mememechez's picture
Deploy final cleaned source code
ca28016
raw
history blame
2.97 kB
"use client";
import React from 'react';
import { Card, CardContent, CardHeader, CardTitle } from "@/components/ui/card";
import { CircuitBoard, Zap, Shield, Atom } from "lucide-react";
export default function ModelSummary() {
return (
<Card className="bg-card/80 backdrop-blur-sm">
<CardHeader>
<CardTitle className="font-headline">Model Architecture Key Features</CardTitle>
</CardHeader>
<CardContent>
<div className="space-y-6">
<div className="flex items-start gap-4">
<div className="p-2 bg-primary/10 rounded-md">
<CircuitBoard className="h-6 w-6 text-primary" />
</div>
<div>
<h3 className="font-semibold text-accent-foreground">Network Structure</h3>
<p className="text-sm text-muted-foreground mt-1">
Convolutional backbone with 4 conv layers (64-128-256-512 channels) and 3 fully connected layers (2048-512-10). Employs GELU activations and SE Blocks for dynamic feature recalibration.
</p>
</div>
</div>
<div className="flex items-start gap-4">
<div className="p-2 bg-primary/10 rounded-md">
<Zap className="h-6 w-6 text-primary" />
</div>
<div>
<h3 className="font-semibold text-accent-foreground">ZPE Flow Layers</h3>
<p className="text-sm text-muted-foreground mt-1">
Zero-Point Energy flow applied after each major block, featuring tunable momentum, strength, noise, and coupling parameters for adaptive network modulation.
</p>
</div>
</div>
<div className="flex items-start gap-4">
<div className="p-2 bg-primary/10 rounded-md">
<Shield className="h-6 w-6 text-primary" />
</div>
<div>
<h3 className="font-semibold text-accent-foreground">Advanced Regularization</h3>
<p className="text-sm text-muted-foreground mt-1">
Integrates skip connections for improved gradient flow, dropout (0.05-0.25), and label smoothing (0.03) to enhance generalization and prevent overfitting.
</p>
</div>
</div>
<div className="flex items-start gap-4">
<div className="p-2 bg-primary/10 rounded-md">
<Atom className="h-6 w-6 text-primary" />
</div>
<div>
<h3 className="font-semibold text-accent-foreground">Quantum Integration</h3>
<p className="text-sm text-muted-foreground mt-1">
Strategic quantum noise injection, simulated via a 32-qubit circuit, primarily targets the 4th convolutional layer to enhance feature representation at high complexity stages.
</p>
</div>
</div>
</div>
</CardContent>
</Card>
);
}