Promote manual_0429 as flagship checkpoint

Update the model card, compatibility metadata, and checksum manifest. Remove five deprecated checkpoints from main.

README.md

@@ -20,7 +20,7 @@ tags:
 
 VesselBoost segments small blood vessels in high-resolution human brain MRI. The primary models target time-of-flight magnetic resonance angiography (TOF-MRA). One checkpoint, `t2s_mod_ep1k2_0728`, provides experimental support for T2*-weighted MRI.
 
-These files are PyTorch state dictionaries for use with the VesselBoost inference, test-time adaptation, and boosting workflows. They are not standalone Hugging Face Transformers models or hosted inference endpoints.
+These files are PyTorch state dictionaries for use with the VesselBoost prediction and test-time adaptation workflows. They are not standalone Hugging Face Transformers models or hosted inference endpoints.
 
 **Research use only. Not validated for clinical diagnosis, treatment planning, or other clinical decision-making.**
 
@@ -30,9 +30,9 @@ The checkpoints use the VesselBoost 3D U-Net with one input channel, one output
 
 The corresponding source release is pinned to:
 
-- VesselBoost version: `2.0.2`
-- Git tag: [`v2.0.2`](https://github.com/KMarshallX/VesselBoost/tree/v2.0.2)
-- Git commit: [`1504b00c91777d5e2c271c1cab7f500078f08c69`](https://github.com/KMarshallX/VesselBoost/commit/1504b00c91777d5e2c271c1cab7f500078f08c69)
+- VesselBoost version: `2.0.5`
+- Git tag: [`v2.0.5`](https://github.com/KMarshallX/VesselBoost/tree/v2.0.5)
+- Git commit: [`3f028bbd6784c8fac82ac872a70aa06de2e162ae`](https://github.com/KMarshallX/VesselBoost/commit/3f028bbd6784c8fac82ac872a70aa06de2e162ae)
 
 See `config.json` for the machine-readable inference configuration.
 
@@ -40,19 +40,14 @@ See `config.json` for the machine-readable inference configuration.
 
 All pretrained checkpoints are stored under `weights/`. Their original filenames and serialization formats are preserved from the original release.
 
-| Checkpoint                            | MRI contrast     | Description                                                                                                                             |
-| ------------------------------------- | ---------------- | --------------------------------------------------------------------------------------------------------------------------------------- |
-| [`BM_VB2_aug_all_ep2k_bat_10_0903`](weights/BM_VB2_aug_all_ep2k_bat_10_0903)   | TOF-MRA          | Primary TOF-MRA checkpoint referenced by the VesselBoost documentation and tests; trained with the combined augmentation configuration. |
-| [`VB2_aug_intensity_ep2k_bat10_0903`](weights/VB2_aug_intensity_ep2k_bat10_0903) | TOF-MRA          | Augmentation ablation using the intensity augmentation configuration.                                                                   |
-| [`VB2_aug_off_ep2k_bat10_0903`](weights/VB2_aug_off_ep2k_bat10_0903)       | TOF-MRA          | Augmentation ablation with augmentation disabled.                                                                                       |
-| [`VB2_aug_random_ep2k_bat10_0903`](weights/VB2_aug_random_ep2k_bat10_0903)    | TOF-MRA          | Augmentation ablation using the random augmentation configuration.                                                                      |
-| [`VB2_aug_spatial_ep2k_bat10_0903`](weights/VB2_aug_spatial_ep2k_bat10_0903)   | TOF-MRA          | Augmentation ablation using the spatial augmentation configuration.                                                                     |
-| [`manual_0429`](weights/manual_0429)                       | TOF-MRA          | Legacy checkpoint associated with the manual-label training run and used in VesselBoost v2.0.2 examples.                                |
-| [`omelette1_0429`](weights/omelette1_0429)                    | TOF-MRA          | Legacy TOF-MRA checkpoint identified as Omelette variant 1.                                                                             |
-| [`omelette2_0429`](weights/omelette2_0429)                    | TOF-MRA          | Legacy TOF-MRA checkpoint identified as Omelette variant 2.                                                                            |
-| [`t2s_mod_ep1k2_0728`](weights/t2s_mod_ep1k2_0728)                | T2*-weighted MRI | Experimental T2*-weighted vessel-segmentation checkpoint. It has not received the same validation as the primary TOF-MRA model.         |
+| Checkpoint                                         | MRI contrast     | Description                                                                                                                     |
+| -------------------------------------------------- | ---------------- | ------------------------------------------------------------------------------------------------------------------------------- |
+| [`manual_0429`](weights/manual_0429)                | TOF-MRA          | Flagship TOF-MRA checkpoint referenced by the current VesselBoost documentation and automated tests.                            |
+| [`omelette1_0429`](weights/omelette1_0429)          | TOF-MRA          | Legacy TOF-MRA checkpoint identified as Omelette variant 1.                                                                     |
+| [`omelette2_0429`](weights/omelette2_0429)          | TOF-MRA          | Legacy TOF-MRA checkpoint identified as Omelette variant 2.                                                                     |
+| [`t2s_mod_ep1k2_0728`](weights/t2s_mod_ep1k2_0728) | T2*-weighted MRI | Experimental T2*-weighted vessel-segmentation checkpoint. It has not received the same validation as the flagship TOF-MRA model. |
 
-The augmentation-specific checkpoints are included to preserve the original model set and support comparison or reproduction of augmentation experiments. For the standard TOF-MRA prediction workflow, use `manual_0429` or `BM_VB2_aug_all_ep2k_bat_10_0903` unless reproducing a specific legacy experiment.
+For standard TOF-MRA prediction and test-time adaptation workflows, use the flagship `manual_0429` checkpoint.
 
