<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:wfw="http://wellformedweb.org/CommentAPI/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	xmlns:atom="http://www.w3.org/2005/Atom"
	xmlns:sy="http://purl.org/rss/1.0/modules/syndication/"
	xmlns:slash="http://purl.org/rss/1.0/modules/slash/"
	>

<channel>
	<title>H2-Blog &#187; ComfyUI</title>
	<atom:link href="https://www.h2-blog.com/category/technologie/comfyui/feed/" rel="self" type="application/rss+xml" />
	<link>https://www.h2-blog.com</link>
	<description>Un blog mêlant actualités et divertissement</description>
	<lastBuildDate>Thu, 05 Nov 2026 20:46:12 +0000</lastBuildDate>
	<language>fr-FR</language>
	<sy:updatePeriod>hourly</sy:updatePeriod>
	<sy:updateFrequency>1</sy:updateFrequency>
	<generator>http://wordpress.org/?v=4.2.2</generator>
	<item>
		<title>FLUX 3 KLEIN vs KREA 2 vs Qwen 2.1 vs Z image turbo</title>
		<link>https://www.h2-blog.com/2026/11/05/flux-3-klein-vs-krea-2-vs-qwen-2-1-vs-z-image-turbo/</link>
		<comments>https://www.h2-blog.com/2026/11/05/flux-3-klein-vs-krea-2-vs-qwen-2-1-vs-z-image-turbo/#comments</comments>
		<pubDate>Thu, 05 Nov 2026 20:44:52 +0000</pubDate>
		<dc:creator><![CDATA[Guillaume R]]></dc:creator>
				<category><![CDATA[ComfyUI]]></category>

		<guid isPermaLink="false">http://www.h2-blog.com/?p=11911</guid>
		<description><![CDATA[<p>J’ai créé deux flux de travail pour comparer FLUX 3, KREA 2, Z-Image Turbo et Qwen 2.1. Ils utilisent tous les version turbo INT8 ce qui fait que cela nécessite seulement 25g de RAM et cela prend environ 27 secondes avec une 5070Ti. Ces deux flux de travail sont disponibles sur Civitai et Github. MANGA [&#8230;]</p><p>Cet article <a href="https://www.h2-blog.com/2026/11/05/flux-3-klein-vs-krea-2-vs-qwen-2-1-vs-z-image-turbo/">FLUX 3 KLEIN vs KREA 2 vs Qwen 2.1 vs Z image turbo</a> provient du site <a href="https://www.h2-blog.com">H2-Blog</a>.</p>]]></description>
				<content:encoded><![CDATA[<p>J’ai créé deux flux de travail pour comparer FLUX 3, KREA 2, Z-Image Turbo et Qwen 2.1. Ils utilisent tous les version turbo INT8 ce qui fait que cela nécessite seulement 25g de RAM et cela prend environ 27 secondes avec une 5070Ti.</p>
<p>Ces deux flux de travail sont disponibles sur <a href="https://civitai.com/user/def813/models?tag=f3vsk2vszvsq21" target="_blank">Civitai</a> et <a href="https://github.com/H2logs/FLUX-3-vs-KREA-2-vs-Z-image-turbo-vs-Qwen-2.1" target="_blank">Github</a>.</p>
<h2>MANGA</h2>
<p><a href="http://www.h2-blog.com/wp-content/uploads/2026/11/flux-3-vs-krea-2-vs-z-image-turbo_00001_.png"><img class="aligncenter size-large wp-image-11913" src="http://www.h2-blog.com/wp-content/uploads/2026/11/flux-3-vs-krea-2-vs-z-image-turbo_00001_-650x650.png" alt="flux-3-vs-krea-2-vs-z-image-turbo_00001_" width="650" height="650" /></a></p>
<div>
<div><div class="wp-custom-code-container"><pre><code>A medium close-up side profile illustration of a young person with pale skin and sharp, angular facial features, gazing upward and to the left. The subject has short, tousled dark blue hair that blends seamlessly into an abstract, fluid background of swirling ocean waves, deep blues, teals, and lighter sky blues. The hair is animated with dynamic, curling strands that extend outward like liquid smoke or water splashes. Scattered throughout the composition are small, glowing yellow four-pointed stars and dots, resembling fireflies or distant stars, adding a magical, ethereal quality. The subject&rsquo;s visible eye is large, almond-shaped, and a striking light blue, detailed with a white catchlight and thin black eyelashes. The mouth is slightly parted, revealing a hint of the tongue, suggesting a sense of wonder or breathless awe. The subject is wearing a simple hooded garment in solid shades of medium and dark blue, which frames the head from the top and left side. The hood&rsquo;s edge is defined by a thick black outline. The background is split: the left side is dominated by the solid blue tones of the hood and shadow, while the right side transitions into the beige/cream-colored canvas where the fluid, abstract water-like patterns and stars are</code></pre><button class="wp-custom-copy-btn" onclick="copyCodeSnippet(this)" aria-label="Copier le code"><svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><rect x="9" y="9" width="13" height="13" rx="2" ry="2"></rect><path d="M5 15H4a2 2 0 0 1-2-2V4a2 2 0 0 1 2-2h9a2 2 0 0 1 2 2v1"></path></svg></button></div></div>
<div>
<h2 class="header-anchor-post">ART</h2>
<p><a href="http://www.h2-blog.com/wp-content/uploads/2026/11/flux-3-vs-krea-2-vs-z-image-turbo_00002_.png"><img class="aligncenter size-large wp-image-11918" src="http://www.h2-blog.com/wp-content/uploads/2026/11/flux-3-vs-krea-2-vs-z-image-turbo_00002_-650x650.png" alt="flux-3-vs-krea-2-vs-z-image-turbo_00002_" width="650" height="650" /></a></p>
<div class="wp-custom-code-container"><pre><code>A vibrant, stylized illustration of a fantastical character resembling a witch or fairy, set against a whimsical, dreamlike background. The subject is a slender figure with a smooth, yellow complexion and a simple, joyful expression featuring closed eyes, rosy cheeks, and a gentle smile. They wear a wide-brimmed, conical purple hat adorned with white geometric patterns (triangles and dots) along the brim. Their attire consists of a short-sleeved, purple tunic with black-and-white striped cuffs and collar, decorated with small white floral motifs.The most striking feature is the expansive, flowing garment that acts as wings or a skirt. It is primarily a vivid magenta-pink color, composed of large, layered panels. These panels are patterned with white circles of varying sizes and pink cross-hatching lines. Extending from the lower section of this garment are three distinct, crescent-shaped elements resembling fish tails or fins. These are bright yellow with intricate blue and green tribal-style patterns, including leafy vines and geometric shapes.The background is a dynamic sky in shades of soft pink and lavender. It features swirling yellow and teal lines suggesting wind or magic, dotted with small yellow specks. There are stylized white clouds with black</code></pre><button class="wp-custom-copy-btn" onclick="copyCodeSnippet(this)" aria-label="Copier le code"><svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><rect x="9" y="9" width="13" height="13" rx="2" ry="2"></rect><path d="M5 15H4a2 2 0 0 1-2-2V4a2 2 0 0 1 2-2h9a2 2 0 0 1 2 2v1"></path></svg></button></div>
<h2 class="header-anchor-post">PORTRAIT</h2>
<p><a href="http://www.h2-blog.com/wp-content/uploads/2026/11/flux-3-vs-krea-2-vs-z-image-turbo_00003_.png"><img class="aligncenter size-large wp-image-11919" src="http://www.h2-blog.com/wp-content/uploads/2026/11/flux-3-vs-krea-2-vs-z-image-turbo_00003_-650x650.png" alt="flux-3-vs-krea-2-vs-z-image-turbo_00003_" width="650" height="650" /></a><div class="wp-custom-code-container"><pre><code>Portrait of a young woman with wavy, blonde-highlighted hair, blue eyes, and light makeup featuring shimmery eyeshadow and nude lipstick. She is gazing upward and slightly to the side with a soft, contemplative expression, her head tilted back slightly. Her lips are parted just enough to reveal the top teeth. She has fair skin with natural freckles across the nose and cheeks. She is wearing a light-colored, possibly beige or cream, top with thin straps or a V-neckline. The background is a soft, out-of-focus gradient of pale blue and white, creating a dreamy, airy atmosphere. The lighting is soft and diffused, highlighting the contours of her face and adding a gentle glow to her skin. The camera angle is a close-up, slightly low-angle shot looking up at her face. The composition focuses tightly on her head and shoulders, with her hair framing her face on both sides. The overall aesthetic is ethereal, soft, and feminine, reminiscent of high-fashion editorial photography or a beauty advertisement.</code></pre><button class="wp-custom-copy-btn" onclick="copyCodeSnippet(this)" aria-label="Copier le code"><svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><rect x="9" y="9" width="13" height="13" rx="2" ry="2"></rect><path d="M5 15H4a2 2 0 0 1-2-2V4a2 2 0 0 1 2-2h9a2 2 0 0 1 2 2v1"></path></svg></button></div></p>
<h2 class="header-anchor-post">CHAT</h2>
<p><a href="http://www.h2-blog.com/wp-content/uploads/2026/11/flux-3-vs-krea-2-vs-z-image-turbo_00004_.png"><img class="aligncenter size-large wp-image-11921" src="http://www.h2-blog.com/wp-content/uploads/2026/11/flux-3-vs-krea-2-vs-z-image-turbo_00004_-650x650.png" alt="flux-3-vs-krea-2-vs-z-image-turbo_00004_" width="650" height="650" /></a><div class="wp-custom-code-container"><pre><code>A fluffy, long-haired cat with a cream and brown tabby coat sits upright on a paved stone surface, facing forward. The cat has a white muzzle and chest, with brown and black fur on its head and back. Its eyes are greenish-yellow and alert, gazing directly at the viewer. The cat&rsquo;s ears are perked up, and its tail is curled around its side. It is not wearing any clothing.The background features a modern building with large glass windows and balconies. A red flag with a white crescent moon and star (the Turkish flag) hangs vertically in the background, slightly out of focus. There are metal railings and street lamps visible along the building facade. The lighting is warm and golden, suggesting late afternoon or early evening sunlight hitting the cat from the side, creating soft highlights on its fur. The camera angle is low, almost eye-level with the seated cat, emphasizing its presence. The composition places the cat centrally in the foreground, with the building and flag providing depth in the background. The color palette consists of warm earth tones (browns, creams, tans) contrasted with the bright red of the flag and the cool blue-grey of the building&rsquo;s glass. The overall aesthetic is realistic, detailed,</code></pre><button class="wp-custom-copy-btn" onclick="copyCodeSnippet(this)" aria-label="Copier le code"><svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><rect x="9" y="9" width="13" height="13" rx="2" ry="2"></rect><path d="M5 15H4a2 2 0 0 1-2-2V4a2 2 0 0 1 2-2h9a2 2 0 0 1 2 2v1"></path></svg></button></div></p>
