Model
Trends
.ai
Model
Trends
.ai
model AI open source terbaik
Jalankan secara lokal
- Flux1 - soothing_atmo_V2.0.safetensors
76 MB · SafeTensor
Unduhan disediakan oleh Civitai. Beberapa kreator meminta login Civitai terlebih dahulu.
AD
Pasang / Jalankan di Graydient
Memerlukan keanggotaan
<soothing-flux>
Tentang Model Ini
FLUX · LoRA
Dibuat oleh
TijuanaSlumlord
di Civitai
by TijuanaSlumlord - Really wish they would put out the Pro version of Flux since this distilled stuff is kinda hard to control and also limited in its flexibility. If you play around with some prompts it always goes to a certain image or something it knows good and is trained on, which results in Flux typical steril look cinematic, photo, certain animals or people etc
Catatan pembuat
zImageBase - V1.0 After quite a few test with different settings, including theory's without any proof, decided to just change all training settings till something works. Did a small test with only 6 images, different styles, photo, illustration etc. After 500 steps it pretty much nailed all images, low style bleeding if any, no flip-flopping of characters changing position, reduced hallucinations, learns faster ( well, it's slow but constant ) and the Lora has more influence even with lower strength ( in some cases you had to crank up the strength over 1.0 which induces artifacts ) This one is just another test with 20 images. Old images with a lot of junk in them, which was wanted to see if it tries to compensate mistakes or learns what is given. This is a version with lower step count. The versions with higher steps where a bit too close to the source images and made the output very rigged even with different seeds. Even this one is kinda strong. I don't trust it fully yet. Currently training stuff with a bigger dataset which takes some time. - AI-Toolkit - Prodigy ( non 8bit, at least i think, if i read the optimizer files correctly ) - Rank 64 ( haven't tested lower ranks yet, just went with it ) - LR: 1 ( Prodigy changes the learning rate dynamically ) - noise_scheduler: "flowmatch" optimizer: "prodigy" timestep_type: "shift" content_or_style: "balanced" optimizer_params: weight_decay: 0.01 ( still needs testing if needed at all, changed it a few times ) d_coef: 1.0 ( also needs testing, changed it to 1.5 for the bigger dataset for now ) decouple: true use_bias_correction: false Take all of this with a big grain of salt ... actually buy a whole salt-mine. With AI it seems no one knows how and why shit works or not. Even the people who made the models don't know what the fuck is going on, all just theory and math. High chance that i'm on the wrong path too, wouldn't be the first time. It did what it's supposed to do for now, so i'll stick with it till problems occur. V.2.0 Added more pictures and higher step count. Recommended to use a low strength or you turn everything into a tentacle monster for whatever reason. Really wish they would put out the Pro version of Flux since this distilled stuff is kinda hard to control and also limited in its flexibility. If you play around with some prompts it always goes to a certain image or something it knows good and is trained on, which results in Flux typical steril look ( cinematic, photo, certain animals or people etc ) I also highly doubt that any of the full checkpoints out there that are trained will ever work. Went trough pretty much every single one of them, but it always comes down to fucked up anatomy or a hardcore bias which you can't negate with negative prompts or prompt weighting since Flux doesn't use this stuff. A picture of a person that is always naked, even if you ask for clothes for example. Doubtful if that even did/does anything in SDXL, never really used that much negative prompts, but it was good enough to get rid of some things you tried to avoid. Strangely enough, all checkpoints that are trained or LoRAfied with a style ( like anime ) work great, even for realistic images, which i pretty much use for all of my pictures ( not here, basic FP8 checkpoint for showcase reason )... so confusing. Trained on a dataset which i planned to use for SDXL, but i never got satisfying results . Just a small test with a few images ( basic capitations for now ) and only 800 steps. Will change that later to natural language. Most used words should be atmospheric, moody, calm, soothing, serene, mysterious ... and assorted ( the images without capitation ... well, that is a capitation, forgot that Kohya picks the name of the folder if no .txt file is present :D ) - Order: First 2 Images with LoRA / without LoRA, after that, reversed order - Has more/less impact in specific cases ( for now ) Did quite some testing with Flux LoRA's I've made and got really crazy results. Even after only like 100 - 200 steps it got the concept and lower/higher strength from base 1 has always a huge impact, but something is always left and it feels like you can kinda pick specific parts from the images it was trained on without it using the whole image ( like you only want the the yellow clothing but nothing else ) Made like 30 LoRA's so far ( only for testing purposes ) and what you can do with just a few images is baffling. Might be just a fluke, who knows.
