Update app.py

app.py

@@ -149,7 +149,7 @@ c1.write("Cached keys and values (GB)")
 acts = round(2*bs*l*(d/h)*h*2*n/1e9, 2)
 c2.write(str(acts))
 
-st.subheader("Multi-Query Attention")
+st.caption("Multi-Query Attention")
 c1, c2 = st.columns([2, 4])
 num_params = (10+2/h)*l*d*d
 c1.write("Num Parameters (in B)")
@@ -160,8 +160,12 @@ c1.write("Cached keys and values (GB)")
 acts = round(2*bs*l*(d/h)*2*n/1e9, 2)
 c2.write(str(acts))
 
-st.subheader("How we approximate the inference time")
-st.write("TODO")
+st.subheader("Approximations")
+st.markdown("We estimate the time for each matrix multiplication")
+
+st.latex("C = A \dot B, A \in R^{NxM}")
+
+
 
 breakdown = st.checkbox("Show breakdown per operation")
 if breakdown: