Added threshold slider for siglip models

app.py

@@ -115,7 +115,9 @@ CLASSIFIER_MAP = {
 }
 
 
-def run_dfine_classify(image, refs_path, dfine_threshold, dfine_model_choice, min_display_conf=0.703, gap_threshold=0.005, classifier_choice="Jina-CLIP-v2 (few-shot)"):
+def run_dfine_classify(image, refs_path, dfine_threshold, dfine_model_choice,
+                       min_display_conf, gap_threshold, siglip_threshold,
+                       classifier_choice="Jina-CLIP-v2 (few-shot)"):
     """Tab 2: D-FINE first, then classify crops.
     Returns (group_crop_gallery, known_crop_gallery, status_message).
     """
@@ -129,16 +131,27 @@ def run_dfine_classify(image, refs_path, dfine_threshold, dfine_model_choice, mi
 
     dfine_model = "large" if dfine_model_choice.strip().lower() == "large" else "medium"
     classifier = CLASSIFIER_MAP.get(classifier_choice, "jina")
+
+    # SigLIP models: use their own threshold, no gap check
+    if classifier in ("siglip", "siglip2_onnx"):
+        conf_thresh = float(siglip_threshold)
+        gap_thresh = 0.0
+        display_conf = float(siglip_threshold)
+    else:
+        conf_thresh = 0.5
+        gap_thresh = float(gap_threshold)
+        display_conf = float(min_display_conf)
+
     group_crops, known_crops, status = run_single_image(
         image,
         refs_dir=refs,
         dfine_model=dfine_model,
         det_threshold=float(dfine_threshold),
-        conf_threshold=0.5,
-        gap_threshold=float(gap_threshold),
+        conf_threshold=conf_thresh,
+        gap_threshold=gap_thresh,
         min_side=24,
         crop_dedup_iou=0.4,
-        min_display_conf=float(min_display_conf),
+        min_display_conf=display_conf,
         classifier=classifier,
     )
 
@@ -298,12 +311,13 @@ with gr.Blocks(title="Small Object Detection") as app:
 
                 with gr.Column(scale=1):
 
+                    # --- Jina thresholds (visible when Jina selected) ---
                     threshold_slider = gr.Slider(
                         minimum=0.0,
                         maximum=1.0,
                         value=0.703,
                         step=0.005,
-                        label="Threshold (min display confidence)",
+                        label="Jina: min display confidence",
                     )
 
                     gap_slider = gr.Slider(
@@ -311,7 +325,17 @@ with gr.Blocks(title="Small Object Detection") as app:
                         maximum=0.02,
                         value=0.005,
                         step=0.001,
-                        label="Gap: how much the top guess must beat the runner-up (higher = stricter, fewer accepted)",
+                        label="Jina: gap (top class must beat runner-up by this much)",
+                    )
+
+                    # --- SigLIP threshold (visible when SigLIP selected) ---
+                    siglip_threshold_slider = gr.Slider(
+                        minimum=0.0,
+                        maximum=1.0,
+                        value=0.05,
+                        step=0.01,
+                        label="SigLIP: min confidence threshold",
+                        visible=False,
                     )
 
                     out_gallery_dfine = gr.Gallery(
@@ -334,9 +358,25 @@ with gr.Blocks(title="Small Object Detection") as app:
                         interactive=False,
                     )
 
+            # Show/hide threshold sliders based on classifier choice
+            def update_threshold_visibility(choice):
+                is_jina = (choice == "Jina-CLIP-v2 (few-shot)")
+                return (
+                    gr.update(visible=is_jina),      # threshold_slider
+                    gr.update(visible=is_jina),      # gap_slider
+                    gr.update(visible=not is_jina),  # siglip_threshold_slider
+                )
+
+            classifier_radio.change(
+                fn=update_threshold_visibility,
+                inputs=[classifier_radio],
+                outputs=[threshold_slider, gap_slider, siglip_threshold_slider],
+            )
+
             btn_dfine.click(
                 fn=run_dfine_classify,
-                inputs=[inp_dfine, refs_path, dfine_threshold_slider, dfine_model_radio, threshold_slider, gap_slider, classifier_radio],
+                inputs=[inp_dfine, refs_path, dfine_threshold_slider, dfine_model_radio,
+                        threshold_slider, gap_slider, siglip_threshold_slider, classifier_radio],
                 outputs=[out_gallery_dfine, out_gallery_known, out_status_dfine],
                 concurrency_limit=1,
             )

siglip2_onnx_zeroshot.py

@@ -163,20 +163,12 @@ class SigLIP2ONNXClassifier:
         conf = float(probs[best_idx])
         gap = float(probs[best_idx] - probs[second_idx])
 
-        conf_ok = conf >= conf_threshold
-        gap_ok = gap >= gap_threshold
-
-        if conf_ok and gap_ok:
+        if conf >= conf_threshold:
             prediction = self.labels[best_idx]
             status = "accepted"
         else:
             prediction = "unknown"
-            reasons = []
-            if not conf_ok:
-                reasons.append(f"conf {conf:.4f} < {conf_threshold}")
-            if not gap_ok:
-                reasons.append(f"gap {gap:.4f} < {gap_threshold}")
-            status = "rejected: " + ", ".join(reasons)
+            status = f"rejected: conf {conf:.4f} < {conf_threshold}"
 
         return {
             "prediction": prediction,

siglip_zeroshot.py

@@ -62,20 +62,12 @@ class SigLIPClassifier:
         conf = float(probs[best_idx])
         gap = float(probs[best_idx] - probs[second_idx])
 
-        conf_ok = conf >= conf_threshold
-        gap_ok = gap >= gap_threshold
-
-        if conf_ok and gap_ok:
+        if conf >= conf_threshold:
             prediction = self.labels[best_idx]
             status = "accepted"
         else:
             prediction = "unknown"
-            reasons = []
-            if not conf_ok:
-                reasons.append(f"conf {conf:.4f} < {conf_threshold}")
-            if not gap_ok:
-                reasons.append(f"gap {gap:.4f} < {gap_threshold}")
-            status = "rejected: " + ", ".join(reasons)
+            status = f"rejected: conf {conf:.4f} < {conf_threshold}"
 
         return {
             "prediction": prediction,