5,809 unduhan · 543 suka di Civitai
SHA256
CD7599ECAF0A68ECD08E15615DD28E0002E07FCA76EA3BFAB46E542771B0C249
ID Model ModelTrends.ai
#15905
Laporkan konten
Contoh render
Prompt dan pengaturan yang dibagikan oleh orang-orang yang membuat render ini di Civitai. Pilih sebuah render untuk melihat cara membuatnya.
Pengaturan
Sampler
Euler
Langkah
30
Guidance
1
Seed
2745709515
Ukuran
896x1152
Prompt
hyperrealistic art <lora:FluxMythD1gitalL1nes:1.4> D1gitalL1nes, Night in the garden of the samurai castle. Moonlight breaks through the branches of cherry blossoms. In the background, there are dark silhouettes of towers and andon lanterns. <lora:- Flux1 - soothing_atmo_V2.0:1> atmospheric, moody, mysterious, assorted, Ninja (Attacking): At the highest point of a powerful jump, diving down at an angle of 45 degrees. The body is stretched into a string, the legs are grouped for impact. Classic black shinobi shozoku (ninja costume) with a mask covering the entire face except the eyes. The fabric is torn by the wind. A short sharp kunai or tanto (ninja dagger) is held in the right hand in a reverse grip (the blade along the forearm), pointing the tip down at the target. The blade glitters coldly in the moonlight. incredible speed and power of the jump. Behind him is a plume of dust and falling sakura petals, lifted by a push. Sharp focus on the ninja especially on the dagger and the faces. The background is slightly blurred, Contrasting. Bright moonlight illuminates the stage from above, creating harsh shadows. The warm glow of the andon lanterns illuminates the general from below. Flying sakura petals, dust, perhaps sparks from the incipient movement of the general's sword. Textures of fabric, metal of armor and wood of scabbard. Epic, dynamic, in the style of action anime (senen) or historical manga. The emphasis is on speed, danger and skill of warriors. . extremely high-resolution details, photographic, realism pushed to extreme, fine texture, incredibly lifelike
Prompt negatif
uncensored, clothes on legs, double hand, panties, overexposed, blurred details, subtitles, style, artwork, painting, picture, still, overall gray, worst quality, low quality, JPEG compression residue, ugly, incomplete, extra fingers, poorly drawn hands, poorly drawn faces, deformed, disfigured, deformed limbs, fused fingers, still picture, cluttered background, three legs, many people in the background, walking backwards, zoom in, zoom out, talking, (((small breast size))), fat, big pussy, ((girl without head on image)), tattoo, nipples on clothes,, simplified, abstract, unrealistic, impressionistic, low resolution
Pengaturan
Sampler
DPM++ 2M
Langkah
40
Guidance
3.5
Seed
897232042303832
Ukuran
832x1216
Clip skip
1
Prompt
A bioluminescent alien with soft, jellyfish-like features, glowing in shades of pink and blue, and flowing tendrils cascading like hair <lora:Dever_Flux_Enhancer.safetensors:0.5>, <lora:Giger_2_0.safetensors:0.5>, <lora:Movie_Portrait.safetensors:0.4>, <lora:- Flux1 - soothing_atmo_V2.0.safetensors:0.2>, <lora:lith_mimosa_kiel_grw134_FLUX.safetensors:0.35>, <lora:black_fantasy.safetensors:0.5>,
Pengaturan
Sampler
Undefined
Langkah
34
Guidance
3.5
Seed
1963045758
Ukuran
832x1216
Clip skip
2
Bobot yang dipakai
0.4
Prompt
A hyperdetailed hyperrealistic moody photorealistic Portrait of a Woman, an analog Photography with deep depth that feels Like A surreal and hyperdetailed Monochrome Line artwork made of lines, created entirely from unbroken bold, flowing thick lines, forming an optical illusion. The image depicts a close-up portrait of a faceless ethereal being, with a busty figure and long, flowing hair seamlessly integrated into the linework. The head is a smooth, black shiny polished orb in place of a face, surrounded by black and white hues. The lines weave together to create intricate patterns and cinematic depth, drawing the viewer's eye through the composition with hypnotic precision, resembling a cyclops-like presence.