 ## Downloading checkpoints
 
@@ -66,11 +61,11 @@ Download the primary TOF-MRA checkpoint:
 
 ```bash
 hf download BrainVascuLab/VesselBoost \
-  weights/BM_VB2_aug_all_ep2k_bat_10_0903 \
+  weights/manual_0429 \
   --local-dir saved_models
 ```
 
-The downloaded checkpoint will be available at `saved_models/weights/BM_VB2_aug_all_ep2k_bat_10_0903`.
+The downloaded checkpoint will be available at `saved_models/weights/manual_0429`.
 
 Download every pretrained checkpoint and the checksum manifest:
 
@@ -84,7 +79,7 @@ For reproducible automated workflows, pass `--revision` with a specific Hugging
 
 ## Preprocessing and inference
 
-VesselBoost v2.0.2 performs the following inference operations:
+VesselBoost v2.0.5 performs the following inference operations:
 
 1. Load a single-channel NIfTI MRI volume.
 2. Resize each spatial dimension to at least 64 voxels and to a multiple of 64, using nearest-neighbor interpolation.
@@ -106,7 +101,7 @@ cd saved_models/weights
 sha256sum --check MANIFEST.sha256
 ```
 
-All nine checkpoints should report `OK`.
+All four checkpoints should report `OK`.
 
 Load the checkpoints with the pinned VesselBoost source and map tensors to the intended device. When supported by the installed PyTorch version, use `weights_only=True` when loading these state dictionaries.
 

config.json

@@ -48,9 +48,9 @@
   },
   "checkpoint_format": "PyTorch state_dict ZIP serialization",
   "compatible_vesselboost": {
-    "version": "2.0.2",
-    "git_tag": "v2.0.2",
-    "git_commit": "1504b00c91777d5e2c271c1cab7f500078f08c69",
-    "source_url": "https://github.com/KMarshallX/VesselBoost/tree/v2.0.2"
+    "version": "2.0.5",
+    "git_tag": "v2.0.5",
+    "git_commit": "3f028bbd6784c8fac82ac872a70aa06de2e162ae",
+    "source_url": "https://github.com/KMarshallX/VesselBoost/tree/v2.0.5"
   }
 }

weights/BM_VB2_aug_all_ep2k_bat_10_0903

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:ff51c8e6c79947f13bf7c24ac2ae2242eb4b23b7739e62dc3f8344944bde28f1
-size 26397502

weights/MANIFEST.sha256

@@ -1,9 +1,3 @@
-ff51c8e6c79947f13bf7c24ac2ae2242eb4b23b7739e62dc3f8344944bde28f1  BM_VB2_aug_all_ep2k_bat_10_0903
-73b3cabf38dcb4266507f375c67f7f6e1fa3e57747ff9ae5b9a2285175462ada  VB2_aug_intensity_ep2k_bat10_0903
-056849b426a669a7536d57b270803fa9a2dab69d9f8abbfcb865d028864db961  VB2_aug_off_ep2k_bat10_0903
-09b28d972172939f554f3884de9570f9584720b5c8fd225d620e812e48507461  VB2_aug_random_ep2k_bat10_0903
-a5bb099eaa70b1c61f948c5475002083b7637c118ed5525124dece20100de4d4  VB2_aug_spatial_ep2k_bat10_0903
-fb318efe161b0036f2af2585fb282dd92ab45c6f0258422154340f05a2b3ef80  manual_0429
-980cbbd4cdd85fc731cc5bc621b402ebcc1f0abba028ced4b6035bf92ebaa212  omelette1_0429
-d18245068d8a196d52d9cbaa841a0eff314c5673e5bef85af60bf482d28d1633  omelette2_0429
-794d8e333ae0b26578824debc375950a0637868deca3f16656dc81e0076951ad  t2s_mod_ep1k2_0728
+version https://git-lfs.github.com/spec/v1
+oid sha256:74b0316c2fe09e75f8f41c3d0972723215a5029537924cd10d404e17960e1dc7
+size 325

weights/VB2_aug_intensity_ep2k_bat10_0903

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:73b3cabf38dcb4266507f375c67f7f6e1fa3e57747ff9ae5b9a2285175462ada
-size 26398202

weights/VB2_aug_off_ep2k_bat10_0903

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:056849b426a669a7536d57b270803fa9a2dab69d9f8abbfcb865d028864db961
-size 26397362

weights/VB2_aug_random_ep2k_bat10_0903

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:09b28d972172939f554f3884de9570f9584720b5c8fd225d620e812e48507461
-size 26397782

weights/VB2_aug_spatial_ep2k_bat10_0903

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:a5bb099eaa70b1c61f948c5475002083b7637c118ed5525124dece20100de4d4
-size 26397922