<h2 class="header-anchor-post">MOTOCROSS</h2>
<p><a href="http://www.h2-blog.com/wp-content/uploads/2026/11/flux-3-vs-krea-2-vs-z-image-turbo_00005_.png"><img class="aligncenter size-large wp-image-11923" src="http://www.h2-blog.com/wp-content/uploads/2026/11/flux-3-vs-krea-2-vs-z-image-turbo_00005_-650x650.png" alt="flux-3-vs-krea-2-vs-z-image-turbo_00005_" width="650" height="650" /></a><div class="wp-custom-code-container"><pre><code>A full-body, eye-level shot of a motocross rider in dynamic motion on a dirt track, captured with sharp focus on the subject and a shallow depth of field blurring the background. The rider is positioned on the right side of the frame, leaning forward aggressively into a turn or straightaway, knees bent and gripping handlebars equipped with black perforated handguards. They wear a full-face helmet featuring a high-contrast design of matte black and neon yellow geometric patterns, paired with reflective blue-tinted goggles. Their attire consists of a long-sleeved black and white jersey, black protective pants with white diagonal stripes on the thighs, and heavy-duty black motocross boots. The dirt bike is a black dual-sport model with knobby tires kicking up a spray of brown dust and dirt particles at the rear wheel, indicating speed and traction.The background features a hazy, overcast sky in pale grey-blue tones. To the left, out of focus, stands a large air traffic control tower topped with a prominent white spherical radar dome, suggesting an airport setting adjacent to the off-road course. Two weathered wooden posts stand vertically on either side of the path. The ground is composed of dry, brownish grass and packed earth. The lighting is soft and</code></pre><button class="wp-custom-copy-btn" onclick="copyCodeSnippet(this)" aria-label="Copier le code"><svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><rect x="9" y="9" width="13" height="13" rx="2" ry="2"></rect><path d="M5 15H4a2 2 0 0 1-2-2V4a2 2 0 0 1 2-2h9a2 2 0 0 1 2 2v1"></path></svg></button></div></p>
<h2 class="header-anchor-post">DANSE</h2>
<p><a href="http://www.h2-blog.com/wp-content/uploads/2026/11/flux-3-vs-krea-2-vs-z-image-turbo_00006_.png"><img class="aligncenter size-large wp-image-11925" src="http://www.h2-blog.com/wp-content/uploads/2026/11/flux-3-vs-krea-2-vs-z-image-turbo_00006_-650x650.png" alt="flux-3-vs-krea-2-vs-z-image-turbo_00006_" width="650" height="650" /></a><div class="wp-custom-code-container"><pre><code>A low-angle, full-body shot of a young woman with light skin and long, thin dreadlocks pulled back, wearing large over-ear headphones. She is dressed in a black sports bra with white trim and matching black cargo pants. Her left arm is raised, hand near her head adjusting the headphones, while her right hand rests on a dark object to her side. Visible tattoos adorn her left forearm. She wears multiple bracelets on her right wrist and a necklace. Her expression is serious and focused, eyes looking downward. The background features dark green square tiles on the walls and floor, creating a grid pattern. Above her, a digital LED sign displays the word &laquo;&nbsp;MOTO&nbsp;&raquo; in bright orange dots. Flanking the space are vertical panels of small, glowing lights. A large, round, mesh-textured light fixture hangs from the ceiling. The lighting is warm and moody, dominated by amber and orange tones that cast strong shadows and highlight the subject against the darker background. The composition is centered, emphasizing the subject&rsquo;s pose and the enclosed, club-like atmosphere. **Prompt:** low angle shot, full body, young woman, light skin, long thin dreadlocks, wearing large beige over-ear headphones, black sports bra with white trim, black cargo pants,</code></pre><button class="wp-custom-copy-btn" onclick="copyCodeSnippet(this)" aria-label="Copier le code"><svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><rect x="9" y="9" width="13" height="13" rx="2" ry="2"></rect><path d="M5 15H4a2 2 0 0 1-2-2V4a2 2 0 0 1 2-2h9a2 2 0 0 1 2 2v1"></path></svg></button></div></p>
<h2 class="header-anchor-post">UNIVERS</h2>
<p><a href="http://www.h2-blog.com/wp-content/uploads/2026/11/flux-3-vs-krea-2-vs-z-image-turbo_00007_.png"><img class="aligncenter size-large wp-image-11927" src="http://www.h2-blog.com/wp-content/uploads/2026/11/flux-3-vs-krea-2-vs-z-image-turbo_00007_-650x650.png" alt="flux-3-vs-krea-2-vs-z-image-turbo_00007_" width="650" height="650" /></a><div class="wp-custom-code-container"><pre><code>A side-profile illustration of a young person with short, voluminous hair styled in a textured, cloud-like or nebula formation. The hair is composed of deep navy blue, medium blue, and cyan watercolor-style washes, interspersed with small yellow starbursts and white dots resembling stars, giving it a cosmic or dreamy appearance. The skin is pale beige with smooth shading. The eye is large, almond-shaped, and open, featuring a dark iris with a white catchlight, looking slightly upward and to the right. The expression is calm and serene. The ear is visible on the left side of the head. The clothing consists of a high-collared garment or jacket that matches the hair&rsquo;s color palette—blues and teals—with a similar abstract, splattered pattern. A hand is extended from the bottom right, palm up, holding a small, swirling orb of the same blue, cloud-like energy with stars within it. The background is a solid, flat beige or light tan color. The lighting is soft and even, highlighting the contours of the face and the texture of the hair without harsh shadows. The composition is a close-up portrait, filling most of the frame with the subject&rsquo;s head and upper torso. The artistic style is</code></pre><button class="wp-custom-copy-btn" onclick="copyCodeSnippet(this)" aria-label="Copier le code"><svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><rect x="9" y="9" width="13" height="13" rx="2" ry="2"></rect><path d="M5 15H4a2 2 0 0 1-2-2V4a2 2 0 0 1 2-2h9a2 2 0 0 1 2 2v1"></path></svg></button></div></p>
<h2 class="header-anchor-post">CERISES</h2>
<p><a href="http://www.h2-blog.com/wp-content/uploads/2026/11/flux-3-vs-krea-2-vs-z-image-turbo_00008_.png"><img class="aligncenter size-large wp-image-11928" src="http://www.h2-blog.com/wp-content/uploads/2026/11/flux-3-vs-krea-2-vs-z-image-turbo_00008_-650x650.png" alt="flux-3-vs-krea-2-vs-z-image-turbo_00008_" width="650" height="650" /></a><div class="wp-custom-code-container"><pre><code>A close-up, eye-level shot of a white porcelain teacup tipped over onto its side, spilling a cascade of glistening, deep red cherries with long stems. The cherries are piled high on the cup&rsquo;s rim and onto a matching white saucer below. Some cherries spill out from the saucer onto a sheet of aged, yellowed musical score paper printed with black staves and notes. The saucer and cup rest on a crisp white tablecloth. In the background, out of focus, is a lush garden bed filled with vibrant purple and pink flowers and green foliage, creating a soft bokeh effect. The lighting is bright, natural sunlight coming from the upper left, casting soft shadows to the right and highlighting the glossy texture of the fruit. The composition uses a shallow depth of field to isolate the sharp foreground subject against the blurred floral backdrop. The color palette is dominated by the vivid red of the cherries, the stark white of the ceramics, and the complementary purples and greens of the background. The aesthetic is fresh, summery, and appetizing, with a focus on texture and natural beauty.</code></pre><button class="wp-custom-copy-btn" onclick="copyCodeSnippet(this)" aria-label="Copier le code"><svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><rect x="9" y="9" width="13" height="13" rx="2" ry="2"></rect><path d="M5 15H4a2 2 0 0 1-2-2V4a2 2 0 0 1 2-2h9a2 2 0 0 1 2 2v1"></path></svg></button></div></p>
<h2 class="header-anchor-post">PHOTO DE STUDIOS</h2>
<p><a href="http://www.h2-blog.com/wp-content/uploads/2026/11/flux-3-vs-krea-2-vs-z-image-turbo_00009_.png"><img class="aligncenter size-large wp-image-11931" src="http://www.h2-blog.com/wp-content/uploads/2026/11/flux-3-vs-krea-2-vs-z-image-turbo_00009_-650x650.png" alt="flux-3-vs-krea-2-vs-z-image-turbo_00009_" width="650" height="650" /></a></p>
