Spaces:
Running
Running
Update app.py
Browse files
app.py
CHANGED
|
@@ -105,13 +105,14 @@ def predict_chat(message: str, history: list):
|
|
| 105 |
generated_text += token
|
| 106 |
yield generated_text
|
| 107 |
|
| 108 |
-
else:
|
| 109 |
input_text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
|
| 110 |
inputs = tokenizer.encode(input_text, return_tensors="pt").to(device)
|
| 111 |
|
| 112 |
outputs = model.generate(
|
| 113 |
inputs,
|
| 114 |
-
max_new_tokens
|
|
|
|
| 115 |
temperature=TEMPERATURE,
|
| 116 |
top_k=TOP_K,
|
| 117 |
top_p=TOP_P,
|
|
@@ -157,7 +158,7 @@ if __name__ == "__main__":
|
|
| 157 |
["What's the best way to stay motivated?"],
|
| 158 |
],
|
| 159 |
cache_examples=False,
|
| 160 |
-
#
|
| 161 |
)
|
| 162 |
|
| 163 |
demo.chatbot.value = initial_messages_for_value
|
|
|
|
| 105 |
generated_text += token
|
| 106 |
yield generated_text
|
| 107 |
|
| 108 |
+
else: # This is the block where the error occurred
|
| 109 |
input_text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
|
| 110 |
inputs = tokenizer.encode(input_text, return_tensors="pt").to(device)
|
| 111 |
|
| 112 |
outputs = model.generate(
|
| 113 |
inputs,
|
| 114 |
+
# Changed max_new_tokens to max_length for broader compatibility
|
| 115 |
+
max_length=inputs.shape[-1] + MAX_NEW_TOKENS,
|
| 116 |
temperature=TEMPERATURE,
|
| 117 |
top_k=TOP_K,
|
| 118 |
top_p=TOP_P,
|
|
|
|
| 158 |
["What's the best way to stay motivated?"],
|
| 159 |
],
|
| 160 |
cache_examples=False,
|
| 161 |
+
# clear_btn="Clear Chat" was removed in the previous step
|
| 162 |
)
|
| 163 |
|
| 164 |
demo.chatbot.value = initial_messages_for_value
|