Update app.py

app.py

@@ -4,7 +4,8 @@ import os
 
 tg.set_backward_engine(tg.get_engine("gpt-4o"))
 
-initial_solution = """To solve the equation 3x^2 - 7x + 2 = 0, we use the quadratic formula:
+# Hardcoded examples
+default_initial_solution = """To solve the equation 3x^2 - 7x + 2 = 0, we use the quadratic formula:
 x = (-b ± √(b^2 - 4ac)) / 2a
 a = 3, b = -7, c = 2
 x = (7 ± √((-7)^2 + 4(3)(2))) / 6
@@ -13,21 +14,36 @@ The solutions are:
 x1 = (7 + √73)
 x2 = (7 - √73)"""
 
+default_loss_system_prompt = """You will evaluate a solution to a math question. 
+Do not attempt to solve it yourself, do not give a solution, only identify errors. Be super concise."""
+
+# Display clickable examples
+if st.button("Use Example Initial Solution"):
+    st.session_state.initial_solution = default_initial_solution
+
+if st.button("Use Example Loss System Prompt"):
+    st.session_state.loss_system_prompt = default_loss_system_prompt
+
+# Input boxes
+initial_solution = st.text_area("Initial Solution", st.session_state.get("initial_solution", ""))
+loss_system_prompt = st.text_area("Loss System Prompt", st.session_state.get("loss_system_prompt", ""))
+num_epochs = st.number_input("Epochs", min_value=1, value=1)
+
+# Set up the textgrad variables
 solution = tg.Variable(initial_solution,
                        requires_grad=True,
                        role_description="solution to the math question")
 
-loss_system_prompt = tg.Variable("""You will evaluate a solution to a math question. 
-Do not attempt to solve it yourself, do not give a solution, only identify errors. Be super concise.""",
-                                 requires_grad=False,
-                                 role_description="system prompt")
-                              
-loss_fn = tg.TextLoss(loss_system_prompt)
+loss_fn = tg.TextLoss(tg.Variable(loss_system_prompt,
+                                  requires_grad=False,
+                                  role_description="system prompt"))
 optimizer = tg.TGD([solution])
 
-loss = loss_fn(solution)
-st.write(loss.value)
+# Training loop
+for i in range(num_epochs):
+    loss = loss_fn(solution)
+    loss.backward()
+    optimizer.step()
 
-loss.backward()
-optimizer.step()
-st.write(solution.value)
+# Output box
+st.text_area("Result", solution.value)