<div class="wp-custom-code-container"><pre><code>A high-angle, close-up portrait of a young East Asian woman lying on her side, propped up on her right elbow and chin. She has fair skin with a soft, warm glow on her cheek and forehead, contrasting with the deep black background. Her dark brown hair is styled in a loose, slightly messy updo with wispy strands framing her face and falling over her left shoulder. She wears large, round pearl earrings that catch the light. Her makeup is polished: defined eyebrows, long eyelashes, dark eyeliner, and bright coral-pink lipstick. Her eyes are wide and looking directly at the camera with an intense, slightly melancholic gaze. She is wearing a voluminous, lavender-purple gown made of a sheer, flowing fabric like chiffon or organza. The fabric drapes heavily around her body, creating soft folds and ruffles that extend into the foreground. A delicate silver chain necklace is visible against her collarbone. On her right hand, resting under her chin, she wears a simple silver ring on her ring finger. Her fingernails are painted white. The lighting is dramatic and directional, likely studio lighting, coming from the upper left, illuminating the right side of her face and body while leaving the left side in shadow.</code></pre><button class="wp-custom-copy-btn" onclick="copyCodeSnippet(this)" aria-label="Copier le code"><svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><rect x="9" y="9" width="13" height="13" rx="2" ry="2"></rect><path d="M5 15H4a2 2 0 0 1-2-2V4a2 2 0 0 1 2-2h9a2 2 0 0 1 2 2v1"></path></svg></button></div>
<h2 class="header-anchor-post">ANGE</h2>
<p><a href="http://www.h2-blog.com/wp-content/uploads/2026/11/flux-3-vs-krea-2-vs-z-image-turbo_00010_.png"><img class="aligncenter size-large wp-image-11932" src="http://www.h2-blog.com/wp-content/uploads/2026/11/flux-3-vs-krea-2-vs-z-image-turbo_00010_-650x650.png" alt="flux-3-vs-krea-2-vs-z-image-turbo_00010_" width="650" height="650" /></a><div class="wp-custom-code-container"><pre><code>A vertical digital illustration featuring a young girl with reddish-brown hair and large white feathered wings, falling backward through a vast, abstract beige void. She wears a dark blue school skirt, a light grey long-sleeved top, black knee-high socks, and brown shoes. Her pose is dynamic and inverted, with her legs kicked up towards the sky and her left arm reaching upwards as if grasping for something. Her expression is one of serene determination or peaceful surrender. Above her, a swirling vortex of deep blues, teals, and lighter blues forms a celestial scene, containing a large pale yellow moon, scattered stars (some four-pointed), and small yellow dots resembling distant lights or stardust. The vortex has a liquid, fluid texture with sharp, angular edges where it meets the beige background. The lighting is soft and diffuse, emanating from the celestial body above, casting gentle shadows on the girl&rsquo;s back and wings. The composition is asymmetrical, with the vortex dominating the upper third and the girl positioned in the lower center. The color palette consists of muted beiges, cool blues, teals, soft yellows, and skin tones. The overall aesthetic style is reminiscent of modern anime art with a painterly, illustrative quality, utilizing flat</code></pre><button class="wp-custom-copy-btn" onclick="copyCodeSnippet(this)" aria-label="Copier le code"><svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><rect x="9" y="9" width="13" height="13" rx="2" ry="2"></rect><path d="M5 15H4a2 2 0 0 1-2-2V4a2 2 0 0 1 2-2h9a2 2 0 0 1 2 2v1"></path></svg></button></div></p>
<h2 class="header-anchor-post">LUNETTES</h2>
<p><a href="http://www.h2-blog.com/wp-content/uploads/2026/11/flux-3-vs-krea-2-vs-z-image-turbo_00011_.png"><img class="aligncenter size-large wp-image-11933" src="http://www.h2-blog.com/wp-content/uploads/2026/11/flux-3-vs-krea-2-vs-z-image-turbo_00011_-650x975.png" alt="flux-3-vs-krea-2-vs-z-image-turbo_00011_" width="650" height="975" /></a><div class="wp-custom-code-container"><pre><code>An anime-style close-up portrait of a young woman with short, messy white hair and pale skin, lying down or reclining. She wears large, round sunglasses with reflective lenses showing a vivid abstract pattern of blue sky, white fluffy clouds, and red sunsets. Her lips are full, glossy pink, slightly parted to reveal a white tooth. Her fingernails are painted with a dark blue polish featuring red and white abstract splatter designs. She is wearing a dark blue top with a similar cloud/sky print. The background is a soft, out-of-focus light gray or white. The lighting is bright and high-contrast, creating strong highlights on her cheekbones, nose, and the metallic frame of the glasses, with deep shadows under her chin and along her neck. The camera angle is a low-angle shot looking up at her face. The composition is tight, focusing on her upper face, hands adjusting the glasses, and the reflection in the lenses. The aesthetic is vibrant, stylish, and modern, reminiscent of digital art or manga illustrations with smooth shading and sharp linework.</code></pre><button class="wp-custom-copy-btn" onclick="copyCodeSnippet(this)" aria-label="Copier le code"><svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><rect x="9" y="9" width="13" height="13" rx="2" ry="2"></rect><path d="M5 15H4a2 2 0 0 1-2-2V4a2 2 0 0 1 2-2h9a2 2 0 0 1 2 2v1"></path></svg></button></div></p>
<h2 class="header-anchor-post">CONCLUSION</h2>
<p>Dans la plupart de mes tests, les résultats sont similaires entre ces 4 modèles. Clairement, les autres modèles sont au même niveau que Z-Image Turbo.</p>
<p><strong>Z-Image Turbo</strong> : toujours le meilleur pour le réalisme.</p>
<p><strong>Krea 2 Turbo</strong> : légèrement en tête pour l&rsquo;illustration.</p>
<p><strong>Qwen Image 2.1</strong> : correct, mais le problème de bruit doit être corrigé.</p>
<p><strong>FLUX 3 Klein</strong> : a du potentiel, surtout avec la première image. Le premier prompt peut être un bon point de départ pour créer de superbes images.</p>
</div>
</div>
<p>Cet article <a href="https://www.h2-blog.com/2026/11/05/flux-3-klein-vs-krea-2-vs-qwen-2-1-vs-z-image-turbo/">FLUX 3 KLEIN vs KREA 2 vs Qwen 2.1 vs Z image turbo</a> provient du site <a href="https://www.h2-blog.com">H2-Blog</a>.</p>]]></content:encoded>
			<wfw:commentRss>https://www.h2-blog.com/2026/11/05/flux-3-klein-vs-krea-2-vs-qwen-2-1-vs-z-image-turbo/feed/</wfw:commentRss>
		<slash:comments>0</slash:comments>
		</item>
		<item>
		<title>Commet régler le problème de la peau en plastique sur les images avec l’IA et ComfyUI</title>
		<link>https://www.h2-blog.com/2026/11/05/commet-regler-probleme-peau-en-plastique-les-images-lia-comfyui/</link>
		<comments>https://www.h2-blog.com/2026/11/05/commet-regler-probleme-peau-en-plastique-les-images-lia-comfyui/#comments</comments>
		<pubDate>Thu, 05 Nov 2026 19:50:53 +0000</pubDate>
		<dc:creator><![CDATA[Guillaume R]]></dc:creator>
				<category><![CDATA[ComfyUI]]></category>

		<guid isPermaLink="false">http://www.h2-blog.com/?p=11901</guid>
		<description><![CDATA[<p>Est-ce que les images que vous générez avec l&#8217;IA (ComfyUI) ont l&#8217;air de AI slop avec une peau en plastique ? Cet article va vous aider à résoudre ce problème. Les flux de travail présentés dans cet article sont basés sur l&#8217;upscaler 2K de Z Image Turbo. La partie de mise à l&#8217;échelle a été [&#8230;]</p><p>Cet article <a href="https://www.h2-blog.com/2026/11/05/commet-regler-probleme-peau-en-plastique-les-images-lia-comfyui/">Commet régler le problème de la peau en plastique sur les images avec l’IA et ComfyUI</a> provient du site <a href="https://www.h2-blog.com">H2-Blog</a>.</p>]]></description>
				<content:encoded><![CDATA[<p>Est-ce que les images que vous générez avec l&rsquo;IA (ComfyUI) ont l&rsquo;air de AI slop avec une peau en plastique ?</p>
<p><a href="http://www.h2-blog.com/wp-content/uploads/2026/11/KREA2_peau-en-plastique.jpg"><img class="aligncenter size-large wp-image-11903" src="http://www.h2-blog.com/wp-content/uploads/2026/11/KREA2_peau-en-plastique-650x325.jpg" alt="KREA2_peau-en plastique" width="650" height="325" /></a></p>
<p>Cet article va vous aider à résoudre ce problème. Les flux de travail présentés dans cet article sont basés sur l&rsquo;upscaler 2K de Z Image Turbo. La partie de mise à l&rsquo;échelle a été supprimée et le modèle a été modifié pour la version INT8.</p>
<p><a href="http://www.h2-blog.com/wp-content/uploads/2026/11/z-image-turbo-realisme.png"><img class="aligncenter size-large wp-image-11904" src="http://www.h2-blog.com/wp-content/uploads/2026/11/z-image-turbo-realisme-650x344.png" alt="z-image-turbo-realisme" width="650" height="344" /></a></p>
<p>Les performances pour une image 1024&#215;1024 : 3 secondes avec une 5070 Ti, et 6 secondes avec une 5060 Ti 16 Go.</p>
<p>Vous pouvez les ajouter facilement au flux de travail de Z Image Turbo, avec une légère augmentation du temps de création de l&rsquo;image.</p>
<p><a href="http://www.h2-blog.com/wp-content/uploads/2026/11/Krea-2-trubo-avec-z-image.png"><img class="aligncenter size-large wp-image-11905" src="http://www.h2-blog.com/wp-content/uploads/2026/11/Krea-2-trubo-avec-z-image-650x228.png" alt="Krea-2-trubo-avec-z-image" width="650" height="228" /></a></p>
<p>Les performances pour une image 1024&#215;1024 KREA2 Turbo + Z Image Turbo : 9,5 secondes avec une 5070 Ti, et 18 secondes avec une 5060 Ti 16 Go.</p>
<p>Ces deux flux de travail sont disponibles sur <a href="https://civitai.com/user/def813/models?sort=Highest+Rated&amp;tag=realism">Civitai</a> et <a href="https://github.com/H2logs/Fix-plastic-skin-improve-realism">Github</a>.</p>
