Upload app.py with huggingface_hub

app.py

@@ -0,0 +1,40 @@
+import gradio as gr
+import json
+
+MODELS = [
+    {"name": "SmolLM2-135M-Instruct-mobile", "size_mb": 101, "ram_mb": 400, "tps": 59.7, "phone_tps": 46.0},
+    {"name": "Qwen2.5-0.5B-Instruct-mobile-int4", "size_mb": 469, "ram_mb": 550, "tps": 12.5, "phone_tps": 23.2},
+    {"name": "Llama-3.2-1B-Instruct-Q4-mobile", "size_mb": 770, "ram_mb": 1100, "tps": 11.3, "phone_tps": 5.4},
+    {"name": "TinyLlama-1.1B-Chat-Q5-mobile", "size_mb": 800, "ram_mb": 1200, "tps": 10.9, "phone_tps": ""},
+    {"name": "Gemma-2-2B-IT-Q5-mobile", "size_mb": 1500, "ram_mb": 2200, "tps": 5.7, "phone_tps": ""},
+    {"name": "Llama-3.2-3B-Instruct-Q5-mobile", "size_mb": 2100, "ram_mb": 2700, "tps": 2.6, "phone_tps": ""},
+]
+
+def check_ram(ram_mb: int) -> str:
+    """Check which dispatchAI models fit in a given RAM budget.
+    
+    Args:
+        ram_mb: Available RAM in MB
+    
+    Returns:
+        JSON with models that fit and their specs
+    """
+    fitting = [m for m in MODELS if m["ram_mb"] <= ram_mb]
+    not_fitting = [m for m in MODELS if m["ram_mb"] > ram_mb]
+    
+    return json.dumps({
+        "ram_budget_mb": ram_mb,
+        "models_that_fit": len(fitting),
+        "best_option": fitting[0] if fitting else None,
+        "all_fitting": fitting,
+        "too_large": [{"name": m["name"], "needs_mb": m["ram_mb"]} for m in not_fitting],
+    }, indent=2)
+
+with gr.Blocks(theme=gr.themes.Soft(primary_hue="blue"), title="RAM Calculator") as demo:
+    gr.Markdown("# 💾 Phone RAM Calculator (MCP)")
+    ram = gr.Slider(512, 8192, value=2048, step=256, label="Available RAM (MB)")
+    btn = gr.Button("Check", variant="primary")
+    out = gr.Textbox(label="Results (JSON)", lines=15)
+    btn.click(fn=check_ram, inputs=ram, outputs=out)
+    gr.Markdown("---\n🚀 [dispatchAI](https://huggingface.co/dispatchAI)")
+    demo.launch(mcp_server=True)