Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -32,102 +32,111 @@ def load_model_and_tokenizer(model_name):
|
|
32 |
|
33 |
return model_cache[model_name], tokenizer_cache[model_name]
|
34 |
|
35 |
-
|
36 |
-
|
37 |
-
|
38 |
-
|
39 |
-
|
40 |
-
|
41 |
-
|
42 |
-
top_p,
|
43 |
-
):
|
44 |
# Load selected model and tokenizer
|
45 |
-
|
46 |
-
model, tokenizer = load_model_and_tokenizer(model_name)
|
47 |
|
48 |
# Build conversation messages
|
49 |
messages = [{"role": "system", "content": system_message}]
|
50 |
-
for user_msg, assistant_msg in
|
51 |
-
|
52 |
-
|
53 |
-
|
54 |
-
messages.append({"role": "assistant", "content": assistant_msg})
|
55 |
|
56 |
-
|
|
|
57 |
|
58 |
-
# Format prompt using chat template
|
59 |
-
prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
|
60 |
-
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
|
61 |
|
62 |
-
# Set up streaming
|
63 |
-
streamer = TextIteratorStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True)
|
64 |
|
65 |
-
# Configure generation parameters
|
66 |
-
|
67 |
-
|
68 |
-
|
69 |
-
|
70 |
-
|
71 |
-
|
72 |
-
|
73 |
-
|
74 |
-
|
75 |
-
)
|
76 |
|
77 |
-
# Start generation in separate thread
|
78 |
-
thread = Thread(target=model.generate, kwargs=generation_kwargs)
|
79 |
-
thread.start()
|
80 |
|
81 |
-
# Stream response
|
82 |
-
|
83 |
-
for
|
84 |
-
|
85 |
-
yield
|
86 |
-
|
87 |
-
|
88 |
-
available_models = [
|
89 |
-
"GoofyLM/BrainrotLM-Assistant-362M",
|
90 |
-
"GoofyLM/BrainrotLM2-Assistant-362M"
|
91 |
-
]
|
92 |
|
93 |
-
# Create Gradio interface
|
94 |
with gr.Blocks() as demo:
|
95 |
gr.Markdown("# BrainrotLM Chat Interface")
|
96 |
|
97 |
with gr.Row():
|
98 |
-
with gr.Column(scale=
|
99 |
chatbot = gr.Chatbot(height=600)
|
100 |
-
|
101 |
-
|
102 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
103 |
with gr.Column(scale=1):
|
104 |
model_dropdown = gr.Dropdown(
|
105 |
choices=available_models,
|
106 |
value=available_models[0],
|
107 |
label="Select Model"
|
108 |
)
|
|
|
109 |
system_message = gr.Textbox(
|
110 |
-
value="Your name is BrainrotLM, an AI assistant trained by GoofyLM.",
|
111 |
label="System message",
|
112 |
lines=4
|
113 |
)
|
|
|
114 |
max_tokens = gr.Slider(1, 512, value=72, label="Max new tokens")
|
115 |
temperature = gr.Slider(0.1, 2.0, value=0.65, label="Temperature")
|
116 |
top_p = gr.Slider(0.1, 1.0, value=0.95, label="Top-p (nucleus sampling)")
|
117 |
|
118 |
-
|
119 |
-
|
120 |
-
|
121 |
-
|
122 |
-
|
123 |
-
|
124 |
-
|
125 |
-
|
126 |
-
|
127 |
-
|
128 |
-
|
129 |
-
],
|
130 |
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
131 |
|
132 |
if __name__ == "__main__":
|
133 |
demo.launch()
|
|
|
32 |
|
33 |
return model_cache[model_name], tokenizer_cache[model_name]
|
34 |
|
35 |
+
# Define available models
|
36 |
+
available_models = [
|
37 |
+
"GoofyLM/BrainrotLM-Assistant-362M",
|
38 |
+
"GoofyLM/BrainrotLM2-Assistant-362M"
|
39 |
+
]
|
40 |
+
|
41 |
+
def respond(message, chat_history, model_choice, system_message, max_tokens, temperature, top_p):
|
|
|
|
|
42 |
# Load selected model and tokenizer
|
43 |
+
model, tokenizer = load_model_and_tokenizer(model_choice)
|
|
|
44 |
|
45 |
# Build conversation messages
|
46 |
messages = [{"role": "system", "content": system_message}]
|
47 |
+
for user_msg, assistant_msg in chat_history:
