Upload 3 files

fixed token issue

README.md

@@ -1,15 +1,16 @@
 ---
-title: Spotpredator
-emoji: 💻
+title: SpotPredator Predator Detection
+emoji: 🦅
 colorFrom: green
-colorTo: green
+colorTo: blue
 sdk: gradio
-sdk_version: 6.19.0
-python_version: '3.13'
 app_file: app.py
+python_version: "3.11"
 pinned: false
-license: agpl-3.0
-short_description: Just a location to test my vision models on predators
 ---
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+# SpotPredator — predator_v2_fp16
+
+Browser demo for an off-grid farm predator detector (coyote / fox / raptor).
+Runs the TFLite model directly via LiteRT — no ultralytics required.
+Drag in a field image to see detections.

app.py

@@ -10,47 +10,72 @@ Create the Space:
 Or from CLI after `pip install huggingface_hub`:
   huggingface-cli repo create predator-demo --type space --space_sdk gradio
 """
+import os
 import numpy as np
 from PIL import Image, ImageDraw
 import gradio as gr
 from huggingface_hub import hf_hub_download, list_repo_files
 
 MODEL_REPO = "JZVince/predator_v2_fp16"
-CLASS_NAMES = ["coyote", "fox", "raptor"]   # confirm this matches training order
+# confirm this matches training order
+CLASS_NAMES = ["coyote", "fox", "raptor"]
+
+# Optional: set HF_TOKEN as a Space secret for higher rate limits / faster pulls.
+HF_TOKEN = os.environ.get("spcae")
 
 # ---- fetch the .tflite from the model repo (auto-detect the filename) --------
-tflite_name = next(f for f in list_repo_files(MODEL_REPO) if f.endswith(".tflite"))
-MODEL_PATH = hf_hub_download(MODEL_REPO, tflite_name)
+tflite_name = next(f for f in list_repo_files(MODEL_REPO, token=HF_TOKEN)
+                   if f.endswith(".tflite"))
+MODEL_PATH = hf_hub_download(MODEL_REPO, tflite_name, token=HF_TOKEN)
 
+# LiteRT is the successor to tflite-runtime; fall back through older runtimes.
 try:
-    from tflite_runtime.interpreter import Interpreter
+    from ai_edge_litert.interpreter import Interpreter
 except ImportError:
-    from tensorflow.lite import Interpreter
-interp = Interpreter(model_path=MODEL_PATH); interp.allocate_tensors()
-INP = interp.get_input_details()[0]; OUT = interp.get_output_details()[0]
-IN_SIZE = int(INP["shape"][1]); NC = len(CLASS_NAMES)
+    try:
+        from tflite_runtime.interpreter import Interpreter
+    except ImportError:
+        from tensorflow.lite import Interpreter
+interp = Interpreter(model_path=MODEL_PATH)
+interp.allocate_tensors()
+INP = interp.get_input_details()[0]
+OUT = interp.get_output_details()[0]
+IN_SIZE = int(INP["shape"][1])
+NC = len(CLASS_NAMES)
 
 # ---- inference core (same verified decode as detect_tflite.py) ---------------
+
+
 def letterbox(img, new, color=(114, 114, 114)):
-    w, h = img.size; gain = min(new/w, new/h)
+    w, h = img.size
+    gain = min(new/w, new/h)
     nw, nh = round(w*gain), round(h*gain)
     canvas = Image.new("RGB", (new, new), color)
     px, py = (new-nw)//2, (new-nh)//2
     canvas.paste(img.resize((nw, nh), Image.BILINEAR), (px, py))
     return canvas, gain, (px, py)
 
+
 def nms(boxes, scores, iou_thr=0.45):
-    x1, y1, x2, y2 = boxes.T; areas = (x2-x1)*(y2-y1)
-    order = scores.argsort()[::-1]; keep = []
+    x1, y1, x2, y2 = boxes.T
+    areas = (x2-x1)*(y2-y1)
+    order = scores.argsort()[::-1]
+    keep = []
     while order.size:
-        i = order[0]; keep.append(i)
-        xx1 = np.maximum(x1[i], x1[order[1:]]); yy1 = np.maximum(y1[i], y1[order[1:]])
-        xx2 = np.minimum(x2[i], x2[order[1:]]); yy2 = np.minimum(y2[i], y2[order[1:]])
-        w = np.maximum(0, xx2-xx1); h = np.maximum(0, yy2-yy1); inter = w*h
+        i = order[0]
+        keep.append(i)
+        xx1 = np.maximum(x1[i], x1[order[1:]])
+        yy1 = np.maximum(y1[i], y1[order[1:]])
+        xx2 = np.minimum(x2[i], x2[order[1:]])
+        yy2 = np.minimum(y2[i], y2[order[1:]])
+        w = np.maximum(0, xx2-xx1)
+        h = np.maximum(0, yy2-yy1)
+        inter = w*h
         iou = inter/(areas[i]+areas[order[1:]]-inter+1e-9)
         order = order[1:][iou <= iou_thr]
     return keep
 