<p>Cet article <a href="https://www.h2-blog.com/2026/11/05/commet-regler-probleme-peau-en-plastique-les-images-lia-comfyui/">Commet régler le problème de la peau en plastique sur les images avec l’IA et ComfyUI</a> provient du site <a href="https://www.h2-blog.com">H2-Blog</a>.</p>]]></content:encoded>
			<wfw:commentRss>https://www.h2-blog.com/2026/11/05/commet-regler-probleme-peau-en-plastique-les-images-lia-comfyui/feed/</wfw:commentRss>
		<slash:comments>0</slash:comments>
		</item>
		<item>
		<title>Des images générées avec Moss image</title>
		<link>https://www.h2-blog.com/2026/10/20/images-generees-moss-image/</link>
		<comments>https://www.h2-blog.com/2026/10/20/images-generees-moss-image/#comments</comments>
		<pubDate>Tue, 20 Oct 2026 11:35:37 +0000</pubDate>
		<dc:creator><![CDATA[Guillaume R]]></dc:creator>
				<category><![CDATA[ComfyUI]]></category>

		<guid isPermaLink="false">http://www.h2-blog.com/?p=11868</guid>
		<description><![CDATA[<p>Moss image est le nouveau modèle à la mode qui vient de sortir voyons ce qu&#8217;il a dans le ventre. Le flux de travail est disponible Civitai et Github. les fichiers nécessaires sont disponible sur huggingface. CONCLUSION Les résultats sont très bons. Les gens sous-estiment les capacités de Moss image pour générer des illustrations. Le seul point négatif [&#8230;]</p><p>Cet article <a href="https://www.h2-blog.com/2026/10/20/images-generees-moss-image/">Des images générées avec Moss image</a> provient du site <a href="https://www.h2-blog.com">H2-Blog</a>.</p>]]></description>
				<content:encoded><![CDATA[<p>Moss image est le nouveau modèle à la mode qui vient de sortir voyons ce qu&rsquo;il a dans le ventre.</p>
<p>Le flux de travail est disponible <a href="https://civitai.com/models/3009145/moss-image-int8-workflow">Civitai</a> et <a href="https://github.com/H2logs/Workflows-archive-from-H2logs.substack.com/blob/main/moss_image_t2i.json">Github</a>. les fichiers nécessaires sont disponible sur <a href="https://huggingface.co/Comfy-Org/MOSS-Image">huggingface</a>.</p>
<p><a href="http://www.h2-blog.com/wp-content/uploads/2026/10/moss_image_t2i_00019_.png"><img class="aligncenter size-large wp-image-11869" src="http://www.h2-blog.com/wp-content/uploads/2026/10/moss_image_t2i_00019_-650x650.png" alt="moss_image_t2i_00019_" width="650" height="650" /></a></p>
<div>
<div><div class="wp-custom-code-container"><pre><code>A vertical digital painting in a whimsical, anime-inspired fantasy style. A young woman with short black bob haircut and pale skin sits cross-legged on the right side of a large, anthropomorphic crescent moon that dominates the left half of the frame. She wears a traditional black kimono with wide sleeves, accented by light teal undergarments visible at the collar and hem, and dark knee-high boots. Her expression is serene and contemplative as she gazes upward toward the stars. In her right hand, she holds a long-handled paintbrush with a glowing blue tip; in her left, she grips a slender white staff or wand. The brush emits a trail of sparkling starlight as it paints new stars into the night sky — one bright six-pointed star appears mid-air near the moon’s eye, another fainter star floats nearby.The crescent moon itself has a human face: a large expressive eye with long lashes and iridescent blue-green iris, a soft blush on its cheek, and a gentle smile. Its surface is textured with golden craters and painted highlights, giving it a metallic, celestial appearance. The background is a deep cosmic void rendered in rich shades of navy, indigo, and violet, speckled with tiny white stars</code></pre><button class="wp-custom-copy-btn" onclick="copyCodeSnippet(this)" aria-label="Copier le code"><svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><rect x="9" y="9" width="13" height="13" rx="2" ry="2"></rect><path d="M5 15H4a2 2 0 0 1-2-2V4a2 2 0 0 1 2-2h9a2 2 0 0 1 2 2v1"></path></svg></button></div></div>
</div>
<div></div>
<div><a href="http://www.h2-blog.com/wp-content/uploads/2026/10/moss_image_t2i_00012_.png"><img class="aligncenter size-large wp-image-11871" src="http://www.h2-blog.com/wp-content/uploads/2026/10/moss_image_t2i_00012_-650x650.png" alt="moss_image_t2i_00012_" width="650" height="650" /></a></div>
<div>
<div>
<div><div class="wp-custom-code-container"><pre><code>A medium close-up side profile illustration of a young person with pale skin and sharp, angular facial features, gazing upward and to the left. The subject has short, tousled dark blue hair that blends seamlessly into an abstract, fluid background of swirling ocean waves, deep blues, teals, and lighter sky blues. The hair is animated with dynamic, curling strands that extend outward like liquid smoke or water splashes. Scattered throughout the composition are small, glowing yellow four-pointed stars and dots, resembling fireflies or distant stars, adding a magical, ethereal quality. The subject&rsquo;s visible eye is large, almond-shaped, and a striking light blue, detailed with a white catchlight and thin black eyelashes. The mouth is slightly parted, revealing a hint of the tongue, suggesting a sense of wonder or breathless awe. The subject is wearing a simple hooded garment in solid shades of medium and dark blue, which frames the head from the top and left side. The hood&rsquo;s edge is defined by a thick black outline. The background is split: the left side is dominated by the solid blue tones of the hood and shadow, while the right side transitions into the beige/cream-colored canvas where the fluid, abstract water-like patterns and stars are</code></pre><button class="wp-custom-copy-btn" onclick="copyCodeSnippet(this)" aria-label="Copier le code"><svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><rect x="9" y="9" width="13" height="13" rx="2" ry="2"></rect><path d="M5 15H4a2 2 0 0 1-2-2V4a2 2 0 0 1 2-2h9a2 2 0 0 1 2 2v1"></path></svg></button></div></div>
</div>
</div>
<div><a href="http://www.h2-blog.com/wp-content/uploads/2026/10/moss_image_t2i_00039_.png"><img class="aligncenter size-large wp-image-11872" src="http://www.h2-blog.com/wp-content/uploads/2026/10/moss_image_t2i_00039_-650x650.png" alt="moss_image_t2i_00039_" width="650" height="650" /></a></div>
<div>
<div>
<div><div class="wp-custom-code-container"><pre><code>A vibrant, stylized illustration of a fantastical character resembling a witch or fairy, set against a whimsical, dreamlike background. The subject is a slender figure with a smooth, yellow complexion and a simple, joyful expression featuring closed eyes, rosy cheeks, and a gentle smile. They wear a wide-brimmed, conical purple hat adorned with white geometric patterns (triangles and dots) along the brim. Their attire consists of a short-sleeved, purple tunic with black-and-white striped cuffs and collar, decorated with small white floral motifs.The most striking feature is the expansive, flowing garment that acts as wings or a skirt. It is primarily a vivid magenta-pink color, composed of large, layered panels. These panels are patterned with white circles of varying sizes and pink cross-hatching lines. Extending from the lower section of this garment are three distinct, crescent-shaped elements resembling fish tails or fins. These are bright yellow with intricate blue and green tribal-style patterns, including leafy vines and geometric shapes.The background is a dynamic sky in shades of soft pink and lavender. It features swirling yellow and teal lines suggesting wind or magic, dotted with small yellow specks. There are stylized white clouds with black</code></pre><button class="wp-custom-copy-btn" onclick="copyCodeSnippet(this)" aria-label="Copier le code"><svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><rect x="9" y="9" width="13" height="13" rx="2" ry="2"></rect><path d="M5 15H4a2 2 0 0 1-2-2V4a2 2 0 0 1 2-2h9a2 2 0 0 1 2 2v1"></path></svg></button></div></div>
</div>
</div>
<div></div>
<div><a href="http://www.h2-blog.com/wp-content/uploads/2026/10/moss_image_t2i_00040_.png"><img class="aligncenter size-large wp-image-11874" src="http://www.h2-blog.com/wp-content/uploads/2026/10/moss_image_t2i_00040_-650x650.png" alt="moss_image_t2i_00040_" width="650" height="650" /></a></div>
<div>
<div>
<div><div class="wp-custom-code-container"><pre><code>Portrait of a young woman with wavy, blonde-highlighted hair, blue eyes, and light makeup featuring shimmery eyeshadow and nude lipstick. She is gazing upward and slightly to the side with a soft, contemplative expression, her head tilted back slightly. Her lips are parted just enough to reveal the top teeth. She has fair skin with natural freckles across the nose and cheeks. She is wearing a light-colored, possibly beige or cream, top with thin straps or a V-neckline. The background is a soft, out-of-focus gradient of pale blue and white, creating a dreamy, airy atmosphere. The lighting is soft and diffused, highlighting the contours of her face and adding a gentle glow to her skin. The camera angle is a close-up, slightly low-angle shot looking up at her face. The composition focuses tightly on her head and shoulders, with her hair framing her face on both sides. The overall aesthetic is ethereal, soft, and feminine, reminiscent of high-fashion editorial photography or a beauty advertisement.</code></pre><button class="wp-custom-copy-btn" onclick="copyCodeSnippet(this)" aria-label="Copier le code"><svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><rect x="9" y="9" width="13" height="13" rx="2" ry="2"></rect><path d="M5 15H4a2 2 0 0 1-2-2V4a2 2 0 0 1 2-2h9a2 2 0 0 1 2 2v1"></path></svg></button></div></div>
