Update app.py

app.py

@@ -13,7 +13,7 @@ n = st.sidebar.number_input('End seq', value=1024)
 
 st.sidebar.header("GPU parameters")
 
-GPU = st.selectbox('GPU', ('A100', 'V100'))
+GPU = st.sidebar.selectbox('GPU', ('A100', 'V100'))
 
 if GPU == 'A100':
   # A100 specs
@@ -97,7 +97,7 @@ def mlp_exec(bs, h, n, d):
   
 def print_kernel_execution(flop, mem_bytes):
   c1, c2 = st.columns([2, 3])
-  exec_time = calc_exec_time(comp_flop, mem_bytes, include_overhead=False)
+  exec_time = calc_exec_time(flop, nbytes, include_overhead=False)
   flop = round(flop/1e9, 2)
   nbytes = round(nbytes/1e6, 2)
   
@@ -167,7 +167,6 @@ if breakdown:
   flop, nbytes, exec_time = qkv_mha_exec(bs, h, n, d)
   print_kernel_execution(flop, nbytes)
  
-  
   st.caption("Multi-Query Attention")
   flop, nbytes, exec_time = qkv_mqa_exec(bs, h, n, d)
   print_kernel_execution(flop, nbytes)