Add fold seed prompts

README.md

@@ -13,7 +13,7 @@ pretty_name: Lurcher 10x Microscopy
 
 # Lurcher 10x Microscopy Dataset
 
-This dataset contains 10x microscopy images for the Lurcher classification experiments used by Active Prompt Tuning (APT). Images are organized once by class and case id, while the experiment folds are provided as JSONL manifests.
+This dataset contains 10x microscopy images for the Lurcher classification experiments used by Active Prompt Tuning (APT). Images are organized once by class and case id, while the experiment folds are provided as JSONL manifests plus fold-specific seed prompts.
 
 ## Image layout
 
@@ -32,6 +32,7 @@ folds/microscopy_lurcher/fold-<fold>/
   train.jsonl
   val.jsonl
   test.jsonl
+  seed.json
 ```
 
 Each row has a path relative to the dataset repository root and a normalized class label:
@@ -40,7 +41,7 @@ Each row has a path relative to the dataset repository root and a normalized cla
 {"image_path":"images/microscopy_lurcher/lurcher/5917/example.jpg","class":"lurcher"}
 ```
 
-`train.jsonl` is the active-learning candidate pool. `val.jsonl` is the validation split used by the APT code. `test.jsonl` is the held-out evaluation split.
+`train.jsonl` is the active-learning candidate pool. `val.jsonl` is the validation split used by the APT code. `test.jsonl` is the held-out evaluation split. `seed.json` contains the initial seed prompt examples used before active selection starts.
 
 ## Metadata
 
@@ -64,6 +65,8 @@ hf download USF-CS-Microscopy-Image-Analysis/Lurcher_10x \
   --local-dir .
 ```
 
+The APT repository also includes `setup_lurcher_data.py`, which downloads the images, fold manifests, seed prompts, metadata, and prompt bank into the expected local layout automatically.
+
 ## Intended use
 
 Use these fixed folds when comparing against the APT paper or repository results. Do not regenerate splits for direct comparison to reported fold-level numbers.

metadata/folds.csv

@@ -2,12 +2,16 @@ dataset,fold,split,num_examples,wild,lurcher,file
 microscopy_lurcher,5,train,2238,1588,650,folds/microscopy_lurcher/fold-5/train.jsonl
 microscopy_lurcher,5,val,174,100,74,folds/microscopy_lurcher/fold-5/val.jsonl
 microscopy_lurcher,5,test,146,96,50,folds/microscopy_lurcher/fold-5/test.jsonl
+microscopy_lurcher,5,seed,10,5,5,folds/microscopy_lurcher/fold-5/seed.json
 microscopy_lurcher,6,train,2420,1696,724,folds/microscopy_lurcher/fold-6/train.jsonl
 microscopy_lurcher,6,val,158,108,50,folds/microscopy_lurcher/fold-6/val.jsonl
 microscopy_lurcher,6,test,150,100,50,folds/microscopy_lurcher/fold-6/test.jsonl
+microscopy_lurcher,6,seed,10,5,5,folds/microscopy_lurcher/fold-6/seed.json
 microscopy_lurcher,8,train,2110,1439,671,folds/microscopy_lurcher/fold-8/train.jsonl
 microscopy_lurcher,8,val,166,107,59,folds/microscopy_lurcher/fold-8/val.jsonl
 microscopy_lurcher,8,test,142,100,42,folds/microscopy_lurcher/fold-8/test.jsonl
+microscopy_lurcher,8,seed,10,5,5,folds/microscopy_lurcher/fold-8/seed.json
 microscopy_lurcher,10,train,2184,1605,579,folds/microscopy_lurcher/fold-10/train.jsonl
 microscopy_lurcher,10,val,326,193,133,folds/microscopy_lurcher/fold-10/val.jsonl
 microscopy_lurcher,10,test,214,110,104,folds/microscopy_lurcher/fold-10/test.jsonl
+microscopy_lurcher,10,seed,10,5,5,folds/microscopy_lurcher/fold-10/seed.json