<?xml version="1.0"?>
<feed xmlns="http://www.w3.org/2005/Atom" xml:lang="en">
	<id>https://onnocenter.or.id/wiki/index.php?action=history&amp;feed=atom&amp;title=ComfyUI%3A_Instalasi_venv_GPU</id>
	<title>ComfyUI: Instalasi venv GPU - Revision history</title>
	<link rel="self" type="application/atom+xml" href="https://onnocenter.or.id/wiki/index.php?action=history&amp;feed=atom&amp;title=ComfyUI%3A_Instalasi_venv_GPU"/>
	<link rel="alternate" type="text/html" href="https://onnocenter.or.id/wiki/index.php?title=ComfyUI:_Instalasi_venv_GPU&amp;action=history"/>
	<updated>2026-05-10T10:29:29Z</updated>
	<subtitle>Revision history for this page on the wiki</subtitle>
	<generator>MediaWiki 1.35.4</generator>
	<entry>
		<id>https://onnocenter.or.id/wiki/index.php?title=ComfyUI:_Instalasi_venv_GPU&amp;diff=73523&amp;oldid=prev</id>
		<title>Onnowpurbo at 23:32, 25 March 2026</title>
		<link rel="alternate" type="text/html" href="https://onnocenter.or.id/wiki/index.php?title=ComfyUI:_Instalasi_venv_GPU&amp;diff=73523&amp;oldid=prev"/>
		<updated>2026-03-25T23:32:58Z</updated>

		<summary type="html">&lt;p&gt;&lt;/p&gt;
&lt;table class=&quot;diff diff-contentalign-left diff-editfont-monospace&quot; data-mw=&quot;interface&quot;&gt;
				&lt;col class=&quot;diff-marker&quot; /&gt;
				&lt;col class=&quot;diff-content&quot; /&gt;
				&lt;col class=&quot;diff-marker&quot; /&gt;
				&lt;col class=&quot;diff-content&quot; /&gt;
				&lt;tr class=&quot;diff-title&quot; lang=&quot;en&quot;&gt;
				&lt;td colspan=&quot;2&quot; style=&quot;background-color: #fff; color: #202122; text-align: center;&quot;&gt;← Older revision&lt;/td&gt;
				&lt;td colspan=&quot;2&quot; style=&quot;background-color: #fff; color: #202122; text-align: center;&quot;&gt;Revision as of 23:32, 25 March 2026&lt;/td&gt;
				&lt;/tr&gt;&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot; id=&quot;mw-diff-left-l3&quot; &gt;Line 3:&lt;/td&gt;
&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot;&gt;Line 3:&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;Catatan penting dulu: untuk desktop, '''RTX 5070 = 12 GB''', sedangkan '''RTX 5070 Ti = 16 GB'''. Jadi kalau Anda menulis '''5070 16 GB''', besar kemungkinan yang dimaksud adalah '''5070 Ti''', atau GPU laptop kelas 16 GB. Saya buat panduan ini aman untuk '''VRAM 12–16 GB'''. PyTorch 2.7 juga sudah menambahkan dukungan untuk arsitektur '''NVIDIA Blackwell''' dan wheel '''CUDA 12.8''', jadi jalur yang paling masuk akal untuk seri 50 adalah pakai PyTorch baru, bukan stack lama. ([NVIDIA][1])&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;Catatan penting dulu: untuk desktop, '''RTX 5070 = 12 GB''', sedangkan '''RTX 5070 Ti = 16 GB'''. Jadi kalau Anda menulis '''5070 16 GB''', besar kemungkinan yang dimaksud adalah '''5070 Ti''', atau GPU laptop kelas 16 GB. Saya buat panduan ini aman untuk '''VRAM 12–16 GB'''. PyTorch 2.7 juga sudah menambahkan dukungan untuk arsitektur '''NVIDIA Blackwell''' dan wheel '''CUDA 12.8''', jadi jalur yang paling masuk akal untuk seri 50 adalah pakai PyTorch baru, bukan stack lama. ([NVIDIA][1])&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;== Rekomendasi singkat&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;== Rekomendasi singkat&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;==&lt;/ins&gt;&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;Untuk '''gambar''', mulai dari '''SDXL''' dulu karena paling stabil untuk 12–16 GB VRAM. ComfyUI sendiri memang dipasang lewat alur '''venv → clone repo → install dependencies → run''', dan model diletakkan di folder `models/...` sesuai tipenya. ([ComfyUI Documentation][2])&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;Untuk '''gambar''', mulai dari '''SDXL''' dulu karena paling stabil untuk 12–16 GB VRAM. ComfyUI sendiri memang dipasang lewat alur '''venv → clone repo → install dependencies → run''', dan model diletakkan di folder `models/...` sesuai tipenya. ([ComfyUI Documentation][2])&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;/table&gt;</summary>
