Upload folder using huggingface_hub

CODE_CHANGE_LOG.txt

@@ -139,3 +139,13 @@ Validation
   this session. User's earlier 12/12 GPU run remains the GPU evidence.
   Public source/staged filename and text scan, Python syntax, manifest config
   references, artifact hashes and strict checkpoint loading all passed.
+
+2026-09-30 — Pin refreshed public weight bundle
+==============================================
+- User supplied successful HF upload commit
+  aa51bb453ec2aac74e74e1d84e2d0903625575e7 after the terminology cleanup.
+- Pinned all six manifest URLs to that revision using set_weight_urls.py.
+  Verified staged checksum/size metadata and preserved all weight hashes.
+- Updated README.md, docs/release_audit.md and docs/publishing.md status;
+  refreshed staged manifest/docs/log. Public download verification remains
+  pending. User executes remaining metadata upload and GitHub commit/push.

docs/publishing.md

@@ -16,8 +16,9 @@ This follows the CLI's folder synchronization behavior: local configs are
 uploaded and obsolete remote configs are removed. Existing repository history
 is retained. After the upload, use its new commit hash with
 `scripts/set_weight_urls.py`, refresh/upload the manifest, and then commit/push
-the code changes. The QL2 segmentation URL is intentionally unset until then;
-its former URL would download a file with the old metadata and checksum.
+the code changes. The user completed this upload at commit
+`aa51bb453ec2aac74e74e1d84e2d0903625575e7`; all six URLs are now pinned to
+that revision. Upload the refreshed manifest/docs before pushing GitHub code.
 
 - Code: https://github.com/gdslab/alma
 - Model weights: https://huggingface.co/gdslab/alma
@@ -114,8 +115,9 @@ before publishing the code release. The verification directory holds about
 Record the weight commit, completed verification, and manifest change in
 CODE_CHANGE_LOG.txt and the workspace's global audit log before committing.
 The initial upload commit was `5307569f57e16b59eca53484c42304d8ddd8c38e`.
-The terminology update above requires a new upload commit before repinning
-the manifest and verifying fresh public downloads.
+The terminology update is uploaded at
+`aa51bb453ec2aac74e74e1d84e2d0903625575e7`, and the manifest is repinned.
+Fresh public-download verification remains to be completed.
 
 ## 5. Publish the GitHub repository
 

docs/release_audit.md

@@ -62,6 +62,7 @@ The initial weight bundle was uploaded at Hugging Face commit
 `5307569f57e16b59eca53484c42304d8ddd8c38e`. A subsequent terminology cleanup
 uses a public experiment identifier for the QL2 segmentation artifact. This
 changes its metadata and release hash, but preserves every tensor and source
-checkpoint hash. Its download URL awaits a new upload commit; the other five
-files retain their initial links. Fresh public-download verification remains
-pending. Training data and original split manifests are not distributed.
+checkpoint hash. The user uploaded the refreshed bundle at commit
+`aa51bb453ec2aac74e74e1d84e2d0903625575e7`; all six download URLs now point
+to that revision. Fresh public-download verification remains pending.
+Training data and original split manifests are not distributed.

manifest.yaml

@@ -4,7 +4,7 @@ artifacts:
     kind: encoder
     regime: ql1
     filename: ALMA-QL1-Indiana-encoder.pth
-    url: https://huggingface.co/gdslab/alma/resolve/5307569f57e16b59eca53484c42304d8ddd8c38e/ALMA-QL1-Indiana-encoder.pth
+    url: https://huggingface.co/gdslab/alma/resolve/aa51bb453ec2aac74e74e1d84e2d0903625575e7/ALMA-QL1-Indiana-encoder.pth
     sha256: 8af44fb35a409f9e8d05ec21408f4054127833dbeb131101c2f7e21164584c65
     size_bytes: 137629057
     architecture_yaml: configs/ql1_encoder_exp6.yaml
@@ -15,7 +15,7 @@ artifacts:
     kind: encoder
     regime: ql2
     filename: ALMA-QL2-Indiana-encoder.pth
-    url: https://huggingface.co/gdslab/alma/resolve/5307569f57e16b59eca53484c42304d8ddd8c38e/ALMA-QL2-Indiana-encoder.pth
+    url: https://huggingface.co/gdslab/alma/resolve/aa51bb453ec2aac74e74e1d84e2d0903625575e7/ALMA-QL2-Indiana-encoder.pth
     sha256: c6c6860432004bbdd2fd34f36042384f6747410204386bc63b4037c1a2672540
     size_bytes: 265581515
     architecture_yaml: configs/ql2_encoder_exp10_stage2.yaml
@@ -26,7 +26,7 @@ artifacts:
     kind: segmentation
     regime: ql1
     filename: ALMA-QL1-Indiana-seg.pth
-    url: https://huggingface.co/gdslab/alma/resolve/5307569f57e16b59eca53484c42304d8ddd8c38e/ALMA-QL1-Indiana-seg.pth
+    url: https://huggingface.co/gdslab/alma/resolve/aa51bb453ec2aac74e74e1d84e2d0903625575e7/ALMA-QL1-Indiana-seg.pth
     sha256: 1e8c6bc90926e285f15c961c085dce051ea459d8c6f6c295a113468d9b2ed486
     size_bytes: 179926373
     architecture_yaml: configs/ql1_encoder_exp6.yaml
@@ -40,7 +40,7 @@ artifacts:
     kind: segmentation
     regime: ql2
     filename: ALMA-QL2-Indiana-seg.pth
-    url: null
+    url: https://huggingface.co/gdslab/alma/resolve/aa51bb453ec2aac74e74e1d84e2d0903625575e7/ALMA-QL2-Indiana-seg.pth
     sha256: 16c31924136c602a7f138f5ecb50e1883213f21bbf67856d3cfc40563bc381c4
     size_bytes: 339171479
     architecture_yaml: configs/ql2_encoder_exp10_stage2.yaml
@@ -54,7 +54,7 @@ artifacts:
     kind: mae
     regime: ql1
     filename: ALMA-QL1-Indiana-mae.pth
-    url: https://huggingface.co/gdslab/alma/resolve/5307569f57e16b59eca53484c42304d8ddd8c38e/ALMA-QL1-Indiana-mae.pth
+    url: https://huggingface.co/gdslab/alma/resolve/aa51bb453ec2aac74e74e1d84e2d0903625575e7/ALMA-QL1-Indiana-mae.pth
     sha256: bdd5b35de0a7177ad83e295e38acb6fcae6f711733c68d16fd7c2bd3d5633e1c
     size_bytes: 177972088
     architecture_yaml: configs/ql1_encoder_exp6.yaml
@@ -65,7 +65,7 @@ artifacts:
     kind: mae
     regime: ql2
     filename: ALMA-QL2-Indiana-mae.pth
-    url: https://huggingface.co/gdslab/alma/resolve/5307569f57e16b59eca53484c42304d8ddd8c38e/ALMA-QL2-Indiana-mae.pth
+    url: https://huggingface.co/gdslab/alma/resolve/aa51bb453ec2aac74e74e1d84e2d0903625575e7/ALMA-QL2-Indiana-mae.pth
     sha256: 1e4a7e367cac2e5e103e6abbec7893003bdc806405c78b67fd7b8f05b6f0177a
     size_bytes: 337052586
     architecture_yaml: configs/ql2_encoder_exp10_stage2.yaml