Add README.md

README.md

@@ -21,7 +21,7 @@ identification). This repo hosts a swappable model package consumed via a
 
 | File | What |
 |---|---|
-| `image_encoder.onnx` | BioCLIP v1 **image encoder only** (OpenCLIP ViT-B/16), **fp16**, ONNX (~173 MB). fp16 I/O — cast the preprocessed tensor to float16 before the run. |
+| `image_encoder.onnx` | BioCLIP v1 **image encoder only** (OpenCLIP ViT-B/16), fp16 weights with **fp32 input/output** (~173 MB). Feed a plain float32 `[1,3,224,224]` tensor; boundary Cast nodes handle fp16 internally. |
 | `text_embeddings_v1.safetensors` | Precomputed, L2-normalised text embeddings for the candidate species (fp16, 512-dim). |
 | `species_index.json` | `species_id → row` mapping into the table. |
 | `manifest.json` | Model manifest (architecture, preprocessing, files + sha256). |