</div>
</div>
<div></div>
<div><a href="http://www.h2-blog.com/wp-content/uploads/2026/10/moss_image_t2i_00041_.png"><img class="aligncenter size-large wp-image-11875" src="http://www.h2-blog.com/wp-content/uploads/2026/10/moss_image_t2i_00041_-650x650.png" alt="moss_image_t2i_00041_" width="650" height="650" /></a></div>
<div>
<div>
<div><div class="wp-custom-code-container"><pre><code>A fluffy, long-haired cat with a cream and brown tabby coat sits upright on a paved stone surface, facing forward. The cat has a white muzzle and chest, with brown and black fur on its head and back. Its eyes are greenish-yellow and alert, gazing directly at the viewer. The cat&rsquo;s ears are perked up, and its tail is curled around its side. It is not wearing any clothing.The background features a modern building with large glass windows and balconies. A red flag with a white crescent moon and star (the Turkish flag) hangs vertically in the background, slightly out of focus. There are metal railings and street lamps visible along the building facade. The lighting is warm and golden, suggesting late afternoon or early evening sunlight hitting the cat from the side, creating soft highlights on its fur. The camera angle is low, almost eye-level with the seated cat, emphasizing its presence. The composition places the cat centrally in the foreground, with the building and flag providing depth in the background. The color palette consists of warm earth tones (browns, creams, tans) contrasted with the bright red of the flag and the cool blue-grey of the building&rsquo;s glass. The overall aesthetic is realistic, detailed,</code></pre><button class="wp-custom-copy-btn" onclick="copyCodeSnippet(this)" aria-label="Copier le code"><svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><rect x="9" y="9" width="13" height="13" rx="2" ry="2"></rect><path d="M5 15H4a2 2 0 0 1-2-2V4a2 2 0 0 1 2-2h9a2 2 0 0 1 2 2v1"></path></svg></button></div></div>
<div><a href="http://www.h2-blog.com/wp-content/uploads/2026/10/moss_image_t2i_00037_.png"><img class="aligncenter size-large wp-image-11879" src="http://www.h2-blog.com/wp-content/uploads/2026/10/moss_image_t2i_00037_-650x650.png" alt="moss_image_t2i_00037_" width="650" height="650" /></a><div class="wp-custom-code-container"><pre><code>Portrait of a young woman in her late teens to early twenties, captured in a high-resolution, cinematic close-up shot. She is facing slightly away from the camera but turning her head to look directly over her left shoulder with an engaging, soft smile. Her dark brown hair is styled in a messy, voluminous top knot (bun) with loose strands framing her face and falling naturally around her neck and shoulders; some strands are damp with rain. Her facial features are symmetrical and refined: well-defined eyebrows, large expressive hazel-brown eyes with subtle eyeliner and natural lashes, a straight nose, and full lips with a glossy finish. She wears minimal jewelry — a small stud earring visible in her left ear. She is wearing a bright mustard-yellow hooded raincoat made of waterproof nylon or similar synthetic material, heavily textured with glistening water droplets across the surface, especially prominent on the hood, shoulders, and front panel. The coat has a white drawstring with plastic toggles at the collar, silver-toned snap buttons down the front, and visible stitching details. Underneath, a light gray crewneck sweatshirt is partially visible at the neckline. Her pose is dynamic yet relaxed — body angled away, head turned toward viewer, creating depth</code></pre><button class="wp-custom-copy-btn" onclick="copyCodeSnippet(this)" aria-label="Copier le code"><svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><rect x="9" y="9" width="13" height="13" rx="2" ry="2"></rect><path d="M5 15H4a2 2 0 0 1-2-2V4a2 2 0 0 1 2-2h9a2 2 0 0 1 2 2v1"></path></svg></button></div></div>
<div><a href="http://www.h2-blog.com/wp-content/uploads/2026/10/moss_image_t2i_00038_-1.png"><img class="aligncenter size-large wp-image-11880" src="http://www.h2-blog.com/wp-content/uploads/2026/10/moss_image_t2i_00038_-1-650x650.png" alt="moss_image_t2i_00038_ 1" width="650" height="650" /></a><div class="wp-custom-code-container"><pre><code>A full-body, eye-level shot of a motocross rider in dynamic motion on a dirt track, captured with sharp focus on the subject and a shallow depth of field blurring the background. The rider is positioned on the right side of the frame, leaning forward aggressively into a turn or straightaway, knees bent and gripping handlebars equipped with black perforated handguards. They wear a full-face helmet featuring a high-contrast design of matte black and neon yellow geometric patterns, paired with reflective blue-tinted goggles. Their attire consists of a long-sleeved black and white jersey, black protective pants with white diagonal stripes on the thighs, and heavy-duty black motocross boots. The dirt bike is a black dual-sport model with knobby tires kicking up a spray of brown dust and dirt particles at the rear wheel, indicating speed and traction.The background features a hazy, overcast sky in pale grey-blue tones. To the left, out of focus, stands a large air traffic control tower topped with a prominent white spherical radar dome, suggesting an airport setting adjacent to the off-road course. Two weathered wooden posts stand vertically on either side of the path. The ground is composed of dry, brownish grass and packed earth. The lighting is soft and</code></pre><button class="wp-custom-copy-btn" onclick="copyCodeSnippet(this)" aria-label="Copier le code"><svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><rect x="9" y="9" width="13" height="13" rx="2" ry="2"></rect><path d="M5 15H4a2 2 0 0 1-2-2V4a2 2 0 0 1 2-2h9a2 2 0 0 1 2 2v1"></path></svg></button></div></div>
<div><a href="http://www.h2-blog.com/wp-content/uploads/2026/10/moss_image_t2i_00042_.png"><img class="aligncenter size-large wp-image-11881" src="http://www.h2-blog.com/wp-content/uploads/2026/10/moss_image_t2i_00042_-650x650.png" alt="moss_image_t2i_00042_" width="650" height="650" /></a><div class="wp-custom-code-container"><pre><code>A low-angle, full-body shot of a young woman with light skin and long, thin dreadlocks pulled back, wearing large over-ear headphones. She is dressed in a black sports bra with white trim and matching black cargo pants. Her left arm is raised, hand near her head adjusting the headphones, while her right hand rests on a dark object to her side. Visible tattoos adorn her left forearm. She wears multiple bracelets on her right wrist and a necklace. Her expression is serious and focused, eyes looking downward. The background features dark green square tiles on the walls and floor, creating a grid pattern. Above her, a digital LED sign displays the word &laquo;&nbsp;MOTO&nbsp;&raquo; in bright orange dots. Flanking the space are vertical panels of small, glowing lights. A large, round, mesh-textured light fixture hangs from the ceiling. The lighting is warm and moody, dominated by amber and orange tones that cast strong shadows and highlight the subject against the darker background. The composition is centered, emphasizing the subject&rsquo;s pose and the enclosed, club-like atmosphere. **Prompt:** low angle shot, full body, young woman, light skin, long thin dreadlocks, wearing large beige over-ear headphones, black sports bra with white trim, black cargo pants,</code></pre><button class="wp-custom-copy-btn" onclick="copyCodeSnippet(this)" aria-label="Copier le code"><svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><rect x="9" y="9" width="13" height="13" rx="2" ry="2"></rect><path d="M5 15H4a2 2 0 0 1-2-2V4a2 2 0 0 1 2-2h9a2 2 0 0 1 2 2v1"></path></svg></button></div></div>
</div>
<div><a href="http://www.h2-blog.com/wp-content/uploads/2026/10/moss_image_t2i_00017_.png"><img class="aligncenter size-large wp-image-11882" src="http://www.h2-blog.com/wp-content/uploads/2026/10/moss_image_t2i_00017_-650x650.png" alt="moss_image_t2i_00017_" width="650" height="650" /></a><div class="wp-custom-code-container"><pre><code>**Subject Appearance:**  A stylized, silhouette-based humanoid figure with a single glowing red eye visible on the left side of its face. The head has two small, upward-pointing tufts resembling horns or styled hair buns. The body is slender and elongated, standing upright in profile view facing left. **Clothing:**  The figure wears a long, flowing robe or cloak that drapes asymmetrically — one shoulder bare, the other covered by fabric that billows dramatically behind it as if caught in wind. The garment ends mid-calf, revealing bare legs. No facial features are rendered except the singular red eye; no mouth, nose, or eyebrows are depicted. **Pose &amp; Expression:**&lt;br /&gt;
