Release anonymous generator checkpoints

README.md

@@ -0,0 +1,61 @@
+# SpineFairBench Generator Checkpoints
+
+This anonymous repository provides generator checkpoint artifacts for reviewer
+inspection. Reviewers do not need to run the generator to verify the benchmark
+endpoints: the fixed counterfactual images used for evaluation are already
+released in `anon-submission7979/spinefairbench-artifacts`.
+
+## Recommended checkpoint
+
+Use `spinefairbench_sd15_lora.safetensors` for reviewer inference inspection.
+This is a clean LoRA-only export derived from the production checkpoint.
+
+## Optional archival checkpoint
+
+`latest.pt` is the full production PyTorch training checkpoint. It is provided
+only for archival/reproducibility inspection. It contains training state such as
+optimizer and discriminator components and is not the recommended inference
+artifact.
+
+## Not included
+
+Raw source radiographs, source masks, runnable training data, credentials,
+private run roots, and provider-client orchestration are not included.
+
+## Download
+
+Recommended LoRA only:
+
+```bash
+hf download anon-submission7979/spinefairbench-generator \
+  spinefairbench_sd15_lora.safetensors \
+  generator_config.yaml \
+  SHA256SUMS.txt \
+  --repo-type model \
+  --local-dir generator_assets
+```
+
+All generator assets, including the optional archival checkpoint:
+
+```bash
+hf download anon-submission7979/spinefairbench-generator \
+  --repo-type model \
+  --local-dir generator_assets
+```
+
+Verify checksums after download:
+
+```bash
+cd generator_assets
+shasum -a 256 -c SHA256SUMS.txt
+```
+
+## Release files
+
+| File | Purpose |
+|---|---|
+| `spinefairbench_sd15_lora.safetensors` | Recommended LoRA-only checkpoint for reviewer inference inspection. |
+| `latest.pt` | Optional archival PyTorch training checkpoint; not recommended for standard inference. |
+| `generator_config.yaml` | Public generator settings used for reviewer inspection. |
+| `SHA256SUMS.txt` | SHA256 checksums for released files. |
+

SHA256SUMS.txt

@@ -0,0 +1,4 @@
+aa94e46cefe9a29b3353bc4344de35663c4e0e49efc4ee6f8fbd0c43d201a106  README.md
+4e6010f49ddd648f81ac21f0b8c9baa3202896f993ad335f4743ce74c4c97725  generator_config.yaml
+21bbfa4ee50ff389c3a49edc9fa39f4e1f7c691fad85c34cbcce595f62b8d56c  spinefairbench_sd15_lora.safetensors
+bc3f71631f07f3d0ca645808b0622f28da2c863b21622cab0ed65e8e44f329ff  latest.pt

generator_config.yaml

@@ -0,0 +1,30 @@
+base_model: stable-diffusion-v1-5
+recommended_checkpoint:
+  file: spinefairbench_sd15_lora.safetensors
+  type: lora_safetensors
+optional_archival_checkpoint:
+  file: latest.pt
+  type: pytorch_training_checkpoint
+  included: true
+  note: "Contains training state such as optimizer/discriminator components; not recommended for standard inference."
+lora:
+  rank: 64
+  alpha: 128
+inference:
+  steps: 50
+  guidance_scale: 5.0
+  strength: 0.15
+  seed: 42
+  mask_blend: 0.7
+prompts:
+  elderly_female: "Lumbar spine X-ray of a 75-year-old female patient"
+  elderly_male: "Lumbar spine X-ray of a 75-year-old male patient"
+  young_female: "Lumbar spine X-ray of a 25-year-old female patient"
+  young_male: "Lumbar spine X-ray of a 25-year-old male patient"
+release_scope:
+  raw_source_radiographs: not_included
+  source_masks: not_included
+  runnable_training_data: not_included
+  optimizer_state_in_lora_safetensors: not_included
+  discriminator_in_lora_safetensors: not_included
+  private_run_roots: not_included

latest.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bc3f71631f07f3d0ca645808b0622f28da2c863b21622cab0ed65e8e44f329ff
+size 3640479384

spinefairbench_sd15_lora.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:21bbfa4ee50ff389c3a49edc9fa39f4e1f7c691fad85c34cbcce595f62b8d56c
+size 51063160