Update app.py
Browse files
app.py
CHANGED
|
@@ -11,51 +11,39 @@ def respond(
|
|
| 11 |
top_p,
|
| 12 |
hf_token: gr.OAuthToken,
|
| 13 |
):
|
| 14 |
-
# Initialize the Hugging Face inference client for Gemma
|
| 15 |
-
client = InferenceClient(
|
| 16 |
-
model="google/gemma-2b-it",
|
| 17 |
-
token=hf_token.token,
|
| 18 |
-
)
|
| 19 |
|
| 20 |
-
|
|
|
|
| 21 |
messages = [{"role": "system", "content": system_message}]
|
|
|
|
| 22 |
messages.extend(history)
|
| 23 |
-
messages.append({"role": "user", "content": message})
|
| 24 |
|
| 25 |
-
|
| 26 |
-
prompt = ""
|
| 27 |
-
for msg in messages:
|
| 28 |
-
if msg["role"] == "system":
|
| 29 |
-
prompt += f"System: {msg['content']}\n"
|
| 30 |
-
elif msg["role"] == "user":
|
| 31 |
-
prompt += f"User: {msg['content']}\n"
|
| 32 |
-
elif msg["role"] == "assistant":
|
| 33 |
-
prompt += f"Assistant: {msg['content']}\n"
|
| 34 |
-
prompt += "Assistant:"
|
| 35 |
|
| 36 |
response = ""
|
| 37 |
|
| 38 |
-
|
| 39 |
-
|
| 40 |
-
|
| 41 |
-
max_new_tokens=max_tokens,
|
| 42 |
stream=True,
|
| 43 |
temperature=temperature,
|
| 44 |
top_p=top_p,
|
| 45 |
):
|
| 46 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 47 |
yield response
|
| 48 |
|
| 49 |
|
| 50 |
chatbot = gr.ChatInterface(
|
| 51 |
-
|
| 52 |
type="messages",
|
| 53 |
additional_inputs=[
|
| 54 |
-
gr.Textbox(
|
| 55 |
-
|
| 56 |
-
label="System message",
|
| 57 |
-
),
|
| 58 |
-
gr.Slider(minimum=1, maximum=4096, value=2048, step=1, label="Max new tokens"),
|
| 59 |
gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
|
| 60 |
gr.Slider(
|
| 61 |
minimum=0.1,
|
|
@@ -72,5 +60,6 @@ with gr.Blocks() as demo:
|
|
| 72 |
gr.LoginButton()
|
| 73 |
chatbot.render()
|
| 74 |
|
|
|
|
| 75 |
if __name__ == "__main__":
|
| 76 |
demo.launch()
|
|
|
|
| 11 |
top_p,
|
| 12 |
hf_token: gr.OAuthToken,
|
| 13 |
):
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 14 |
|
| 15 |
+
client = InferenceClient(token=hf_token.token, model="openai/gpt-oss-20b")
|
| 16 |
+
|
| 17 |
messages = [{"role": "system", "content": system_message}]
|
| 18 |
+
|
| 19 |
messages.extend(history)
|
|
|
|
| 20 |
|
| 21 |
+
messages.append({"role": "user", "content": message})
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 22 |
|
| 23 |
response = ""
|
| 24 |
|
| 25 |
+
for message in client.chat_completion(
|
| 26 |
+
messages,
|
| 27 |
+
max_tokens=max_tokens,
|
|
|
|
| 28 |
stream=True,
|
| 29 |
temperature=temperature,
|
| 30 |
top_p=top_p,
|
| 31 |
):
|
| 32 |
+
choices = message.choices
|
| 33 |
+
token = ""
|
| 34 |
+
if len(choices) and choices[0].delta.content:
|
| 35 |
+
token = choices[0].delta.content
|
| 36 |
+
|
| 37 |
+
response += token
|
| 38 |
yield response
|
| 39 |
|
| 40 |
|
| 41 |
chatbot = gr.ChatInterface(
|
| 42 |
+
respond,
|
| 43 |
type="messages",
|
| 44 |
additional_inputs=[
|
| 45 |
+
gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
|
| 46 |
+
gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
|
|
|
|
|
|
|
|
|
|
| 47 |
gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
|
| 48 |
gr.Slider(
|
| 49 |
minimum=0.1,
|
|
|
|
| 60 |
gr.LoginButton()
|
| 61 |
chatbot.render()
|
| 62 |
|
| 63 |
+
|
| 64 |
if __name__ == "__main__":
|
| 65 |
demo.launch()
|