		<author><name>Onnowpurbo</name></author>
	</entry>
	<entry>
		<id>https://onnocenter.or.id/wiki/index.php?title=ComfyUI:_Instalasi_venv_GPU&amp;diff=73522&amp;oldid=prev</id>
		<title>Onnowpurbo: /* = Pilihan terbaik untuk mulai: SDXL Base 1.0 */</title>
		<link rel="alternate" type="text/html" href="https://onnocenter.or.id/wiki/index.php?title=ComfyUI:_Instalasi_venv_GPU&amp;diff=73522&amp;oldid=prev"/>
		<updated>2026-03-25T23:32:43Z</updated>

		<summary type="html">&lt;p&gt;&lt;span dir=&quot;auto&quot;&gt;&lt;span class=&quot;autocomment&quot;&gt;= Pilihan terbaik untuk mulai: SDXL Base 1.0&lt;/span&gt;&lt;/span&gt;&lt;/p&gt;
&lt;table class=&quot;diff diff-contentalign-left diff-editfont-monospace&quot; data-mw=&quot;interface&quot;&gt;
				&lt;col class=&quot;diff-marker&quot; /&gt;
				&lt;col class=&quot;diff-content&quot; /&gt;
				&lt;col class=&quot;diff-marker&quot; /&gt;
				&lt;col class=&quot;diff-content&quot; /&gt;
				&lt;tr class=&quot;diff-title&quot; lang=&quot;en&quot;&gt;
				&lt;td colspan=&quot;2&quot; style=&quot;background-color: #fff; color: #202122; text-align: center;&quot;&gt;← Older revision&lt;/td&gt;
				&lt;td colspan=&quot;2&quot; style=&quot;background-color: #fff; color: #202122; text-align: center;&quot;&gt;Revision as of 23:32, 25 March 2026&lt;/td&gt;
				&lt;/tr&gt;&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot; id=&quot;mw-diff-left-l83&quot; &gt;Line 83:&lt;/td&gt;
&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot;&gt;Line 83:&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;== A. Model grafik / gambar==&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;== A. Model grafik / gambar==&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;=== Pilihan terbaik untuk mulai: '''SDXL Base 1.0'''==&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;=== Pilihan terbaik untuk mulai: '''SDXL Base 1.0'''&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;=&lt;/ins&gt;==&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;Ini pilihan paling aman untuk GPU 12–16 GB. Cocok untuk text-to-image umum, stabil, dan sangat didukung workflow ComfyUI. ComfyUI “first generation” juga memang mengajarkan pola dasar image generation dengan memilih model di node loader. ([ComfyUI Documentation][7])&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;Ini pilihan paling aman untuk GPU 12–16 GB. Cocok untuk text-to-image umum, stabil, dan sangat didukung workflow ComfyUI. ComfyUI “first generation” juga memang mengajarkan pola dasar image generation dengan memilih model di node loader. ([ComfyUI Documentation][7])&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot; id=&quot;mw-diff-left-l114&quot; &gt;Line 114:&lt;/td&gt;
&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot;&gt;Line 114:&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;* '''ControlNet''' untuk pose/lineart/depth → taruh di `models/controlnet/`&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;* '''ControlNet''' untuk pose/lineart/depth → taruh di `models/controlnet/`&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;   ComfyUI memang memisahkan model berdasarkan tipe loader seperti `Load Checkpoint`, `Load LoRA`, dan `Load VAE`. ([ComfyUI Documentation][6])&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;   ComfyUI memang memisahkan model berdasarkan tipe loader seperti `Load Checkpoint`, `Load LoRA`, dan `Load VAE`. ([ComfyUI Documentation][6])&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del style=&quot;font-weight: bold; text-decoration: none;&quot;&gt;&lt;/del&gt;&lt;/div&gt;&lt;/td&gt;&lt;td colspan=&quot;2&quot;&gt; &lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;== B. Model audio==&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;== B. Model audio==&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;/table&gt;</summary>
