updated readme

README.md

@@ -82,34 +82,114 @@ configs:
     path: "data/configs/intl_train_US__ood_IN/ood_test-*.parquet"
 ---
 
-# Tree Distribution Shift — Satellite Tree Detection (COCO export)
+# Tree Distribution Shift — Satellite Tree Detection (COCO + HF Datasets)
 
-This dataset is organized as **configs** (distribution shift settings).
-Each config provides 3 splits:
+This dataset is organized as **configs** (distribution shift benchmarks). Each config provides **three splits**:
 
 - `train`
-- `id_test` (same region as train, held-out)
-- `ood_test` (different region)
+- `id_test` (same distribution as train, held-out)
+- `ood_test` (different distribution)
 
-## Super simple COCO UX
+## What is a config?
 
-Export any config into standard COCO folder structure:
+A **config** fully defines a benchmark setting (e.g., country shift, state shift, biome shift).  
+When you select a config, you automatically get:
 
+- `train` split
+- `id_test` split
+- `ood_test` split
+
+No filtering or custom split logic is required.
+
+---
+
+# Export to COCO (recommended for training)
+
+Use this when you want a **standard COCO folder structure** that works with most CV frameworks (Detectron2, MMDetection, torchvision detection, etc.).
+
+## Step 1 — Clone the dataset repository (to get the export tool)
+```bash
+git clone https://huggingface.co/datasets/aadityabuilds/tree-distribution-shift
+cd tree-distribution-shift
+```
+
+## Step 2 — Install dependencies
 ```bash
-python tools/export_coco.py --repo aadityabuilds/tree-distribution-shift --config biome_India_Karnataka_train_DRY__ood_WET --out ./coco_out
+pip install datasets huggingface_hub orjson
 ```
 
-This writes:
+## Step 3 — Pick a config
+
+Go to the dataset page and choose a config from the Configs section:
 
-- `./coco_out/biome_India_Karnataka_train_DRY__ood_WET/train/images/*.tiff`
-- `./coco_out/biome_India_Karnataka_train_DRY__ood_WET/train/annotations/instances_train.json`
-- `./coco_out/biome_India_Karnataka_train_DRY__ood_WET/id_test/...`
-- `./coco_out/biome_India_Karnataka_train_DRY__ood_WET/ood_test/...`
+https://huggingface.co/datasets/aadityabuilds/tree-distribution-shift
+
+Example config:
+```
+biome_India_Karnataka_train_DRY__ood_WET
+```
 
-## Programmatic load (optional)
+## Step 4 — Export to COCO (one command)
+```bash
+python tools/export_coco.py \
+  --repo aadityabuilds/tree-distribution-shift \
+  --config biome_India_Karnataka_train_DRY__ood_WET \
+  --out ./coco_out
+```
 
+## Output structure
+```
+coco_out/
+└── biome_India_Karnataka_train_DRY__ood_WET/
+    ├── train/
+    │   ├── images/
+    │   │   └── *.tiff
+    │   └── annotations/
+    │       └── instances_train.json
+    ├── id_test/
+    │   ├── images/
+    │   └── annotations/
+    │       └── instances_id_test.json
+    └── ood_test/
+        ├── images/
+        └── annotations/
+            └── instances_ood_test.json
+```
+
+---
+
+# Load with Hugging Face Datasets (recommended for analysis / prototyping)
+
+Use this when you want programmatic access (e.g., notebooks, statistics, custom pipelines) without writing files to disk.
+
+## Install
+```bash
+pip install datasets
+```
+
+## Load a config (gets train / id_test / ood_test)
 ```python
 from datasets import load_dataset
-ds = load_dataset("aadityabuilds/tree-distribution-shift", "biome_India_Karnataka_train_DRY__ood_WET")
+
+ds = load_dataset(
+    "aadityabuilds/tree-distribution-shift",
+    "biome_India_Karnataka_train_DRY__ood_WET"
+)
+
+train   = ds["train"]
+id_test = ds["id_test"]
+ood_test= ds["ood_test"]
+
 print(ds)
+print(len(train), len(id_test), len(ood_test))
+print(train[0].keys())
 ```
+
+## What each example contains
+
+Each row includes:
+
+- `image_bytes` (raw image bytes)
+- `coco_annotations` (COCO annotations for that image)
+- `coco_categories` (COCO categories)
+- metadata fields (country, state, zone, biome, etc.)