Standing firmly atop a floating platform, legs straight, feet planted. Head turned slightly toward the viewer but mostly profiled. Expression is neutral yet haunting due to the lone glowing eye — evokes mystery, solitude, or melancholy. No movement suggests dynamism beyond implied motion from flowing clothes. **Background Elements:**   &#8211; A massive, cracked pale-yellow moon dominates the upper center, veined with branching red lines resembling blood vessels or roots. &#8211; Below the moon, swirling clouds transition between teal-blue (left) and crimson-red (right</code></pre><button class="wp-custom-copy-btn" onclick="copyCodeSnippet(this)" aria-label="Copier le code"><svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><rect x="9" y="9" width="13" height="13" rx="2" ry="2"></rect><path d="M5 15H4a2 2 0 0 1-2-2V4a2 2 0 0 1 2-2h9a2 2 0 0 1 2 2v1"></path></svg></button></div></div>
<div><a href="http://www.h2-blog.com/wp-content/uploads/2026/10/moss_image_t2i_00044_.png"><img class="aligncenter size-large wp-image-11885" src="http://www.h2-blog.com/wp-content/uploads/2026/10/moss_image_t2i_00044_-650x650.png" alt="moss_image_t2i_00044_" width="650" height="650" /></a><div class="wp-custom-code-container"><pre><code>A side-profile illustration of a young person with short, voluminous hair styled in a textured, cloud-like or nebula formation. The hair is composed of deep navy blue, medium blue, and cyan watercolor-style washes, interspersed with small yellow starbursts and white dots resembling stars, giving it a cosmic or dreamy appearance. The skin is pale beige with smooth shading. The eye is large, almond-shaped, and open, featuring a dark iris with a white catchlight, looking slightly upward and to the right. The expression is calm and serene. The ear is visible on the left side of the head. The clothing consists of a high-collared garment or jacket that matches the hair&rsquo;s color palette—blues and teals—with a similar abstract, splattered pattern. A hand is extended from the bottom right, palm up, holding a small, swirling orb of the same blue, cloud-like energy with stars within it. The background is a solid, flat beige or light tan color. The lighting is soft and even, highlighting the contours of the face and the texture of the hair without harsh shadows. The composition is a close-up portrait, filling most of the frame with the subject&rsquo;s head and upper torso. The artistic style is</code></pre><button class="wp-custom-copy-btn" onclick="copyCodeSnippet(this)" aria-label="Copier le code"><svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><rect x="9" y="9" width="13" height="13" rx="2" ry="2"></rect><path d="M5 15H4a2 2 0 0 1-2-2V4a2 2 0 0 1 2-2h9a2 2 0 0 1 2 2v1"></path></svg></button></div></div>
<div><a href="http://www.h2-blog.com/wp-content/uploads/2026/10/moss_image_t2i_00045_.png"><img class="aligncenter size-large wp-image-11886" src="http://www.h2-blog.com/wp-content/uploads/2026/10/moss_image_t2i_00045_-650x650.png" alt="moss_image_t2i_00045_" width="650" height="650" /></a><div class="wp-custom-code-container"><pre><code>A close-up, eye-level shot of a white porcelain teacup tipped over onto its side, spilling a cascade of glistening, deep red cherries with long stems. The cherries are piled high on the cup&rsquo;s rim and onto a matching white saucer below. Some cherries spill out from the saucer onto a sheet of aged, yellowed musical score paper printed with black staves and notes. The saucer and cup rest on a crisp white tablecloth. In the background, out of focus, is a lush garden bed filled with vibrant purple and pink flowers and green foliage, creating a soft bokeh effect. The lighting is bright, natural sunlight coming from the upper left, casting soft shadows to the right and highlighting the glossy texture of the fruit. The composition uses a shallow depth of field to isolate the sharp foreground subject against the blurred floral backdrop. The color palette is dominated by the vivid red of the cherries, the stark white of the ceramics, and the complementary purples and greens of the background. The aesthetic is fresh, summery, and appetizing, with a focus on texture and natural beauty.</code></pre><button class="wp-custom-copy-btn" onclick="copyCodeSnippet(this)" aria-label="Copier le code"><svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><rect x="9" y="9" width="13" height="13" rx="2" ry="2"></rect><path d="M5 15H4a2 2 0 0 1-2-2V4a2 2 0 0 1 2-2h9a2 2 0 0 1 2 2v1"></path></svg></button></div></div>
<div><a href="http://www.h2-blog.com/wp-content/uploads/2026/10/moss_image_t2i_00046_.png"><img class="aligncenter size-large wp-image-11887" src="http://www.h2-blog.com/wp-content/uploads/2026/10/moss_image_t2i_00046_-650x650.png" alt="moss_image_t2i_00046_" width="650" height="650" /></a><div class="wp-custom-code-container"><pre><code>A high-angle, close-up portrait of a young East Asian woman lying on her side, propped up on her right elbow and chin. She has fair skin with a soft, warm glow on her cheek and forehead, contrasting with the deep black background. Her dark brown hair is styled in a loose, slightly messy updo with wispy strands framing her face and falling over her left shoulder. She wears large, round pearl earrings that catch the light. Her makeup is polished: defined eyebrows, long eyelashes, dark eyeliner, and bright coral-pink lipstick. Her eyes are wide and looking directly at the camera with an intense, slightly melancholic gaze. She is wearing a voluminous, lavender-purple gown made of a sheer, flowing fabric like chiffon or organza. The fabric drapes heavily around her body, creating soft folds and ruffles that extend into the foreground. A delicate silver chain necklace is visible against her collarbone. On her right hand, resting under her chin, she wears a simple silver ring on her ring finger. Her fingernails are painted white. The lighting is dramatic and directional, likely studio lighting, coming from the upper left, illuminating the right side of her face and body while leaving the left side in shadow.</code></pre><button class="wp-custom-copy-btn" onclick="copyCodeSnippet(this)" aria-label="Copier le code"><svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><rect x="9" y="9" width="13" height="13" rx="2" ry="2"></rect><path d="M5 15H4a2 2 0 0 1-2-2V4a2 2 0 0 1 2-2h9a2 2 0 0 1 2 2v1"></path></svg></button></div></div>
<div><a href="http://www.h2-blog.com/wp-content/uploads/2026/10/moss_image_t2i_00021_.png"><img class="aligncenter size-large wp-image-11888" src="http://www.h2-blog.com/wp-content/uploads/2026/10/moss_image_t2i_00021_-650x650.png" alt="moss_image_t2i_00021_" width="650" height="650" /></a><div class="wp-custom-code-container"><pre><code>High-resolution, close-up portrait of a young woman with fair skin and prominent freckles across her nose and cheeks. Her eyes are striking light green or hazel, large and expressive, gazing directly at the viewer with an intense, slightly melancholic expression. Lips are full, softly colored in muted rose-pink, slightly parted. Hair is dark brown, pulled back but with loose strands framing her face. She wears an elaborate black Victorian-style ensemble: a wide-brimmed hat made of sheer black lace overlaying a solid black fabric base, adorned with sparkling glitter threads that catch the light; ruffled, translucent black lace collar with intricate floral embroidery and shimmering silver specks cascading down the neckline onto her shoulders. The fabric has a delicate, semi-sheer quality with visible texture and depth. Pose is upright, head tilted slightly forward, chin lifted, creating a dignified yet vulnerable stance. Lighting is soft, directional studio lighting from the front-left, casting gentle shadows on the right side of her face and neck, highlighting facial contours and the sparkle in the fabric. Background is a smooth, neutral gray gradient, subtly darker toward the edges to emphasize the subject. Composition is tightly framed on the upper torso and head, centered vertically and horizontally, with shallow</code></pre><button class="wp-custom-copy-btn" onclick="copyCodeSnippet(this)" aria-label="Copier le code"><svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><rect x="9" y="9" width="13" height="13" rx="2" ry="2"></rect><path d="M5 15H4a2 2 0 0 1-2-2V4a2 2 0 0 1 2-2h9a2 2 0 0 1 2 2v1"></path></svg></button></div></div>
<div><a href="http://www.h2-blog.com/wp-content/uploads/2026/10/moss_image_t2i_00048_.png"><img class="aligncenter size-large wp-image-11890" src="http://www.h2-blog.com/wp-content/uploads/2026/10/moss_image_t2i_00048_-650x650.png" alt="moss_image_t2i_00048_" width="650" height="650" /></a><div class="wp-custom-code-container"><pre><code>A vertical digital illustration featuring a young girl with reddish-brown hair and large white feathered wings, falling backward through a vast, abstract beige void. She wears a dark blue school skirt, a light grey long-sleeved top, black knee-high socks, and brown shoes. Her pose is dynamic and inverted, with her legs kicked up towards the sky and her left arm reaching upwards as if grasping for something. Her expression is one of serene determination or peaceful surrender. Above her, a swirling vortex of deep blues, teals, and lighter blues forms a celestial scene, containing a large pale yellow moon, scattered stars (some four-pointed), and small yellow dots resembling distant lights or stardust. The vortex has a liquid, fluid texture with sharp, angular edges where it meets the beige background. The lighting is soft and diffuse, emanating from the celestial body above, casting gentle shadows on the girl&rsquo;s back and wings. The composition is asymmetrical, with the vortex dominating the upper third and the girl positioned in the lower center. The color palette consists of muted beiges, cool blues, teals, soft yellows, and skin tones. The overall aesthetic style is reminiscent of modern anime art with a painterly, illustrative quality, utilizing flat</code></pre><button class="wp-custom-copy-btn" onclick="copyCodeSnippet(this)" aria-label="Copier le code"><svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><rect x="9" y="9" width="13" height="13" rx="2" ry="2"></rect><path d="M5 15H4a2 2 0 0 1-2-2V4a2 2 0 0 1 2-2h9a2 2 0 0 1 2 2v1"></path></svg></button></div></div>