		<author><name>Onnowpurbo</name></author>
	</entry>
	<entry>
		<id>https://onnocenter.or.id/wiki/index.php?title=ComfyUI:_Instalasi_venv_GPU&amp;diff=73521&amp;oldid=prev</id>
		<title>Onnowpurbo: /* Model yang relevan */</title>
		<link rel="alternate" type="text/html" href="https://onnocenter.or.id/wiki/index.php?title=ComfyUI:_Instalasi_venv_GPU&amp;diff=73521&amp;oldid=prev"/>
		<updated>2026-03-25T23:32:15Z</updated>

		<summary type="html">&lt;p&gt;&lt;span dir=&quot;auto&quot;&gt;&lt;span class=&quot;autocomment&quot;&gt;Model yang relevan&lt;/span&gt;&lt;/span&gt;&lt;/p&gt;
&lt;table class=&quot;diff diff-contentalign-left diff-editfont-monospace&quot; data-mw=&quot;interface&quot;&gt;
				&lt;col class=&quot;diff-marker&quot; /&gt;
				&lt;col class=&quot;diff-content&quot; /&gt;
				&lt;col class=&quot;diff-marker&quot; /&gt;
				&lt;col class=&quot;diff-content&quot; /&gt;
				&lt;tr class=&quot;diff-title&quot; lang=&quot;en&quot;&gt;
				&lt;td colspan=&quot;2&quot; style=&quot;background-color: #fff; color: #202122; text-align: center;&quot;&gt;← Older revision&lt;/td&gt;
				&lt;td colspan=&quot;2&quot; style=&quot;background-color: #fff; color: #202122; text-align: center;&quot;&gt;Revision as of 23:32, 25 March 2026&lt;/td&gt;
				&lt;/tr&gt;&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot; id=&quot;mw-diff-left-l81&quot; &gt;Line 81:&lt;/td&gt;
&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot;&gt;Line 81:&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;==Model yang relevan==&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;==Model yang relevan==&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;== A. Model grafik / gambar&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;== A. Model grafik / gambar&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;==&lt;/ins&gt;&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;=== Pilihan terbaik untuk mulai: '''SDXL Base 1.0'''&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;=== Pilihan terbaik untuk mulai: '''SDXL Base 1.0'''&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;==&lt;/ins&gt;&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;Ini pilihan paling aman untuk GPU 12–16 GB. Cocok untuk text-to-image umum, stabil, dan sangat didukung workflow ComfyUI. ComfyUI “first generation” juga memang mengajarkan pola dasar image generation dengan memilih model di node loader. ([ComfyUI Documentation][7])&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;Ini pilihan paling aman untuk GPU 12–16 GB. Cocok untuk text-to-image umum, stabil, dan sangat didukung workflow ComfyUI. ComfyUI “first generation” juga memang mengajarkan pola dasar image generation dengan memilih model di node loader. ([ComfyUI Documentation][7])&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot; id=&quot;mw-diff-left-l98&quot; &gt;Line 98:&lt;/td&gt;
&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot;&gt;Line 98:&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  &amp;quot;https://huggingface.co/stabilityai/stable-diffusion-xl-refiner-1.0/resolve/main/sd_xl_refiner_1.0.safetensors?download=true&amp;quot;&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;  &amp;quot;https://huggingface.co/stabilityai/stable-diffusion-xl-refiner-1.0/resolve/main/sd_xl_refiner_1.0.safetensors?download=true&amp;quot;&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;=== Kapan pakai '''FLUX.1-dev'''&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;=== Kapan pakai '''FLUX.1-dev'''&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;===&lt;/ins&gt;&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;FLUX.1-dev resmi tersedia untuk ComfyUI dan merupakan model 12B untuk text-to-image, tetapi akses file-nya mensyaratkan persetujuan lisensi di Hugging Face, dan lisensinya '''non-commercial'''. Di GPU 16 GB dia bisa dicoba, tetapi untuk pengalaman awal saya tetap menyarankan '''SDXL dulu''' karena lebih ringan dan lebih sederhana. ([Hugging Face][8])&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;FLUX.1-dev resmi tersedia untuk ComfyUI dan merupakan model 12B untuk text-to-image, tetapi akses file-nya mensyaratkan persetujuan lisensi di Hugging Face, dan lisensinya '''non-commercial'''. Di GPU 16 GB dia bisa dicoba, tetapi untuk pengalaman awal saya tetap menyarankan '''SDXL dulu''' karena lebih ringan dan lebih sederhana. ([Hugging Face][8])&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot; id=&quot;mw-diff-left-l109&quot; &gt;Line 109:&lt;/td&gt;
&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot;&gt;Line 109:&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;4. letakkan di folder model yang diminta workflow/nodenya. ([Hugging Face][8])&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;4. letakkan di folder model yang diminta workflow/nodenya. ([Hugging Face][8])&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;=== Tambahan yang relevan setelah SDXL jalan&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;=== Tambahan yang relevan setelah SDXL jalan&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;===&lt;/ins&gt;&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;* '''LoRA''' untuk style/karakter → taruh di `models/loras/`&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;* '''LoRA''' untuk style/karakter → taruh di `models/loras/`&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot; id=&quot;mw-diff-left-l116&quot; &gt;Line 116:&lt;/td&gt;
&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot;&gt;Line 116:&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;== B. Model audio&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;== B. Model audio&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;==&lt;/ins&gt;&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;=== Pilihan paling relevan di ComfyUI: '''Stable Audio Open 1.0'''&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;=== Pilihan paling relevan di ComfyUI: '''Stable Audio Open 1.0'''&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;===&lt;/ins&gt;&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;ComfyUI punya contoh audio resmi untuk '''Stable Audio Open 1.0'''. Yang dibutuhkan:&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;ComfyUI punya contoh audio resmi untuk '''Stable Audio Open 1.0'''. Yang dibutuhkan:&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot; id=&quot;mw-diff-left-l132&quot; &gt;Line 132:&lt;/td&gt;
&lt;td colspan=&quot;2&quot; class=&quot;diff-lineno&quot;&gt;Line 132:&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;Workflow audio resminya bisa di-load dari contoh audio ComfyUI. ([Comfy Anonymous][3])&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;Workflow audio resminya bisa di-load dari contoh audio ComfyUI. ([Comfy Anonymous][3])&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;=== Kalau butuh TTS / voice&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt;+&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #a3d3ff; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;=== Kalau butuh TTS / voice&lt;ins class=&quot;diffchange diffchange-inline&quot;&gt;===&lt;/ins&gt;&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;Untuk TTS, ComfyUI bisa