Update app_tsukasa_remote.py

app_tsukasa_remote.py

@@ -317,7 +317,7 @@ yielding reasonably good results.
 </p>
 
 <p>
-Overall, I'm confident that with a bit of experimentation, you can achieve reasonbaly good results. <br>
+Overall, I'm confident that with a bit of experimentation, you can achieve impressive results. <br>
 The model should work well out of the box 90% of the time without the need for extensive tweaking.<br>
 However, here are some tips in case you encounter issues:
 </p>
@@ -329,6 +329,10 @@ However, here are some tips in case you encounter issues:
     Ensure that your input closely matches your reference (audio or text prompt) in terms of tone,<br>
     non-verbal cues, duration, etc.
   </li>
+
+  <li>
+    **Do not use very short inputs**
+  </li>
   
   <li>
     If your audio is too long but the input is too short, the speech rate will be slow, and vice versa.
@@ -336,7 +340,8 @@ However, here are some tips in case you encounter issues:
   
   <li>
     Experiment with the <b>alpha</b>, <b>beta</b>, and <b>Intensity</b> parameters. The Diffusion<br>
-    sampler is non-deterministic, so regenerate a few times if you're not satisfied with the output.
+    sampler is non-deterministic, so regenerate a few times if you're not satisfied with the output. <br> 
+    don't use crazy values though! turn it off if it was bad by setting both alpha and beta to zero.
   </li>
   
   <li>