new prompts

app.py

@@ -235,12 +235,14 @@ def inference(
         autocast = torch.cuda.amp.autocast
     with autocast(fp16):
         print(f"Generating for prompt: {prompt}")
+        progress_bar = st.progress(0)
         sequences, tokens = generator.sample_sequences(
             num_samples=1,
             input_prompt=prompt,
             max_length=20,
             caif_period=1,
             entropy=entropy_threshold,
+            progress_bar=progress_bar,
             **kwargs
         )
         print(f"Output for prompt: {sequences}")

generator.py

@@ -44,6 +44,7 @@ class Generator:
         caif_period: int,
         caif_tokens_num: Union[int, None] = None,
         entropy: float = None,
+        progress_bar = None,
         **sampler_kwargs
     ):
         self.entropy = entropy
@@ -65,6 +66,7 @@ class Generator:
                 caif_tokens_num=caif_tokens_num,
                 **sampler_kwargs
             )
+            progress_bar.step((i+1)/max_length)
             if ended_sequences.all():
                 break