Update app.py

app.py

@@ -52,6 +52,7 @@ pipe.vae = AutoencoderTiny.from_pretrained("madebyollin/taesdxl", torch_dtype=to
 pipe.scheduler = EulerDiscreteScheduler.from_config(pipe.scheduler.config, timestep_spacing="trailing")
 
 pipe.to(device=DEVICE)
+pipe.set_progress_bar_config(disable=True)
 
 @spaces.GPU
 def compile_em():
@@ -111,11 +112,11 @@ compile_em()
 @spaces.GPU
 def generate(prompt, in_embs=None,):
   if prompt != '':
-    print(prompt)
+    # # print(prompt)
     in_embs = in_embs / in_embs.abs().max() * .15 if in_embs != None else None
     in_embs = .9 * in_embs.to('cuda') + .5 * autoencoder.embed(prompt).to('cuda') if in_embs != None else autoencoder.embed(prompt).to('cuda')
   else:
-    print('From embeds.')
+    # print('From embeds.')
   in_embs = in_embs / in_embs.abs().max() * .15
   text = autoencoder.generate_from_latent(in_embs.to('cuda').to(dtype=torch.bfloat16), temperature=.3, top_p=.99, min_new_tokens=5)
   in_embs = autoencoder.embed(text).to('cuda')
@@ -175,7 +176,7 @@ def get_coeff(embs_local, ys):
     # this ends up adding a rating but losing an embedding, it seems.
     # let's take off a rating if so to continue without indexing errors.
     if len(ys) > len(embs_local):
-        print('ys are longer than embs; popping latest rating')
+        # print('ys are longer than embs; popping latest rating')
         ys.pop(-1)
     
     # also add the latest 0 and the latest 1
@@ -194,7 +195,7 @@ def get_coeff(embs_local, ys):
     feature_embs = np.array(torch.cat([embs_local[i].to('cpu') for i in indices]).to('cpu'))
     scaler = preprocessing.StandardScaler().fit(feature_embs)
     feature_embs = scaler.transform(feature_embs)
-    print(len(feature_embs), len(ys))
+    # print(len(feature_embs), len(ys))
     
     lin_class = SVC(max_iter=50000, kernel='linear',  class_weight='balanced', C=.1).fit(feature_embs, np.array([ys[i] for i in indices]))
     coef_ = torch.tensor(lin_class.coef_, dtype=torch.double)
@@ -222,16 +223,16 @@ def next_image(embs, img_embs, ys, calibrate_prompts):
         
     with torch.no_grad():
         if len(calibrate_prompts) > 0:
-            print('######### Calibrating with sample prompts #########')
+            # print('######### Calibrating with sample prompts #########')
             prompt = calibrate_prompts.pop(0)
-            print(prompt)
+            # print(prompt)
             image, img_emb = predict(prompt)
             im_emb = autoencoder.embed(prompt)
             embs.append(im_emb)
             img_embs.append(img_emb)
             return image, embs, img_embs, ys, calibrate_prompts, prompt
         else:
-            print('######### Roaming #########')
+            # print('######### Roaming #########')
             
             pos_indices = [i for i in range(len(embs)) if ys[i] == 1]
             neg_indices = [i for i in range(len(embs)) if ys[i] == 0]
@@ -251,7 +252,7 @@ def next_image(embs, img_embs, ys, calibrate_prompts):
             
             prompt= '' if not glob_idx % 3 == 0 else rng_prompt
             prompt, _ = generate(prompt, in_embs=im_s)
-            print(prompt)
+            # print(prompt)
             im_emb = autoencoder.embed(prompt)
             embs.append(im_emb)
             
@@ -298,7 +299,7 @@ def choose(img, choice, embs, img_embs, ys, calibrate_prompts):
         choice = 0
     
     if img is None:
-        print('NSFW -- choice is disliked')
+        # print('NSFW -- choice is disliked')
         choice = 0
         
     ys.append(choice)