<div><a href="http://www.h2-blog.com/wp-content/uploads/2026/10/moss_image_t2i_00049_.png"><img class="aligncenter size-large wp-image-11891" src="http://www.h2-blog.com/wp-content/uploads/2026/10/moss_image_t2i_00049_-650x650.png" alt="moss_image_t2i_00049_" width="650" height="650" /></a><div class="wp-custom-code-container"><pre><code>An anime-style close-up portrait of a young woman with short, messy white hair and pale skin, lying down or reclining. She wears large, round sunglasses with reflective lenses showing a vivid abstract pattern of blue sky, white fluffy clouds, and red sunsets. Her lips are full, glossy pink, slightly parted to reveal a white tooth. Her fingernails are painted with a dark blue polish featuring red and white abstract splatter designs. She is wearing a dark blue top with a similar cloud/sky print. The background is a soft, out-of-focus light gray or white. The lighting is bright and high-contrast, creating strong highlights on her cheekbones, nose, and the metallic frame of the glasses, with deep shadows under her chin and along her neck. The camera angle is a low-angle shot looking up at her face. The composition is tight, focusing on her upper face, hands adjusting the glasses, and the reflection in the lenses. The aesthetic is vibrant, stylish, and modern, reminiscent of digital art or manga illustrations with smooth shading and sharp linework.</code></pre><button class="wp-custom-copy-btn" onclick="copyCodeSnippet(this)" aria-label="Copier le code"><svg width="16" height="16" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round"><rect x="9" y="9" width="13" height="13" rx="2" ry="2"></rect><path d="M5 15H4a2 2 0 0 1-2-2V4a2 2 0 0 1 2-2h9a2 2 0 0 1 2 2v1"></path></svg></button></div></div>
</div>
<h2>CONCLUSION</h2>
<p>Les résultats sont très bons. Les gens sous-estiment les capacités de Moss image pour générer des illustrations.</p>
<p>Le seul point négatif est la vitesse de Moss image. Avec le flux de travail INT8, il faut 12 secondes pour générer une image 1024×1024 avec une 5070 Ti, au lieu de 3 à 5 secondes avec les autres modèles (14 secondes avec une 5060 Ti 16 Go).</p>
<p>Cet article <a href="https://www.h2-blog.com/2026/10/20/images-generees-moss-image/">Des images générées avec Moss image</a> provient du site <a href="https://www.h2-blog.com">H2-Blog</a>.</p>]]></content:encoded>
			<wfw:commentRss>https://www.h2-blog.com/2026/10/20/images-generees-moss-image/feed/</wfw:commentRss>
		<slash:comments>0</slash:comments>
		</item>
		<item>
		<title>300 000 vues sur Twitter/x avec cette app ComfyUI</title>
		<link>https://www.h2-blog.com/2026/10/18/300-000-vues-twitterx-app-comfyui/</link>
		<comments>https://www.h2-blog.com/2026/10/18/300-000-vues-twitterx-app-comfyui/#comments</comments>
		<pubDate>Sun, 18 Oct 2026 11:40:12 +0000</pubDate>
		<dc:creator><![CDATA[Guillaume R]]></dc:creator>
				<category><![CDATA[ComfyUI]]></category>

		<guid isPermaLink="false">http://www.h2-blog.com/?p=11837</guid>
		<description><![CDATA[<p>J&#8217;essayais de poster des images populaires sur les réseaux sociaux pour tester ComfyUI dans la vraie vie. Ma première idée ne fonctionnait pas, alors j&#8217;ai cherché quelles étaient les images populaires sur X. Il s&#8217;avère que ce sont les images SFW de personnages d&#8217;animés japonais (femme ou homme transformé en femme). Comment j’ai eu 300K [&#8230;]</p><p>Cet article <a href="https://www.h2-blog.com/2026/10/18/300-000-vues-twitterx-app-comfyui/">300 000 vues sur Twitter/x avec cette app ComfyUI</a> provient du site <a href="https://www.h2-blog.com">H2-Blog</a>.</p>]]></description>
				<content:encoded><![CDATA[<p>J&rsquo;essayais de poster des images populaires sur les réseaux sociaux pour tester ComfyUI dans la vraie vie. Ma première idée ne fonctionnait pas, alors j&rsquo;ai cherché quelles étaient les images populaires sur X. Il s&rsquo;avère que ce sont les images SFW de personnages d&rsquo;animés japonais (femme ou homme transformé en femme).</p>
<h2>Comment j’ai eu 300K vues sur Twitter</h2>
<p>J’ai remarqué ce tweet de <a href="https://x.com/NaruDiffusion/status/2091590748105736326">Narukami (@NaruDiffusion)</a> sur X qui avait 100k vues. C’est le personnage Kenjaku de Jujutsu Kaisen, un homme transformer en femme.</p>
<p>Donc, j&rsquo;ai utilisé mon application, ajouté l&rsquo;image du tweet, ainsi que le nom du personnage et celui de la série.</p>
<p><img class="aligncenter size-large wp-image-11840" src="http://www.h2-blog.com/wp-content/uploads/2026/10/Capture-d’écran-2026-10-08-à-13.36.33-650x461.png" alt="APP comfyui" width="650" height="461" /></p>
<p>J’ai publié cette image sur X sur mon compte <a href="https://x.com/AndroidraFr/status/2092466035345576067">Androidra (@AndroidraFR) </a>et j’ai eu 300k vues en 2 jours. Ce tweet à était principalement populaire au japon and en Amérique (Nord et Sud).</p>
<p><img class="aligncenter size-large wp-image-11839" src="http://www.h2-blog.com/wp-content/uploads/2026/10/Capture-d’écran-2026-10-08-à-17.44.07-650x379.png" alt="Capture d’écran 2026-10-08 à 17.44.07" width="650" height="379" /></p>
<p>Ne me demandez pas pourquoi ce type d&rsquo;image est populaire sur les réseaux sociaux ; en tant qu&rsquo;Européen, cela m&rsquo;est incompréhensible.</p>
<h2>Qui à-il à l’intérieure de cette application.</h2>
<p>Mon application comfyui est basée sur deux flux de travail : <a href="https://comfy.org/fr/workflows/llm_qwen3_5_text_gen-aeb71e383642/">Qwen 3.5</a> qui créer un partie du prompt et <a href="https://comfy.org/fr/workflows/ab2f354cd396-ab2f354cd396/">Anima 1</a>. Je recommande vivement de tester ces deux flux de travail séparément afin de vous éviter des maux de crâne..</p>
<p>L&rsquo;idée derrière cette application n&rsquo;est pas de copier-coller des images populaires des réseaux sociaux, mais plutôt de laisser Qwen 3.5 halluciner pour obtenir de bons résultats. Je ne sais pas pourquoi, mais ce fonctionnement donne de très bons résultats avec Anima.</p>
<p>Il n&rsquo;y a ni plugin ni nœuds (code) personnalisés dans cette application. Tout est natif, donc moins de risques de faille de sécurité et moins de risques de casser ComfyUI lors d&rsquo;une mise à jour</p>
<h2>Télécharger l’application</h2>
<p>Cette application est disponible sur <a href="https://civitai.red/models/2991623/300k-views-on-x-with-this-comfyui-app-anima">Civitai Red </a>et <a href="https://github.com/H2logs/300K-view-ComfyUI-APP/blob/main/300K_APP_ANiMA_h2logs-substack-com.json">GitHub</a>.</p>
<h2>Comment utiliser l’application</h2>
<p>Entrez le nom du personnage, le nom de la série, ajouter une image and appuyer sur <strong>Exécuter</strong>.</p>
<p>Paramètres recommandés : LLM seed +1, batch size 4, turbo mode, Nombre d’exécution 4, Auto-size, max height 1280.</p>
<p><strong>Attention :</strong> si vous voulez publier les images sans les vérifier, je vous recommande de désactiver le mode turbo. Cela réduira le risque de publier du contenu inapproprié.</p>
<h2>Explication des paramètres de L’APP</h2>
<p>Character name: Vous pouvez entrer plusieurs nom de personnage (ex : Yor Forger left and Fiona Frost right).</p>
<p>Auto size ON: Utilisez les dimension de l’image source ; le hauteur maximum est apliquée pour réduire les temps de généation.</p>
<p>Auto size OFF: utilise la largeur et la hauteur que vous avez entrées manuellement</p>
<p>Character Swap: Demande au LLM de ne pas mentionné les cheveux, bijoux et couleur de peau.</p>
<p>Force curvy woman : demande au LLM de décrire le personnage comme une femme aux formes généreuses.</p>
<h2>Recommandation pour les réseaux sociaux.</h2>
<p>Ne créez pas de nouveau compte sur les réseaux sociaux. Sur la plupart d&rsquo;entre eux, les nouveaux comptes sont shadow ban.</p>
<p>Utilisez plutôt un compte inactif, même s&rsquo;il n&rsquo;a pas d&rsquo;abonnés (ne changez pas le nom ni l&rsquo;alias), et commencez à publier des images pour voir si cela rencontre du succès. Essayez de poster une image tous les deux jours pour réduire le risque d&rsquo;être shadow ban. Si cela ne fonctionne pas après 10 images, essayez une autre idée.</p>
<p>Vous pouvez aussi acheter un ancien compte en ligne pour environ 5 € ; vous n&rsquo;avez pas besoin de payer plus pour un compte avec des abonnés. Mon ancien compte Twitter avait 500 abonnés, dont la plupart devaient être des comptes inactifs.</p>
<p>Acheter X Premium n&rsquo;augmente pas le nombre de vues : j&rsquo;ai eu 300 000 vues sans X Premium.</p>
<p>Cette application Comfy fonctionne aussi sur les autres réseaux sociaux, pas seulement sur Twitter/X.</p>
<h2>C’est quoi ComfyUI</h2>
<p>Pour ceux qui sont au fond de la classe : ComfyUI est une application (logiciel sur Windows) qui permet de faire fonctionner des IA pour générer des images ou des vidéos avec des cartes graphiques grand public ou professionnelles (5070 Ti, 5060 Ti 16 Go…). En termes d&rsquo;interface graphique, elle fonctionne sur le principe du flux de travail, ce qui permet notamment de combiner plusieurs IA pour réaliser des tâches plus avancées et surtout de gagner du temps.</p>