diperluas dengan custom node seperti '''TTS-Audio-Suite''', yang menggabungkan beberapa engine TTS dan voice conversion. Namun ini lebih cocok setelah instalasi dasar Anda stabil, karena custom node berarti tambahan dependency. Secara resmi instal custom node di ComfyUI memang selalu dua langkah: clone ke `custom_nodes` lalu install dependensi Python-nya. ([GitHub][9])&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;Untuk TTS, ComfyUI bisa diperluas dengan custom node seperti '''TTS-Audio-Suite''', yang menggabungkan beberapa engine TTS dan voice conversion. Namun ini lebih cocok setelah instalasi dasar Anda stabil, karena custom node berarti tambahan dependency. Secara resmi instal custom node di ComfyUI memang selalu dua langkah: clone ke `custom_nodes` lalu install dependensi Python-nya. ([GitHub][9])&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt;−&lt;/td&gt;&lt;td style=&quot;color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #ffe49c; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;&lt;del style=&quot;font-weight: bold; text-decoration: none;&quot;&gt;&lt;/del&gt;&lt;/div&gt;&lt;/td&gt;&lt;td colspan=&quot;2&quot;&gt; &lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;tr&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;==Urutan instalasi yang saya sarankan==&lt;/div&gt;&lt;/td&gt;&lt;td class='diff-marker'&gt; &lt;/td&gt;&lt;td style=&quot;background-color: #f8f9fa; color: #202122; font-size: 88%; border-style: solid; border-width: 1px 1px 1px 4px; border-radius: 0.33em; border-color: #eaecf0; vertical-align: top; white-space: pre-wrap;&quot;&gt;&lt;div&gt;==Urutan instalasi yang saya sarankan==&lt;/div&gt;&lt;/td&gt;&lt;/tr&gt;
&lt;/table&gt;</summary>
		<author><name>Onnowpurbo</name></author>
	</entry>
	<entry>
		<id>https://onnocenter.or.id/wiki/index.php?title=ComfyUI:_Instalasi_venv_GPU&amp;diff=73520&amp;oldid=prev</id>
		<title>Onnowpurbo: Created page with &quot;Berikut '''panduan lengkap tapi singkat''' untuk '''Ubuntu 24.04 + GPU kelas RTX 5070/5070 Ti''' menjalankan '''ComfyUI''' dengan model '''grafik/gambar''' dan '''audio'''.  C...&quot;</title>
		<link rel="alternate" type="text/html" href="https://onnocenter.or.id/wiki/index.php?title=ComfyUI:_Instalasi_venv_GPU&amp;diff=73520&amp;oldid=prev"/>
		<updated>2026-03-25T23:26:35Z</updated>

		<summary type="html">&lt;p&gt;Created page with &amp;quot;Berikut &amp;#039;&amp;#039;&amp;#039;panduan lengkap tapi singkat&amp;#039;&amp;#039;&amp;#039; untuk &amp;#039;&amp;#039;&amp;#039;Ubuntu 24.04 + GPU kelas RTX 5070/5070 Ti&amp;#039;&amp;#039;&amp;#039; menjalankan &amp;#039;&amp;#039;&amp;#039;ComfyUI&amp;#039;&amp;#039;&amp;#039; dengan model &amp;#039;&amp;#039;&amp;#039;grafik/gambar&amp;#039;&amp;#039;&amp;#039; dan &amp;#039;&amp;#039;&amp;#039;audio&amp;#039;&amp;#039;&amp;#039;.  C...&amp;quot;&lt;/p&gt;
&lt;p&gt;&lt;b&gt;New page&lt;/b&gt;&lt;/p&gt;&lt;div&gt;Berikut '''panduan lengkap tapi singkat''' untuk '''Ubuntu 24.04 + GPU kelas RTX 5070/5070 Ti''' menjalankan '''ComfyUI''' dengan model '''grafik/gambar''' dan '''audio'''.&lt;br /&gt;
&lt;br /&gt;
Catatan penting dulu: untuk desktop, '''RTX 5070 = 12 GB''', sedangkan '''RTX 5070 Ti = 16 GB'''. Jadi kalau Anda menulis '''5070 16 GB''', besar kemungkinan yang dimaksud adalah '''5070 Ti''', atau GPU laptop kelas 16 GB. Saya buat panduan ini aman untuk '''VRAM 12–16 GB'''. PyTorch 2.7 juga sudah menambahkan dukungan untuk arsitektur '''NVIDIA Blackwell''' dan wheel '''CUDA 12.8''', jadi jalur yang paling masuk akal untuk seri 50 adalah pakai PyTorch baru, bukan stack lama. ([NVIDIA][1])&lt;br /&gt;
&lt;br /&gt;
== Rekomendasi singkat&lt;br /&gt;
&lt;br /&gt;
