a

app.py

@@ -32,10 +32,10 @@ def infer(input_ids, bad_words_ids, max_tokens, temperature, top_k, top_p):
 
     return output_sequences
 
-default_value = "We are building the world's first"
+default_value = "We are building the world's first "
     
 #prompts
-st.title("Write with vc.meraGPT.com 🦄")
+st.title("Write with a VC 🦄")
 st.write("This is a LLM that was fine-tuned on a dataset of investment memos to help you generate your next pitch.")
 
 sent = st.text_area("Text", default_value)
@@ -50,7 +50,7 @@ top_p = st.sidebar.slider("Top-p", min_value = 0.0, max_value=1.0, step = 0.05,
 encoded_prompt = tokenizer.encode(tokenizer.eos_token+sent, max_length=1024, return_tensors="pt", truncation=True)
 
 # get tokens of words that should not be generated
-bad_words_ids = tokenizer(["confidential", "angel.co", "angellist.com"], add_special_tokens=False).input_ids
+bad_words_ids = tokenizer(["confidential", "angel.co", "angellist.com", "angellist"], add_special_tokens=False).input_ids
 
 if encoded_prompt.size()[-1] == 0:
     input_ids = None