+
 def detect(image, conf_thr=0.25, iou_thr=0.45):
     if image is None:
         return None, "Upload an image."
@@ -58,38 +83,49 @@ def detect(image, conf_thr=0.25, iou_thr=0.45):
     lb, gain, pad = letterbox(img, IN_SIZE)
     x = np.asarray(lb, np.float32)[None] / 255.0
     if INP["dtype"] == np.uint8:
-        s, z = INP["quantization"]; x = (x/s + z).astype(np.uint8)
-    interp.set_tensor(INP["index"], x); interp.invoke()
+        s, z = INP["quantization"]
+        x = (x/s + z).astype(np.uint8)
+    interp.set_tensor(INP["index"], x)
+    interp.invoke()
     y = interp.get_tensor(OUT["index"])
     if OUT["dtype"] == np.uint8:
-        s, z = OUT["quantization"]; y = (y.astype(np.float32)-z)*s
+        s, z = OUT["quantization"]
+        y = (y.astype(np.float32)-z)*s
     o = y[0]
     if o.shape[0] == 4 + NC:
         o = o.T
     xywh = o[:, :4].astype(np.float32).copy()
-    cls = o[:, 4:4+NC]; cid = cls.argmax(1); conf = cls.max(1)
-    m = conf >= conf_thr; xywh, conf, cid = xywh[m], conf[m], cid[m]
-    draw = ImageDraw.Draw(img); lines = []
+    cls = o[:, 4:4+NC]
+    cid = cls.argmax(1)
+    conf = cls.max(1)
+    m = conf >= conf_thr
+    xywh, conf, cid = xywh[m], conf[m], cid[m]
+    draw = ImageDraw.Draw(img)
+    lines = []
     colors = [(220, 70, 60), (240, 160, 40), (60, 130, 200)]
     if len(xywh):
         if xywh.max() <= 1.5:
             xywh *= IN_SIZE
         cx, cy, w, h = xywh.T
-        x1 = (cx-w/2-pad[0])/gain; y1 = (cy-h/2-pad[1])/gain
-        x2 = (cx+w/2-pad[0])/gain; y2 = (cy+h/2-pad[1])/gain
+        x1 = (cx-w/2-pad[0])/gain
+        y1 = (cy-h/2-pad[1])/gain
+        x2 = (cx+w/2-pad[0])/gain
+        y2 = (cy+h/2-pad[1])/gain
         xyxy = np.stack([x1, y1, x2, y2], 1)
         xyxy[:, [0, 2]] = xyxy[:, [0, 2]].clip(0, img.width)
         xyxy[:, [1, 3]] = xyxy[:, [1, 3]].clip(0, img.height)
         for c in np.unique(cid):
             idx = np.where(cid == c)[0]
             for k in nms(xyxy[idx], conf[idx], iou_thr):
-                b = xyxy[idx][k]; col = colors[int(c) % 3]
+                b = xyxy[idx][k]
+                col = colors[int(c) % 3]
                 draw.rectangle(list(b), outline=col, width=3)
                 draw.text((b[0]+3, max(0, b[1]-12)),
                           f"{CLASS_NAMES[int(c)]} {conf[idx][k]:.2f}", fill=col)
                 lines.append(f"{CLASS_NAMES[int(c)]}: {conf[idx][k]:.2f}")
     return img, ("\n".join(lines) if lines else "No predators detected.")
 
+
 demo = gr.Interface(
     fn=detect,
     inputs=[gr.Image(type="pil", label="Field image"),

requirements.txt

@@ -1,5 +1,5 @@
 gradio
 numpy
 pillow
-tensorflow-cpu
-huggingface_hub
+ai-edge-litert
+huggingface_hub