Spaces:
Running
Running
space update
Browse files
app.py
CHANGED
@@ -182,30 +182,12 @@ def load_model_from_file(uploaded_file):
|
|
182 |
model_status = error_msg
|
183 |
return error_msg
|
184 |
|
185 |
-
def
|
186 |
-
|
187 |
-
history: list[tuple[str, str]],
|
188 |
-
system_message,
|
189 |
-
max_tokens,
|
190 |
-
temperature,
|
191 |
-
top_p,
|
192 |
-
model_file,
|
193 |
-
model_url,
|
194 |
-
load_file_btn,
|
195 |
-
load_url_btn,
|
196 |
-
status_display
|
197 |
-
):
|
198 |
-
"""
|
199 |
-
Generate a response using the UstaModel
|
200 |
-
"""
|
201 |
if model is None or tokenizer is None:
|
202 |
-
|
203 |
-
return
|
204 |
|
205 |
try:
|
206 |
-
# For UstaModel, we'll use the message directly (ignoring system_message for now)
|
207 |
-
# since it's a simpler model focused on geographical knowledge
|
208 |
-
|
209 |
# Encode the input message
|
210 |
tokens = tokenizer.encode(message)
|
211 |
|
@@ -215,7 +197,6 @@ def respond(
|
|
215 |
|
216 |
# Generate response
|
217 |
with torch.no_grad():
|
218 |
-
# Use max_tokens parameter, but cap it at reasonable limit for this model
|
219 |
actual_max_tokens = min(max_tokens, 32 - len(tokens))
|
220 |
generated_tokens = model.generate(tokens, actual_max_tokens)
|
221 |
|
@@ -233,104 +214,76 @@ def respond(
|
|
233 |
if not response:
|
234 |
response = "I'm not sure how to respond to that with my geographical knowledge."
|
235 |
|
236 |
-
#
|
237 |
-
|
|
|
238 |
|
239 |
except Exception as e:
|
240 |
-
|
|
|
241 |
|
242 |
-
# Create
|
243 |
-
with gr.Blocks(title="π€ Usta Model Chat"
|
244 |
gr.Markdown("# π€ Usta Model Chat")
|
245 |
-
gr.Markdown("Chat with a custom transformer language model built from scratch!
|
246 |
|
247 |
-
#
|
248 |
-
|
249 |
-
|
250 |
-
with gr.Column():
|
251 |
-
gr.Markdown("### π Upload Model File")
|
252 |
-
model_file = gr.File(label="Upload Model File (.pth)", file_types=[".pth", ".pt"])
|
253 |
-
load_file_btn = gr.Button("Load from File", variant="primary")
|
254 |
-
|
255 |
-
with gr.Column():
|
256 |
-
gr.Markdown("### π Load from URL")
|
257 |
-
model_url = gr.Textbox(label="Model URL", placeholder="https://github.com/malibayram/llm-from-scratch/raw/main/u_model_4000.pth")
|
258 |
-
load_url_btn = gr.Button("Load from URL", variant="primary")
|
259 |
-
|
260 |
-
status_display = gr.Textbox(label="Model Status", value=model_status, interactive=False)
|
261 |
|
262 |
-
|
263 |
-
|
264 |
-
|
265 |
|
266 |
# Generation settings
|
267 |
-
|
268 |
-
|
269 |
-
|
270 |
-
|
|
|
|
|
|
|
|
|
|
|
271 |
)
|
272 |
-
|
273 |
-
temperature = gr.Slider(minimum=0.1, maximum=2.0, value=1.0, step=0.1, label="Temperature")
|
274 |
-
top_p = gr.Slider(minimum=0.1, maximum=1.0, value=0.95, step=0.05, label="Top-p")
|
275 |
|
276 |
-
# Button controls
|
277 |
with gr.Row():
|
278 |
-
|
279 |
-
|
|
|
|
|
280 |
|
281 |
# Event handlers
|
282 |
-
def
|
283 |
if not message.strip():
|
284 |
return history, ""
|
285 |
-
|
286 |
-
# Convert messages format for our respond function
|
287 |
-
tuple_history = [(h["role"], h["content"]) for h in history if h["role"] != "system"]
|
288 |
-
|
289 |
-
# Generate response using our existing function
|
290 |
-
response_gen = respond(
|
291 |
-
message, tuple_history, sys_msg, max_tok, temp, top_p_val,
|
292 |
-
None, None, None, None, None # Dummy values for unused params
|
293 |
-
)
|
294 |
-
|
295 |
-
# Get the response
|
296 |
-
response = ""
|
297 |
-
for r in response_gen:
|
298 |
-
response = r
|
299 |
-
|
300 |
-
# Add to history in messages format
|
301 |
-
history.append({"role": "user", "content": message})
|
302 |
-
history.append({"role": "assistant", "content": response})
|
303 |
-
|
304 |
-
return history, ""
|
305 |
|
306 |
-
|
307 |
-
|
308 |
-
|
309 |
-
inputs=[msg, chatbot, system_msg, max_tokens, temperature, top_p],
|
310 |
outputs=[chatbot, msg]
|
311 |
)
|
312 |
|
313 |
msg.submit(
|
314 |
-
|
315 |
-
inputs=[msg, chatbot,
|
316 |
outputs=[chatbot, msg]
|
317 |
)
|
318 |
|
319 |
-
clear_btn.click(
|
320 |
-
|
321 |
-
|
|
|
|
|
|
|
322 |
)
|
323 |
|
324 |
load_file_btn.click(
|
325 |
load_model_from_file,
|
326 |
inputs=[model_file],
|
327 |
-
outputs=[
|
328 |
-
)
|
329 |
-
|
330 |
-
load_url_btn.click(
|
331 |
-
load_model_from_url,
|
332 |
-
inputs=[model_url],
|
333 |
-
outputs=[status_display]
|
334 |
)
|
335 |
|
336 |
if __name__ == "__main__":
|
|
|
182 |
model_status = error_msg
|
183 |
return error_msg
|
184 |
|
185 |
+
def chat_with_usta(message, history, max_tokens=20):
|
186 |
+
"""Simple chat function"""
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
187 |
if model is None or tokenizer is None:
|
188 |
+
return history + [["Error", "UstaModel is not available. Please try again later."]]