Untuk '''gambar''', mulai dari '''SDXL''' dulu karena paling stabil untuk 12–16 GB VRAM. ComfyUI sendiri memang dipasang lewat alur '''venv → clone repo → install dependencies → run''', dan model diletakkan di folder `models/...` sesuai tipenya. ([ComfyUI Documentation][2])&lt;br /&gt;
&lt;br /&gt;
Untuk '''audio''', ComfyUI memang punya jalur resmi contoh '''Stable Audio Open 1.0''', dengan file checkpoint di `models/checkpoints/` dan text encoder di `models/text_encoders/`. Jadi audio bisa, tetapi untuk awal saya sarankan pasang setelah image workflow sudah sehat. ([Comfy Anonymous][3])&lt;br /&gt;
&lt;br /&gt;
== 1) Siapkan Ubuntu 24.04==&lt;br /&gt;
&lt;br /&gt;
Update sistem:&lt;br /&gt;
&lt;br /&gt;
 sudo apt update &amp;amp;&amp;amp; sudo apt upgrade -y&lt;br /&gt;
 sudo apt install -y git python3 python3-venv python3-pip build-essential&lt;br /&gt;
&lt;br /&gt;
Untuk driver NVIDIA, paling aman di Ubuntu 24.04 adalah pasang '''driver proprietary''' dari '''Software &amp;amp; Updates → Additional Drivers''', lalu reboot. Setelah reboot, cek:&lt;br /&gt;
&lt;br /&gt;
 nvidia-smi&lt;br /&gt;
&lt;br /&gt;
Kalau `nvidia-smi` keluar normal, lanjut.&lt;br /&gt;
&lt;br /&gt;
== 2) Buat virtual environment dan install ComfyUI==&lt;br /&gt;
&lt;br /&gt;
 cd ~&lt;br /&gt;
 mkdir -p AI&lt;br /&gt;
 cd AI&lt;br /&gt;
 &lt;br /&gt;
 python3 -m venv comfyui-venv&lt;br /&gt;
 source comfyui-venv/bin/activate&lt;br /&gt;
 &lt;br /&gt;
 pip install --upgrade pip setuptools wheel&lt;br /&gt;
 git clone https://github.com/comfy-org/ComfyUI.git&lt;br /&gt;
 cd ComfyUI&lt;br /&gt;
&lt;br /&gt;
Alur ini sesuai dokumentasi resmi ComfyUI manual install untuk Linux. ([ComfyUI Documentation][2])&lt;br /&gt;
&lt;br /&gt;
== 3) Install PyTorch yang cocok untuk RTX 5070 / 5070 Ti==&lt;br /&gt;
&lt;br /&gt;
Karena PyTorch 2.7 menambahkan dukungan Blackwell dan wheel CUDA 12.8, ini yang saya sarankan:&lt;br /&gt;
&lt;br /&gt;
 pip install torch==2.7.0 torchvision torchaudio --index-url https://download.pytorch.org/whl/cu128&lt;br /&gt;
&lt;br /&gt;
Lalu install dependensi ComfyUI:&lt;br /&gt;
&lt;br /&gt;
 pip install -r requirements.txt&lt;br /&gt;
&lt;br /&gt;
PyTorch sendiri menyebut 2.7 membawa dukungan Blackwell dan prebuilt wheel CUDA 12.8; ComfyUI di sisi lain memang meminta Anda memasang dependency GPU dulu baru `requirements.txt`. ([PyTorch][4])&lt;br /&gt;
&lt;br /&gt;
== 4) Jalankan ComfyUI==&lt;br /&gt;
&lt;br /&gt;
 python main.py&lt;br /&gt;
&lt;br /&gt;
Kalau ingin bisa diakses dari browser lain di jaringan:&lt;br /&gt;
 &lt;br /&gt;
 python main.py --listen 0.0.0.0&lt;br /&gt;
&lt;br /&gt;
Secara resmi ComfyUI menjalankan aplikasi dengan `python main.py`. ([ComfyUI Documentation][2])&lt;br /&gt;
&lt;br /&gt;
== 5) Aktifkan Manager==&lt;br /&gt;
&lt;br /&gt;
Sekarang Manager baru sudah built-in di core, tetapi harus diaktifkan dan dependency-nya diinstall:&lt;br /&gt;
&lt;br /&gt;
 pip install -r manager_requirements.txt&lt;br /&gt;
 python main.py --enable-manager&lt;br /&gt;
&lt;br /&gt;
Dokumentasi resmi ComfyUI Manager memang menyebut untuk manual install: aktifkan venv, install `manager_requirements.txt`, lalu jalankan dengan `--enable-manager`. ([ComfyUI Documentation][5])&lt;br /&gt;
&lt;br /&gt;
== 6) Folder model yang perlu ada==&lt;br /&gt;
&lt;br /&gt;
 mkdir -p ~/AI/ComfyUI/models/checkpoints&lt;br /&gt;
 mkdir -p ~/AI/ComfyUI/models/vae&lt;br /&gt;
 mkdir -p ~/AI/ComfyUI/models/loras&lt;br /&gt;
 mkdir -p ~/AI/ComfyUI/models/controlnet&lt;br /&gt;
 mkdir -p ~/AI/ComfyUI/models/text_encoders&lt;br /&gt;
&lt;br /&gt;
ComfyUI menjelaskan model perlu diletakkan di subfolder sesuai tipe, lalu dipilih lewat node loader yang sesuai. ([ComfyUI Documentation][6])&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
==Model yang relevan==&lt;br /&gt;
&lt;br /&gt;
== A. Model grafik / gambar&lt;br /&gt;
&lt;br /&gt;