|
48 |
+
messages.append({"role": "user", "content": user_msg})
|
49 |
+
if assistant_msg: # This might be None during streaming
|
50 |
+
messages.append({"role": "assistant", "content": assistant_msg})
|
|
|
51 |
|
52 |
+
# Add the current message
|
53 |
+
messages.append({"role": "user", "content": message})
|
54 |
|
55 |
+
# Format prompt using chat template
|
56 |
+
prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
|
57 |
+
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
|
58 |
|
59 |
+
# Set up streaming
|
60 |
+
streamer = TextIteratorStreamer(tokenizer, skip_prompt=True, skip_special_tokens=True)
|
61 |
|
62 |
+
# Configure generation parameters
|
63 |
+
generation_kwargs = dict(
|
64 |
+
**inputs,
|
65 |
+
streamer=streamer,
|
66 |
+
max_new_tokens=max_tokens,
|
67 |
+
temperature=temperature,
|
68 |
+
top_p=top_p,
|
69 |
+
do_sample=(temperature > 0 or top_p < 1.0),
|
70 |
+
pad_token_id=tokenizer.pad_token_id
|
71 |
+
)
|
|
|
72 |
|
73 |
+
# Start generation in a separate thread
|
74 |
+
thread = Thread(target=model.generate, kwargs=generation_kwargs)
|
75 |
+
thread.start()
|
76 |
|
77 |
+
# Stream the response
|
78 |
+
partial_message = ""
|
79 |
+
for new_token in streamer:
|
80 |
+
partial_message += new_token
|
81 |
+
yield chat_history + [(message, partial_message)]
|
82 |
+
|
83 |
+
return chat_history + [(message, partial_message)]
|
|
|
|
|
|
|
|
|
84 |
|
85 |
+
# Create the Gradio interface
|
86 |
with gr.Blocks() as demo:
|
87 |
gr.Markdown("# BrainrotLM Chat Interface")
|
88 |
|
89 |
with gr.Row():
|
90 |
+
with gr.Column(scale=3):
|
91 |
chatbot = gr.Chatbot(height=600)
|
92 |
+
|
93 |
+
with gr.Row():
|
94 |
+
msg = gr.Textbox(
|
95 |
+
label="Message",
|
96 |
+
placeholder="Type your message here...",
|
97 |
+
lines=3,
|
98 |
+
show_label=False
|
99 |
+
)
|
100 |
+
submit = gr.Button("Send", variant="primary")
|
101 |
+
|
102 |
+
clear = gr.Button("Clear Conversation")
|
103 |
+
|
104 |
with gr.Column(scale=1):
|
105 |
model_dropdown = gr.Dropdown(
|
106 |
choices=available_models,
|
107 |
value=available_models[0],
|
108 |
label="Select Model"
|
109 |
)
|
110 |
+
|
111 |
system_message = gr.Textbox(
|
112 |
+
value="Your name is BrainrotLM, an AI assistant trained by GoofyLM.",
|
113 |
label="System message",
|
114 |
lines=4
|
115 |
)
|
116 |
+
|
117 |
max_tokens = gr.Slider(1, 512, value=72, label="Max new tokens")
|
118 |
temperature = gr.Slider(0.1, 2.0, value=0.65, label="Temperature")
|
119 |
top_p = gr.Slider(0.1, 1.0, value=0.95, label="Top-p (nucleus sampling)")
|
120 |
|
121 |
+
# Set up event handlers
|
122 |
+
submit_event = msg.submit(
|
123 |
+
respond,
|
124 |
+
inputs=[msg, chatbot, model_dropdown, system_message, max_tokens, temperature, top_p],
|
125 |
+
outputs=chatbot
|
126 |
+
)
|
127 |
+
|
128 |
+
submit_click = submit.click(
|
129 |
+
respond,
|
130 |
+
inputs=[msg, chatbot, model_dropdown, system_message, max_tokens, temperature, top_p],
|
131 |
+
outputs=chatbot
|
|
|
132 |
)
|
133 |
+
|
134 |
+
# Clear message box after sending
|
135 |
+
submit_event.then(lambda: "", None, msg)
|
136 |
+
submit_click.then(lambda: "", None, msg)
|
137 |
+
|
138 |
+
# Clear conversation button
|
139 |
+
clear.click(lambda: None, None, chatbot)
|
140 |
|
141 |
if __name__ == "__main__":
|
142 |
demo.launch()
|