Comparison</h2><p>High-resolution, camera-controllable GT, Language of Motion, and EMAGE previews across all 15 clips. Full native SMPL-X predictions remain downloadable as NPZ.</p></div><a class="section-link" href="cases/index.html" target="_blank" rel="noopener noreferrer">Open full screen</a></div>
       <iframe class="explorer-frame" src="cases/index.html" title="BEAT2 audio-synchronized all-case 3D motion comparison" loading="lazy" allow="autoplay"></iframe>
     </section>
 

speech_to_gesture_results.json

@@ -56,7 +56,7 @@
       "verified": true,
       "submission_status": "public_demo_verified",
       "paper_table_eligible": false,
-      "checkpoint": "https://huggingface.co/JuzeZhang/language_of_motion",
+      "checkpoint": "https://huggingface.co/ZeyuLing/Motius-Language-of-Motion",
       "model_card": "https://github.com/ZeyuLing/Motius/blob/main/docs/model_zoo/language_of_motion.md",
       "paper": "https://arxiv.org/abs/2412.10523",
       "fgd": 5.334046622183627,
@@ -94,6 +94,75 @@
           "output_trim": "paired_reference_frame_count"
         }
       }
+    },
+    {
+      "method": "EMAGE",
+      "version": "official PantoMatrix audio checkpoint",
+      "kind": "method",
+      "verified": true,
+      "submission_status": "official_checkpoint_verified",
+      "paper_table_eligible": false,
+      "checkpoint": "https://huggingface.co/ZeyuLing/Motius-EMAGE-BEAT2",
+      "model_card": "https://github.com/ZeyuLing/Motius/blob/main/docs/model_zoo/emage.md",
+      "paper": "https://arxiv.org/abs/2401.00374",
+      "fgd": 6.19906299336642,
+      "bc": 7.564133316712631,
+      "diversity": 12.475604227512221,
+      "rotation_geodesic": 0.27046809122906024,
+      "expression_l2": 4.743729881237994,
+      "translation_l2": 3.2984432561061547,
+      "raw_metrics": {
+        "FGD": 0.619906299336642,
+        "BC": 0.7564133316712631,
+        "Diversity": 12.475604227512221
+      },
+      "generation_audit": {
+        "max_static_tail_frames": 1,
+        "minimum_active_coverage": 0.9993169398907104,
+        "checkpoint_provenance": {
+          "pytorch_model.bin": {
+            "bytes": 556156301,
+            "sha256": "09c98da6f13445b8a1de1a82cc66cc0582026eb6d1aacdf278f03ef9b665a37a",
+            "revision": "b04466b83bdebd656844a2c416797d3131e34ede"
+          },
+          "emage_vq/face/pytorch_model.bin": {
+            "bytes": 8932994,
+            "sha256": "1ddee4ad1c66a7c21d4f44a4d5ee020b970d9e6532713ec9896de1720f1531c3",
+            "revision": "b04466b83bdebd656844a2c416797d3131e34ede"
+          },
+          "emage_vq/upper/pytorch_model.bin": {
+            "bytes": 8698882,
+            "sha256": "fe2f67701f9fc63c7f1bb0857d37dec3a6cdb082e2fe92b4556d1f9cc70afb88",
+            "revision": "b04466b83bdebd656844a2c416797d3131e34ede"
+          },
+          "emage_vq/lower/pytorch_model.bin": {
+            "bytes": 14870770,
+            "sha256": "eafc580ae68c43ada6f87c83ae96924094ab7e88befbc62bd952655a576f3027",
+            "revision": "b04466b83bdebd656844a2c416797d3131e34ede"
+          },
+          "emage_vq/hands/pytorch_model.bin": {
+            "bytes": 9642306,
+            "sha256": "72872387623e13c45dea2307d61b029506fd08f5b0eb0824600fc138ad0290fb",
+            "revision": "b04466b83bdebd656844a2c416797d3131e34ede"
+          },
+          "emage_vq/global/pytorch_model.bin": {
+            "bytes": 14608303,
+            "sha256": "42d4d883ff611bc82926a246b6e21791ae0eae73efef86248b2221a19c2c5a9f",
+            "revision": "b04466b83bdebd656844a2c416797d3131e34ede"
+          }
+        },
+        "inference_profile": {
+          "protocol": "pantomatrix_emage_audio_main",
+          "official_profile": true,
+          "checkpoint_revision": "b04466b83bdebd656844a2c416797d3131e34ede",
+          "runtime_revision": "c7356f35f8e39e469e510ccd1bf37e44adf8ec0e",
+          "window_frames": 64,
+          "seed_frames": 4,
+          "audio_sample_rate": 16000,
+          "motion_fps": 30,
+          "output_trim": "paired_reference_frame_count"
+        }
+      }
     }
   ]
 }