=== Pilihan terbaik untuk mulai: '''SDXL Base 1.0'''&lt;br /&gt;
&lt;br /&gt;
Ini pilihan paling aman untuk GPU 12–16 GB. Cocok untuk text-to-image umum, stabil, dan sangat didukung workflow ComfyUI. ComfyUI “first generation” juga memang mengajarkan pola dasar image generation dengan memilih model di node loader. ([ComfyUI Documentation][7])&lt;br /&gt;
&lt;br /&gt;
Download:&lt;br /&gt;
&lt;br /&gt;
 cd ~/AI/ComfyUI/models/checkpoints&lt;br /&gt;
 wget -O sd_xl_base_1.0.safetensors \&lt;br /&gt;
 &amp;quot;https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0/resolve/main/sd_xl_base_1.0.safetensors?download=true&amp;quot;&lt;br /&gt;
&lt;br /&gt;
Opsional refiner:&lt;br /&gt;
&lt;br /&gt;
 wget -O sd_xl_refiner_1.0.safetensors \&lt;br /&gt;
 &amp;quot;https://huggingface.co/stabilityai/stable-diffusion-xl-refiner-1.0/resolve/main/sd_xl_refiner_1.0.safetensors?download=true&amp;quot;&lt;br /&gt;
&lt;br /&gt;
=== Kapan pakai '''FLUX.1-dev'''&lt;br /&gt;
&lt;br /&gt;
FLUX.1-dev resmi tersedia untuk ComfyUI dan merupakan model 12B untuk text-to-image, tetapi akses file-nya mensyaratkan persetujuan lisensi di Hugging Face, dan lisensinya '''non-commercial'''. Di GPU 16 GB dia bisa dicoba, tetapi untuk pengalaman awal saya tetap menyarankan '''SDXL dulu''' karena lebih ringan dan lebih sederhana. ([Hugging Face][8])&lt;br /&gt;
&lt;br /&gt;
Kalau Anda tetap ingin FLUX:&lt;br /&gt;
&lt;br /&gt;
1. buka halaman model di Hugging Face,&lt;br /&gt;
2. setujui syarat akses,&lt;br /&gt;
3. unduh file yang direkomendasikan untuk ComfyUI,&lt;br /&gt;
4. letakkan di folder model yang diminta workflow/nodenya. ([Hugging Face][8])&lt;br /&gt;
&lt;br /&gt;
=== Tambahan yang relevan setelah SDXL jalan&lt;br /&gt;
&lt;br /&gt;
* '''LoRA''' untuk style/karakter → taruh di `models/loras/`&lt;br /&gt;
* '''ControlNet''' untuk pose/lineart/depth → taruh di `models/controlnet/`&lt;br /&gt;
  ComfyUI memang memisahkan model berdasarkan tipe loader seperti `Load Checkpoint`, `Load LoRA`, dan `Load VAE`. ([ComfyUI Documentation][6])&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
== B. Model audio&lt;br /&gt;
&lt;br /&gt;
=== Pilihan paling relevan di ComfyUI: '''Stable Audio Open 1.0'''&lt;br /&gt;
&lt;br /&gt;
ComfyUI punya contoh audio resmi untuk '''Stable Audio Open 1.0'''. Yang dibutuhkan:&lt;br /&gt;
&lt;br /&gt;
* `stable_audio_open_1.0.safetensors` → `models/checkpoints/`&lt;br /&gt;
* `t5_base.safetensors` → `models/text_encoders/` ([Comfy Anonymous][3])&lt;br /&gt;
&lt;br /&gt;
Struktur akhirnya:&lt;br /&gt;
&lt;br /&gt;
 ~/AI/ComfyUI/models/checkpoints/stable_audio_open_1.0.safetensors&lt;br /&gt;
 ~/AI/ComfyUI/models/text_encoders/t5_base.safetensors&lt;br /&gt;
&lt;br /&gt;
Workflow audio resminya bisa di-load dari contoh audio ComfyUI. ([Comfy Anonymous][3])&lt;br /&gt;
&lt;br /&gt;
=== Kalau butuh TTS / voice&lt;br /&gt;
&lt;br /&gt;
Untuk TTS, ComfyUI bisa diperluas dengan custom node seperti '''TTS-Audio-Suite''', yang menggabungkan beberapa engine TTS dan voice conversion. Namun ini lebih cocok setelah instalasi dasar Anda stabil, karena custom node berarti tambahan dependency. Secara resmi instal custom node di ComfyUI memang selalu dua langkah: clone ke `custom_nodes` lalu install dependensi Python-nya. ([GitHub][9])&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
==Urutan instalasi yang saya sarankan==&lt;br /&gt;
&lt;br /&gt;
1. Pasang driver NVIDIA, pastikan `nvidia-smi` normal.&lt;br /&gt;
2. Install '''ComfyUI + PyTorch 2.7 cu128'''. ([PyTorch][4])&lt;br /&gt;
3. Aktifkan '''Manager'''. ([ComfyUI Documentation][5])&lt;br /&gt;
4. Download '''SDXL Base 1.0''' dulu.&lt;br /&gt;
5. Tes generate gambar pertama.&lt;br /&gt;
6. Baru tambahkan '''Stable Audio Open 1.0''' untuk audio. ([Comfy Anonymous][3])&lt;br /&gt;
