Update app.py

app.py

@@ -86,8 +86,6 @@ def print_kernel_execution(c1, c2, comp_flop, mem_bytes):
   
   return exec_time
 
-
-
 st.sidebar.header("Transformer parameters")
 col1, col2 = st.sidebar.columns([2, 4])
 
@@ -115,7 +113,7 @@ for i in range(n_start, n):
   mqa_time = shared_time + qkv_mqa_exec(bs, h, i, d)[2] + att1_mqa_exec(bs, h, i, d)[2] + att2_mqa_exec(bs, h, i, d)[2]
   mqa_total_time += l*mqa_time
   
-c1, c2 = st.columns(2, 4)
+c1, c2 = st.columns([2, 4])
 c1.write("Multi-Head Attention:")
 c2.write(str(round(mha_total_time, 2)))
 c1.write("Multi-Query Attention:")
@@ -125,7 +123,7 @@ c2.write(str(round(mha_total_time/mqa_total_time),2))
 
 st.header("Memory consumption")
 st.caption("MHA")
-c1, c2 = st.columns(2, 4)
+c1, c2 = st.columns([2, 4])
 num_params = 12*l*d*d
 c1.write("Num Parameters (in B)")
 c2.write(str(round(num_params/1e9, 3)))