updates

app.py

@@ -114,6 +114,8 @@ col1, col2 = st.columns([5,1])
 
 with col1:
 
+
+
   current_image=ds[st.session_state["img"]]['image'].resize(size=(512,512))
   draw = ImageDraw.Draw(current_image)
 
@@ -145,7 +147,7 @@ with col1:
   with scol2: 
     st.button('Find similar parts', on_click=find)
 
-  st.write("Currently viewing frame "+str(random_i[st.session_state["img"]%20])+" in batch "+str(st.session_state["img"]//20))
+  #st.write("Currently viewing frame "+str(random_i[st.session_state["img"]%20])+" in batch "+str(st.session_state["img"]//20))
 
   #st.write(st.session_state["img"])
   #st.write(st.session_state["point"])
@@ -156,14 +158,10 @@ with col2:
   for i in range(4):
     
     if i==0:
-      st.write("current selection in batch "+str(st.session_state["sideix"][i]))#//(gridsize**2)//20))
-    if i==1:
-      st.write("Best match found in batch "+str(st.session_state["sideix"][i]))#//(gridsize**2)//20))
-    if i==2:
-      st.write("Second best match found in batch "+str(st.session_state["sideix"][i]))#//(gridsize**2)//20))
-    if i==3:
-      st.write("Third best match found in batch "+str(st.session_state["sideix"][i]))#//(gridsize**2)//20))
-
+      st.write("Target in batch "+str(st.session_state["sideix"][i]))#//(gridsize**2)//20))
+    else:
+      st.write("Match #"+str(i)+" in batch "+str(st.session_state["sideix"][i]))#//(gridsize**2)//20))
+   
     st.image(st.session_state["sideimg"][i].resize((128,128)))
     
     #st.write(st.session_state["sideix"][i])
@@ -171,9 +169,9 @@ with col2:
 "The batch indexing machine shakes parts while recording a video."
 "The machine processed 20 batches of random parts, with each batch running for 30 seconds."
 
-#"HOW TO DUMPSTER DIVE DIGITALLY:"
-#"Click in the image to set a target"
-#"Click “Find similar parts” to find the best matches in other batches"
+"INSCTRUCTIONS:"
+"Click in the image to set target part"
+"Click “Find similar parts” to find the best matches in other batches"
 
 "The model is trained completely unsupervised using a CNN with a custom contrastive loss. Open source code to be released soon. "