Colorizer round5: completed checkpoint, model card and evaluation

.gitattributes

@@ -54,3 +54,5 @@ experiments/round5-20260927/structure_distilled_1500.jpg filter=lfs diff=lfs mer
 experiments/round5-20260927/structure_distilled_2000.jpg filter=lfs diff=lfs merge=lfs -text
 experiments/round5-20260927/structure_distilled_2500.jpg filter=lfs diff=lfs merge=lfs -text
 experiments/round5-20260927/structure_distilled_3000.jpg filter=lfs diff=lfs merge=lfs -text
+experiments/round5-20260927/selected_final.jpg filter=lfs diff=lfs merge=lfs -text
+experiments/round5-20260927/v2_final.jpg filter=lfs diff=lfs merge=lfs -text

experiments/round5-20260927/SHA256SUMS.json

@@ -1,6 +1,6 @@
 {
   "baseline_validation.json": "f3609da4c939f1c8b2fd65504c80dd70045afa99e3a1ab9ab05c56acbbba234b",
-  "best/README.md": "fb835cbd31ee2df2d88d83f69354f0d82fc6213f24efbc6b3e92e1238a191493",
+  "best/README.md": "84654bec1c96483b0402761cc3c3556c51a962da0a23dac323c97fe060854aa4",
   "best/config.json": "90299ee25ee3c3e645b362ddbd5a859b9e739c2a0a1c8a5e9ea79df171288bff",
   "best/model.safetensors": "0e4c417375684a044860f8af3ac3a2fb44e1a5729254ca33abda758ea71aea6e",
   "ddcolor_artistic.jpg": "5e12d97fca707877c6e8f3da30dd9f7040d7464b03c506adf72e97b924804d49",
@@ -8,6 +8,8 @@
   "manifest.json": "24bc081bf48f1109d9dfa217ae0d8593c0c06af6785a5cbfc5dc5d9e25461f5c",
   "progress.json": "c83583271cbc65fca1a99dbd3e6d36f57b79f5e518570d69088baab6a3066e38",
   "provenance.json": "8c49012053e9cc6a496f44e0356995c7e93e2a7b388dec9e5c83aa3be58c16a0",
+  "selected_final.jpg": "601853e67398435b8deb53b1db80557c6b16a5633ff70292ab920b07a1ca2d2e",
+  "selected_test.json": "62ecbb7ccaa5f4cd77e73a7b3fc57d80da1855eb7f54e0f936db38db5f542f69",
   "selection.json": "a714182d0a1985452e9ad7bb7a44729325382bde322edcc34b55f6680e2d9004",
   "source/PROTOCOL.md": "cb34624b5f431b859c31f583f96f5ab531f0248dea2f33271947523c5ec4aa07",
   "source/data.py": "eb43b886fe45435841439c36646985f914386a48faa712fb6806accac2ad48e0",
@@ -40,7 +42,7 @@
   "source/vendor_ddcolor/ddcolor/__init__.py": "2606a3377189bda8beb1a2075c7bd789961163bca3e1e53b01bbe325dc561c30",
   "source/vendor_ddcolor/ddcolor/model.py": "e8e30115aa65a9558db33641c999001d1735c444402890898b47b2cc52b37c25",
   "source/vendor_ddcolor/ddcolor/pipeline.py": "61fb3a7642309d1dc5069fb8089dcdff13ab5680c78269013b579bfa654ce005",
-  "status.json": "756d10f0138e6fd95b56a37010d0ea857f12f9d53ca695523ea11f8b75e821b6",
+  "status.json": "102be73380154d09ca97ec02332099a88929fc6edbc75a3a78ba32297de5efb5",
   "structure_control/latest/README.md": "fb835cbd31ee2df2d88d83f69354f0d82fc6213f24efbc6b3e92e1238a191493",
   "structure_control/latest/config.json": "90299ee25ee3c3e645b362ddbd5a859b9e739c2a0a1c8a5e9ea79df171288bff",
   "structure_control/latest/model.safetensors": "37bb9da3946ef195030f757f139d2d2d646c2c0299f76abdb3e668fd1a90cc30",
@@ -75,5 +77,7 @@
   "structure_distilled_500_validation.json": "31d79072c1984e2aaf251122aabbd8eb0f0eb92ba100f0dc36f6f5ffb85e0384",
   "teacher_test.json": "9ee4ce1c1a1f280b6915778399370e96e1cf9c6e5aae34cfcff234aef3e8b076",
   "teacher_validation.json": "4fc2d0b20f09a36385fa74d969ccab7a9f31221199103995fa2c9e6cd14c3385",
-  "v2.jpg": "af5a356407b4abbbe0bea5c5d3f62f007a6f8061bfc15902c44c28dfe3efec7f"
+  "v2.jpg": "af5a356407b4abbbe0bea5c5d3f62f007a6f8061bfc15902c44c28dfe3efec7f",
+  "v2_final.jpg": "ddd7a6d563c30a390d7a48d5617f61bc8a0689192718719ee47544e63afd6bc4",
+  "v2_test.json": "62ecbb7ccaa5f4cd77e73a7b3fc57d80da1855eb7f54e0f936db38db5f542f69"
 }

experiments/round5-20260927/best/README.md

@@ -1,16 +1,7 @@
----
-license: apache-2.0
-pipeline_tag: image-to-image
-tags:
-- classification
-- colorization
-- image-to-image
-- model_hub_mixin
-- pytorch_model_hub_mixin
-- unet
----
+# Mini U-Net round5 candidate
 
-This model has been pushed to the Hub using the [PytorchModelHubMixin](https://huggingface.co/docs/huggingface_hub/package_reference/mixins#huggingface_hub.PyTorchModelHubMixin) integration:
-- Code: [More Information Needed]
-- Paper: [More Information Needed]
-- Docs: [More Information Needed]
+Experimental checkpoint; visual review required before deployment.
+
+Selection: {"arm": "v2_baseline", "step": 0, "accepted": false}
+
+See ../source/PROTOCOL.md, ../provenance.json, ../history.json and ../selected_test.json. Trained on Imagenette and COCO; DDColor teacher uses data with possible evaluation overlap. Predictions are plausible colors, not recovered historical colors. Architecture and decoder remain compatible with the existing Space.

experiments/round5-20260927/status.json

@@ -1,3 +1,4 @@
 {
-  "status": "initializing"
+  "status": "completed",
+  "elapsed_seconds": 4330.74058792
 }