Add files using upload-large-folder tool

EchoMask_all_speakers_results_npz_25spk.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:72add17f387ebf198361732eef54f517d83c54f603534de23dcf4d2b43c00298
+size 1274002235

EchoMask_testnpz.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c20a437b67f19516d1f6e13ca87642e9791e06178244ad1fb9edbb3b6bbc785d
+size 27938262

README.md

@@ -27,6 +27,8 @@ Generated SMPL-X predictions from the official EchoMask checkpoints for comparis
 | `EchoMask_testnpz.zip` | Speaker 2, paper protocol | Generated test `.npz` files |
 | `EchoMask_all_speakers_results_npz_25spk.zip` | 25 English BEAT2 speakers | Generated test `.npz` files |
 
+Verify both downloads with `sha256sum -c SHA256SUMS`.
+
 The two releases use different training protocols and should not be interpreted as a controlled single-speaker versus multi-speaker ablation.
 
 ## Provenance and terms

SHA256SUMS

@@ -0,0 +1,2 @@
+c20a437b67f19516d1f6e13ca87642e9791e06178244ad1fb9edbb3b6bbc785d  EchoMask_testnpz.zip
+72add17f387ebf198361732eef54f517d83c54f603534de23dcf4d2b43c00298  EchoMask_all_speakers_results_npz_25spk.zip

manifest.json

@@ -0,0 +1,20 @@
+{
+  "schema_version": 1,
+  "paper": "https://arxiv.org/abs/2504.09209",
+  "source_code": "https://github.com/Xiangyue-Zhang/EchoMask",
+  "contains_ground_truth": false,
+  "files": [
+    {
+      "path": "EchoMask_testnpz.zip",
+      "protocol": "speaker2-paper",
+      "size_bytes": 27938262,
+      "sha256": "c20a437b67f19516d1f6e13ca87642e9791e06178244ad1fb9edbb3b6bbc785d"
+    },
+    {
+      "path": "EchoMask_all_speakers_results_npz_25spk.zip",
+      "protocol": "all-speakers-25",
+      "size_bytes": 1274002235,
+      "sha256": "72add17f387ebf198361732eef54f517d83c54f603534de23dcf4d2b43c00298"
+    }
+  ]
+}