<h2>Licences</h2>
<p>Qwen 3.5: <a href="https://huggingface.co/datasets/choosealicense/licenses/blob/main/markdown/apache-2.0.md">Apache license 2.0</a></p>
<p>Anima: <a href="https://huggingface.co/circlestone-labs/Anima/blob/main/LICENSE.md">circlestone-labs-non-commercial-license</a></p>
<p>Bonne chance.</p>
<p>Cet article <a href="https://www.h2-blog.com/2026/10/18/300-000-vues-twitterx-app-comfyui/">300 000 vues sur Twitter/x avec cette app ComfyUI</a> provient du site <a href="https://www.h2-blog.com">H2-Blog</a>.</p>]]></content:encoded>
			<wfw:commentRss>https://www.h2-blog.com/2026/10/18/300-000-vues-twitterx-app-comfyui/feed/</wfw:commentRss>
		<slash:comments>0</slash:comments>
		</item>
		<item>
		<title>Qwen Image 2.1 turbo int8 flux de travail : Texte vers image et Edition</title>
		<link>https://www.h2-blog.com/2026/10/13/qwen-image-2-1-turbo-int8-flux-travail-texte-vers-image-edition/</link>
		<comments>https://www.h2-blog.com/2026/10/13/qwen-image-2-1-turbo-int8-flux-travail-texte-vers-image-edition/#comments</comments>
		<pubDate>Tue, 13 Oct 2026 20:57:07 +0000</pubDate>
		<dc:creator><![CDATA[Guillaume R]]></dc:creator>
				<category><![CDATA[ComfyUI]]></category>

		<guid isPermaLink="false">http://www.h2-blog.com/?p=11825</guid>
		<description><![CDATA[<p>Qwen Image 2.1 est le dernier model populaire sur ComfyUI, ce model permet la création d’image et l’édition d’image. Mais ce modèle de génération d’image par IA a vraiment besoin d’un boost de vitesse de type turbo. Pour créer ce flux de travail vous devez télécharger la version Turbo INT8.convrot de Qwen Image 2.1 sur [&#8230;]</p><p>Cet article <a href="https://www.h2-blog.com/2026/10/13/qwen-image-2-1-turbo-int8-flux-travail-texte-vers-image-edition/">Qwen Image 2.1 turbo int8 flux de travail : Texte vers image et Edition</a> provient du site <a href="https://www.h2-blog.com">H2-Blog</a>.</p>]]></description>
				<content:encoded><![CDATA[<p>Qwen Image 2.1 est le dernier model populaire sur ComfyUI, ce model permet la création d’image et l’édition d’image. Mais ce modèle de génération d’image par IA a vraiment besoin d’un boost de vitesse de type turbo.</p>
<p><img class="aligncenter size-large wp-image-11827" src="http://www.h2-blog.com/wp-content/uploads/2026/10/Capture-d’écran-2026-10-13-à-14.28.25-650x241.png" alt="Capture d’écran 2026-10-13 à 14.28.25" width="650" height="241" /></p>
<p>Pour créer ce flux de travail vous devez télécharger la version Turbo INT8.convrot de Qwen Image 2.1 sur <a href="https://huggingface.co/Comfy-Org/Qwen-Image-2.1/tree/main/diffusion_models">HuggingFace</a> and ajouter ce fichier au dossier diffusion_models. Changer le UNet_name et changez le nombre d’étapes à 8.</p>
<p>Performance (1024&#215;1024) Texte vers image : environ 4 seconds pour une deuxième génération avec une RTX 5070 Ti ; environ 8 seconds avec une RTX 5060 Ti 16GB.</p>
<p><img class="aligncenter size-large wp-image-11828" src="http://www.h2-blog.com/wp-content/uploads/2026/10/QWEN2.1-EDIT-650x325.jpg" alt="QWEN2.1-EDIT" width="650" height="325" /></p>
<p>Le prompt : elle porte une casquette « obey » et des lunettes de soleil.</p>
<p>Pour le flux de travail ce sont les même modification que précédemment.</p>
<p><img class="aligncenter size-large wp-image-11829" src="http://www.h2-blog.com/wp-content/uploads/2026/10/Capture-d’écran-2026-10-13-à-14.35.39-650x266.png" alt="Capture d’écran 2026-10-13 à 14.35.39" width="650" height="266" /></p>
<p>Performance (1024&#215;1024) édition d’image : environ 7 seconds pour une deuxième génération avec une RTX 5070 Ti ; environ 14 seconds avec une RTX 5060 Ti 16GB.</p>
<p>Ces flux de travail sont sur <a href="https://civitai.com/user/def813/models?sort=Highest+Rated&amp;tag=base+model">Civitai</a> et <a href="https://github.com/H2logs/Qwen-Image-2.1-turbo-int8-workflow">Github</a>.</p>
<p>License : <a href="https://huggingface.co/Qwen/Qwen-Image-2.1/blob/main/LICENSE">Qwen Research License Agreement</a></p>
<p>Cet article <a href="https://www.h2-blog.com/2026/10/13/qwen-image-2-1-turbo-int8-flux-travail-texte-vers-image-edition/">Qwen Image 2.1 turbo int8 flux de travail : Texte vers image et Edition</a> provient du site <a href="https://www.h2-blog.com">H2-Blog</a>.</p>]]></content:encoded>
			<wfw:commentRss>https://www.h2-blog.com/2026/10/13/qwen-image-2-1-turbo-int8-flux-travail-texte-vers-image-edition/feed/</wfw:commentRss>
		<slash:comments>0</slash:comments>
		</item>
		<item>
		<title>Flux 1 dev FP8 testé sur 13 cartes graphiques avec ComfyUI</title>
		<link>https://www.h2-blog.com/2026/10/05/flux-1-dev-fp8-teste-13-cartes-graphiques-comfyui/</link>
		<comments>https://www.h2-blog.com/2026/10/05/flux-1-dev-fp8-teste-13-cartes-graphiques-comfyui/#comments</comments>
		<pubDate>Mon, 05 Oct 2026 19:27:09 +0000</pubDate>
		<dc:creator><![CDATA[Guillaume R]]></dc:creator>
				<category><![CDATA[ComfyUI]]></category>

		<guid isPermaLink="false">http://www.h2-blog.com/?p=11814</guid>
		<description><![CDATA[<p>Quand on chercher à acheter une carte graphique pour ComfyUI il est très difficile de trouver des données qui permettent se savoir réellement quel est la performance de cette carte graphique. Pourquoi ces tests J’étais sur le point d’acheter une A4500 mais je n’étais pas sûre de ces performance pour la génération d’image avec ComfyUI. [&#8230;]</p><p>Cet article <a href="https://www.h2-blog.com/2026/10/05/flux-1-dev-fp8-teste-13-cartes-graphiques-comfyui/">Flux 1 dev FP8 testé sur 13 cartes graphiques avec ComfyUI</a> provient du site <a href="https://www.h2-blog.com">H2-Blog</a>.</p>]]></description>
				<content:encoded><![CDATA[<p>Quand on chercher à acheter une carte graphique pour ComfyUI il est très difficile de trouver des données qui permettent se savoir réellement quel est la performance de cette carte graphique.</p>
<h2>Pourquoi ces tests</h2>
<p>J’étais sur le point d’acheter une A4500 mais je n’étais pas sûre de ces performance pour la génération d’image avec ComfyUI. J’ai décidé de louer un shadow PC pour test cette A4500, les résultat ne m’ont pas convaincu. Donc, j’ai passé plusieurs heures sur le site web Trooper.ai pour tester toutes les carte graphique qu’il propose à la location.</p>
<p><img class="aligncenter size-large wp-image-11821" src="http://www.h2-blog.com/wp-content/uploads/2026/10/A100_4090_4090_pro_4500_blackwell_4080_super_pro_5070ti_4070_ti_super_4000_blackwell_3090_A4500_a4000_4000_ada_sff_5060ti_16g_comfyui-650x412.png" alt="A100 4090 4090 pro 4500 blackwell 4080 super pro 5070ti 4070 ti super 4000 blackwell 3090 A4500 a4000 4000 ada sff 5060ti 16g comfyui" width="650" height="412" /></p>
<h2>Où est le flux de travail ?</h2>
<p>Si vous voulez tester ce flux de travail sur votre propre carte graphique alors vous trouverez sur <a href="https://trooper.cloud/support_dls.php?f=simple-cute-flux-workflow-trooperai.json">tropper.ai</a> gratuitement.</p>
<p>Prompt : Two cute maltese puppy in gree grass with a colorful hat, sitting in a small wheelbarrow. Largeur : 1152px, Hauteur : 896px, 30 étapes.</p>
<p>Pour télécharger les fichier nécessaire à ce flux de travail vous pouvez vous rendre sur le site web de <a href="https://docs.comfy.org/tutorials/flux/flux-1-text-to-image">documentation de Comfy</a>.</p>
<p>Note : je n’ai aucun liens de quelque sorte que ce soir avec Trooper.ai, j’ai même fini par acheter une 5070ti qui n’est pas disponible sur leur site web.</p>
<h2>Combien ça coute de faire ces tests</h2>
<p>SI on ignore mes cartes graphiques (5060ti, 5070ti, 4000 ada sff), cela m’a couté 70€, 50€ de Shadow PC et 20€ de Trooper.ai. Cela m’a seulement couté 70€ avant d’acheter une carte graphique à 880€.</p>
<p>Avant d’acheter une carte graphique le meilleur conseil est de la louer afin de la tester pour se faire une idée de ces performances. La plupart du temps cela vous coutera quelque chose comme 10€.</p>
<h2>Remarques sur les cartes graphiques</h2>
<p>La 4080 Super Pro et la 4090 Pro sont des cartes graphiques modifiées afin de doubler leur quantité de Vram.</p>
<p>Le bios silencieux sur la 5060ti augmente la vitesse avec ComfyUI de 10% pour la carte ASUS Dual GeForce RTX™ 5060 Ti 16GB GDDR7 OC Edition.</p>
<h2>La liste des cartes graphiques</h2>
<p>Mes carte graphiques : 5060ti, 5070ti, 4000 ada sff avec un ryzen 7700 et 96g de ram DDR5 (acheté 230€ avant augmentation des prix).</p>
<p>Shasow PC : A4500.</p>
<p>Troper.AI : A100, 4090, 4090 pro, 4500 blackwell, 4080 super pro, 4070 ti super, 4000 blackwell, 3090, a4000.</p>
<p>Bonne chasse à la carte graphique.</p>
<p>Version anglaise de l’article : <a href="https://h2logs.substack.com/p/flux-1-dev-fp8-comfyui-benchmark">Flux 1 dev FP8 ComfyUI benchmark across 13 GPUs</a></p>
<p>Cet article <a href="https://www.h2-blog.com/2026/10/05/flux-1-dev-fp8-teste-13-cartes-graphiques-comfyui/">Flux 1 dev FP8 testé sur 13 cartes graphiques avec ComfyUI</a> provient du site <a href="https://www.h2-blog.com">H2-Blog</a>.</p>]]></content:encoded>
			<wfw:commentRss>https://www.h2-blog.com/2026/10/05/flux-1-dev-fp8-teste-13-cartes-graphiques-comfyui/feed/</wfw:commentRss>
		<slash:comments>0</slash:comments>
		</item>
	</channel>
</rss>
