Update app.py

app.py

@@ -64,14 +64,17 @@ with gr.Blocks() as demo:
     hidden_size = gr.Number(label="Hidden Size", value=6144)
     sequence_length = gr.Number(label="Sequence Length", value=2048)
     num_layers = gr.Number(label="Number of Layers", value=44)
-    moe = gr.Checkbox(label="MoE", value=False)
-    num_experts = gr.Number(label="Number of Experts", value=8)
-    expert_interval = gr.Number(label="Expert Interval", value=1)
-    topk = gr.Number(label="Top k Routing", value=1)
     ffn_expansion_factor = gr.Number(label="FFN Expansion Factor", value=4)
     num_mlp_linears = gr.Number(label="Number of Linear Layers per MLP Block", value=2)
     kv_size_ratio = gr.Number(label="KV Size Ratio", value=1.0)
 
+    # MoE Parameters inside an accordion
+    with gr.Accordion("MoE Parameters", open=False):
+        moe = gr.Checkbox(label="MoE", value=False)
+        num_experts = gr.Number(label="Number of Experts", value=8)
+        expert_interval = gr.Number(label="Expert Interval", value=1)
+        topk = gr.Number(label="Top k Routing", value=1)
+
     result = gr.Textbox(label="Output", interactive=False)
 
     def run_calculation(vocab_size, tied_embeddings, hidden_size, sequence_length, num_layers, moe, num_experts, expert_interval, topk, ffn_expansion_factor, num_mlp_linears, kv_size_ratio):