Pengaturan
Sampler
Undefined
Langkah
42
Guidance
3.5
Seed
3845664927
Ukuran
832x1216
Clip skip
2
Bobot yang dipakai
0.3
Prompt
The image is a melancholic photograph of a misty landscape, where the sky is shrouded in dark, brooding clouds that cast a heavy atmosphere over the scene. A large, still body of water stretches across the foreground, reflecting the dull greys of the overcast sky, while distant mountains fade into the mist, their jagged peaks barely visible. On the left side of the image stands a large yellow bird puppet, but its once cheerful, fluffy appearance contrasts sharply with its current mood. The puppet gazes blankly into the distance, its vibrant yellow feathers now dim under the oppressive clouds. Its beak is slightly open, as if lost in a moment of existential questioning, wondering about the meaning of its very existence in this desolate, bleak world. The dark clouds swirl above as if they too are burdened by the weight of uncertainty, adding a sense of impending dread to the scene.
Pengaturan
Sampler
DPM++ 2M
Langkah
40
Guidance
3.5
Seed
40807763792657
Ukuran
832x1216
Clip skip
1
Prompt
Desert nomad, flowing beige robes, veiled face, glowing dagger, sandstorm backdrop <lora:Elden_Ring_-_Yoshitaka_Amano.safetensors:1.0>, <lora:aiai-Glimra-Flux-Dev-v2.safetensors:0.5>, <lora:Movie_Portrait.safetensors:0.4>, <lora:- Flux1 - soothing_atmo_V2.0.safetensors:0.2>, <lora:lith_mimosa_kiel_grw134_FLUX.safetensors:0.35>, <lora:Instant camera portrait photography V2.0.safetensors:0.5>,
Pengaturan
Sampler
Undefined
Langkah
34
Guidance
3.5
Seed
489771649
Ukuran
832x1216
Clip skip
2
Bobot yang dipakai
0.4
Prompt
In the foreground lays the whimsical, decaying, moss-covered, and weathered metal body of a long-defunct robot, its rusted frame blending completely into the natural environment. Beside it, an old, half-rotted sign lies on the ground, bearing the faded text "some buzz pls," hinting at its past use as a begging shield. The robot’s head, overgrown with vines, gives an eerie sense of abandonment and forgotten technology. Behind this rusted relic, a massive abyss hovers in the vastness of outer space, its colossal form seemingly bowing down to the viewer. Two towering structures, forged from black metal, rise ominously into the high clouds. Their sleek, modern-medieval design incorporates both sharp geometric angles and massive windows that reflect the fading light. A bridge of cold, gleaming metal connects the two towers, reinforcing their imposing presence in the sky. The hyper-realistic scene captures a tranquil, mist-shrouded landscape at twilight, with rolling hills and deep valleys bathed in soft orange light. The terrain is covered with delicate shrubs and grasses, their leaves tinged in deep black-pink hues, creating a striking contrast against the cool, misty atmosphere. The decaying robot’s weathered form, set against the towering black metal structures, evokes a haunting sense of the passage of time, where nature slowly reclaims the remnants of a long-lost civilization.
Pengaturan
Sampler
DPM++ 2M
Langkah
30
Guidance
3.5
Seed
2034826968
Ukuran
2496x3648
Bobot yang dipakai
0.4
Prompt
A hyper-realistic, unsettling close-up portrait fusing two enigmatic figures into a single, otherworldly being. The spectral, all-knowing cyborg emerges with a smooth, glowing surface of advanced technology, its face featureless yet alive with intricate digital patterns. Electric blue lines pulse across its surface, shifting and morphing like a living neural network, suggesting an existence within a vast interconnected web. Hollow eyes radiate an eerie, ethereal quality, as if seeing beyond the physical realm, and glowing circuits blend seamlessly into its surroundings, a deep gradient from white to pale blue filled with subtle digital glitches hinting at higher intelligence. From the shadows of this cybernetic form, the rugged presence of a slightly muscular man emerges as a faint duality. His unkempt hair falls like strands of static across the glowing face, revealing one piercing eye that gleams with intensity. A sly, mischievous grin hints at a dangerous charisma, contrasting with the pastel-colored T-shirt and oversized vintage black headphones resting heavily against the figure’s sleek, technological frame. The pale monitor glow casts an eerie light over both organic and digital textures, blending the rugged and the spectral into a single captivating, otherworldly presence. The result is an unnerving harmony of man and machine, a haunting observer that feels alive yet unreachably distant.
Model AI untuk Pembuatan Gambar, Video & Teks
soothing flux adalah LoRA untuk keluarga Flux yang tercantum di ModelTrends.ai, katalog hanya-baca untuk model AI gambar, video, dan teks open source. Bandingkan dengan model Flux lainnya, lihat skor popularitasnya untuk mengetahui trennya, dan buka di PirateDiffusion atau BitVector untuk mencobanya.
Jelajahi berdasarkan Keluarga Model
Jelajahi semua model