7. Setelah semuanya stabil, baru eksperimen '''FLUX''', LoRA, dan ControlNet. ([Hugging Face][8])&lt;br /&gt;
&lt;br /&gt;
==Perintah ringkas yang paling praktis==&lt;br /&gt;
&lt;br /&gt;
 sudo apt update &amp;amp;&amp;amp; sudo apt upgrade -y&lt;br /&gt;
 sudo apt install -y git python3 python3-venv python3-pip build-essential&lt;br /&gt;
&lt;br /&gt;
 cd ~&lt;br /&gt;
 mkdir -p AI&lt;br /&gt;
 cd AI&lt;br /&gt;
 python3 -m venv comfyui-venv&lt;br /&gt;
 source comfyui-venv/bin/activate&lt;br /&gt;
 &lt;br /&gt;
 pip install --upgrade pip setuptools wheel&lt;br /&gt;
 git clone https://github.com/comfy-org/ComfyUI.git&lt;br /&gt;
 cd ComfyUI&lt;br /&gt;
 &lt;br /&gt;
 pip install torch==2.7.0 torchvision torchaudio --index-url https://download.pytorch.org/whl/cu128&lt;br /&gt;
 pip install -r requirements.txt&lt;br /&gt;
 pip install -r manager_requirements.txt&lt;br /&gt;
 &lt;br /&gt;
 mkdir -p models/checkpoints models/vae models/loras models/controlnet models/text_encoders&lt;br /&gt;
 &lt;br /&gt;
 cd models/checkpoints&lt;br /&gt;
 wget -O sd_xl_base_1.0.safetensors \&lt;br /&gt;
 &amp;quot;https://huggingface.co/stabilityai/stable-diffusion-xl-base-1.0/resolve/main/sd_xl_base_1.0.safetensors?download=true&amp;quot;&lt;br /&gt;
 &lt;br /&gt;
 cd ~/AI/ComfyUI&lt;br /&gt;
 python main.py --enable-manager&lt;br /&gt;
&lt;br /&gt;
==Setting awal yang aman untuk GPU 12–16 GB==&lt;br /&gt;
&lt;br /&gt;
Untuk '''SDXL''':&lt;br /&gt;
&lt;br /&gt;
* resolusi awal: '''1024x1024'''&lt;br /&gt;
* batch: '''1'''&lt;br /&gt;
* steps: '''20–30'''&lt;br /&gt;
* sampler: mulai dari default workflow dulu&lt;br /&gt;
&lt;br /&gt;
Untuk '''Stable Audio Open''':&lt;br /&gt;
&lt;br /&gt;
* mulai dari workflow contoh resmi&lt;br /&gt;
* jangan pasang banyak custom node dulu&lt;br /&gt;
&lt;br /&gt;
Untuk '''FLUX''':&lt;br /&gt;
&lt;br /&gt;
* anggap sebagai tahap kedua, bukan model pertama&lt;br /&gt;
&lt;br /&gt;
Yang paling penting: untuk kelas '''5070/5070 Ti''', bottleneck Anda bukan hanya compute, tetapi '''VRAM dan kompleksitas workflow'''. Jadi '''SDXL lebih relevan''' untuk produktivitas awal daripada langsung memaksa '''FLUX'''. ([Hugging Face][8])&lt;br /&gt;
&lt;br /&gt;
Saya bisa lanjutkan dengan '''panduan satu file bash otomatis''' untuk:&lt;br /&gt;
'''Ubuntu 24.04 + NVIDIA + ComfyUI + SDXL + Stable Audio Open''', supaya tinggal copy-paste.&lt;br /&gt;
&lt;br /&gt;
 [1]: https://www.nvidia.com/en-us/geforce/graphics-cards/50-series/rtx-5070-family/ &amp;quot;GeForce RTX 5070 Family Graphics Cards | NVIDIA&amp;quot;&lt;br /&gt;
 [2]: https://docs.comfy.org/installation/manual_install &amp;quot;How to install ComfyUI manually in different systems - ComfyUI&amp;quot;&lt;br /&gt;
 [3]: https://comfyanonymous.github.io/ComfyUI_examples/audio/ &amp;quot;Audio Examples | ComfyUI_examples&amp;quot;&lt;br /&gt;
 [4]: https://pytorch.org/blog/pytorch-2-7/ &amp;quot;PyTorch 2.7 Release – PyTorch&amp;quot;&lt;br /&gt;
 [5]: https://docs.comfy.org/manager/install &amp;quot;ComfyUI-Manager Installation - ComfyUI&amp;quot;&lt;br /&gt;
 [6]: https://docs.comfy.org/development/core-concepts/models?utm_source=chatgpt.com &amp;quot;Models&amp;quot;&lt;br /&gt;
 [7]: https://docs.comfy.org/get_started/first_generation?utm_source=chatgpt.com &amp;quot;Getting Started with AI Image Generation&amp;quot;&lt;br /&gt;
 [8]: https://huggingface.co/black-forest-labs/FLUX.1-dev &amp;quot;black-forest-labs/FLUX.1-dev · Hugging Face&amp;quot;&lt;br /&gt;
 [9]: https://github.com/diodiogod/TTS-Audio-Suite?utm_source=chatgpt.com &amp;quot;diodiogod/TTS-Audio-Suite: A ComfyUI custom node ...&amp;quot;&lt;br /&gt;
&lt;br /&gt;
&lt;br /&gt;
==Pranala Menarik==&lt;br /&gt;
&lt;br /&gt;
* [[LLM]]&lt;/div&gt;</summary>
		<author><name>Onnowpurbo</name></author>
	</entry>
</feed>