|
|
|
189 |
|
190 |
try:
|
|
|
|
|
|
|
191 |
# Encode the input message
|
192 |
tokens = tokenizer.encode(message)
|
193 |
|
|
|
197 |
|
198 |
# Generate response
|
199 |
with torch.no_grad():
|
|
|
200 |
actual_max_tokens = min(max_tokens, 32 - len(tokens))
|
201 |
generated_tokens = model.generate(tokens, actual_max_tokens)
|
202 |
|
|
|
214 |
if not response:
|
215 |
response = "I'm not sure how to respond to that with my geographical knowledge."
|
216 |
|
217 |
+
# Add to history
|
218 |
+
history.append([message, response])
|
219 |
+
return history
|
220 |
|
221 |
except Exception as e:
|
222 |
+
history.append([message, f"Sorry, I encountered an error: {str(e)}"])
|
223 |
+
return history
|
224 |
|
225 |
+
# Create simple interface
|
226 |
+
with gr.Blocks(title="π€ Usta Model Chat") as demo:
|
227 |
gr.Markdown("# π€ Usta Model Chat")
|
228 |
+
gr.Markdown("Chat with a custom transformer language model built from scratch! This model specializes in geographical knowledge.")
|
229 |
|
230 |
+
# Simple chat interface
|
231 |
+
chatbot = gr.Chatbot(height=400)
|
232 |
+
msg = gr.Textbox(label="Your message", placeholder="Ask about countries, capitals, or cities...")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
233 |
|
234 |
+
with gr.Row():
|
235 |
+
send_btn = gr.Button("Send", variant="primary")
|
236 |
+
clear_btn = gr.Button("Clear")
|
237 |
|
238 |
# Generation settings
|
239 |
+
max_tokens = gr.Slider(minimum=1, maximum=30, value=20, step=1, label="Max tokens")
|
240 |
+
|
241 |
+
# Model loading (simplified)
|
242 |
+
gr.Markdown("## π§ Load Custom Model (Optional)")
|
243 |
+
with gr.Row():
|
244 |
+
model_url = gr.Textbox(
|
245 |
+
label="Model URL",
|
246 |
+
placeholder="https://github.com/malibayram/llm-from-scratch/raw/main/u_model_4000.pth",
|
247 |
+
scale=3
|
248 |
)
|
249 |
+
load_url_btn = gr.Button("Load from URL", scale=1)
|
|
|
|
|
250 |
|
|
|
251 |
with gr.Row():
|
252 |
+
model_file = gr.File(label="Upload .pth file", file_types=[".pth"])
|
253 |
+
load_file_btn = gr.Button("Load File", scale=1)
|
254 |
+
|
255 |
+
status = gr.Textbox(label="Status", value=model_status, interactive=False)
|
256 |
|
257 |
# Event handlers
|
258 |
+
def send_message(message, history, max_tok):
|
259 |
if not message.strip():
|
260 |
return history, ""
|
261 |
+
return chat_with_usta(message, history, max_tok), ""
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
262 |
|
263 |
+
send_btn.click(
|
264 |
+
send_message,
|
265 |
+
inputs=[msg, chatbot, max_tokens],
|
|
|
266 |
outputs=[chatbot, msg]
|
267 |
)
|
268 |
|
269 |
msg.submit(
|
270 |
+
send_message,
|
271 |
+
inputs=[msg, chatbot, max_tokens],
|
272 |
outputs=[chatbot, msg]
|
273 |
)
|
274 |
|
275 |
+
clear_btn.click(lambda: [], outputs=[chatbot])
|
276 |
+
|
277 |
+
load_url_btn.click(
|
278 |
+
load_model_from_url,
|
279 |
+
inputs=[model_url],
|
280 |
+
outputs=[status]
|
281 |
)
|
282 |
|
283 |
load_file_btn.click(
|
284 |
load_model_from_file,
|
285 |
inputs=[model_file],
|
286 |
+
outputs=[status]
|
|
|
|
|
|
|
|
|
|
|
|
|
287 |
)
|
288 |
|
289 |
if __name__ == "__main__":
|