nat-wanderer/blog-for-nat 🖼️🔢❓📝✓ → 🖼️
About

Example Output
"
TOK, cinematic ultra-detailed photorealistic side-profile portrait of a young woman, shown from the shoulders up. She is nude, her skin smooth and flawless, softly illuminated by cool, blue-toned professional studio lighting. Her profile is elegant and sculptural, and the light subtly defines the line of her neck, collarbone, and cheek. Her hair is styled in voluminous, glossy curls that cascade down and blend into the shadows, catching subtle cool reflections.
Large, sharp-edged glass shards are positioned across her upper body and in her hair — dramatic, crystalline pieces with jagged, angular forms, like broken windows and crystals. There are a lot of them - the biggest ones look like a crown. These shards are thick and translucent, refracting the icy blue light with striking clarity. Their placement suggests a surreal fusion of vulnerability and danger: embedded gently but prominently along her shoulder, collarbone, neck, and tangled elegantly within her curls. The sharp points and fractured edges sparkle as they catch the light, creating an intense interplay of glow and shadow.
The background is almost completely black, with a deep midnight blue undertone that softly gradients into the darkness. There are no objects or background elements — only the woman, the glass, and the atmosphere. The studio lighting is carefully directed: soft fill from the front, and sharp edge light from behind, all tinted cool blue, making her form glow softly while the glass reflects with dramatic brilliance.
The overall mood is haunting, surreal, and visually striking — like a fashion editorial from a dark fantasy world. The sharp, large shards against bare skin add an emotional charge: beauty meets danger, fragility meets defiance.
Negative prompt additions: no warm light, no background objects, no clothing, no jewelry, no food, no kitchen, no text, no logos, no clutter, no extra limbs, no distortion, no softness in glass — shards must be large and sharp.
"Output

Performance Metrics
All Input Parameters
{ "model": "dev", "prompt": "TOK, cinematic ultra-detailed photorealistic side-profile portrait of a young woman, shown from the shoulders up. She is nude, her skin smooth and flawless, softly illuminated by cool, blue-toned professional studio lighting. Her profile is elegant and sculptural, and the light subtly defines the line of her neck, collarbone, and cheek. Her hair is styled in voluminous, glossy curls that cascade down and blend into the shadows, catching subtle cool reflections.\n\nLarge, sharp-edged glass shards are positioned across her upper body and in her hair — dramatic, crystalline pieces with jagged, angular forms, like broken windows and crystals. There are a lot of them - the biggest ones look like a crown. These shards are thick and translucent, refracting the icy blue light with striking clarity. Their placement suggests a surreal fusion of vulnerability and danger: embedded gently but prominently along her shoulder, collarbone, neck, and tangled elegantly within her curls. The sharp points and fractured edges sparkle as they catch the light, creating an intense interplay of glow and shadow.\n\nThe background is almost completely black, with a deep midnight blue undertone that softly gradients into the darkness. There are no objects or background elements — only the woman, the glass, and the atmosphere. The studio lighting is carefully directed: soft fill from the front, and sharp edge light from behind, all tinted cool blue, making her form glow softly while the glass reflects with dramatic brilliance.\n\nThe overall mood is haunting, surreal, and visually striking — like a fashion editorial from a dark fantasy world. The sharp, large shards against bare skin add an emotional charge: beauty meets danger, fragility meets defiance.\n\nNegative prompt additions: no warm light, no background objects, no clothing, no jewelry, no food, no kitchen, no text, no logos, no clutter, no extra limbs, no distortion, no softness in glass — shards must be large and sharp.", "go_fast": false, "lora_scale": 1, "megapixels": "1", "num_outputs": 1, "aspect_ratio": "1:1", "output_format": "webp", "guidance_scale": 3, "output_quality": 80, "prompt_strength": 0.8, "extra_lora_scale": 1, "num_inference_steps": 28 }
Input Parameters
- mask
- Image mask for image inpainting mode. If provided, aspect_ratio, width, and height inputs are ignored.
- seed
- Random seed. Set for reproducible generation
- image
- Input image for image to image or inpainting mode. If provided, aspect_ratio, width, and height inputs are ignored.
- model
- Which model to run inference with. The dev model performs best with around 28 inference steps but the schnell model only needs 4 steps.
- width
- Width of generated image. Only works if `aspect_ratio` is set to custom. Will be rounded to nearest multiple of 16. Incompatible with fast generation
- height
- Height of generated image. Only works if `aspect_ratio` is set to custom. Will be rounded to nearest multiple of 16. Incompatible with fast generation
- prompt (required)
- Prompt for generated image. If you include the `trigger_word` used in the training process you are more likely to activate the trained object, style, or concept in the resulting image.
- go_fast
- Run faster predictions with model optimized for speed (currently fp8 quantized); disable to run in original bf16
- extra_lora
- Load LoRA weights. Supports Replicate models in the format <owner>/<username> or <owner>/<username>/<version>, HuggingFace URLs in the format huggingface.co/<owner>/<model-name>, CivitAI URLs in the format civitai.com/models/<id>[/<model-name>], or arbitrary .safetensors URLs from the Internet. For example, 'fofr/flux-pixar-cars'
- lora_scale
- Determines how strongly the main LoRA should be applied. Sane results between 0 and 1 for base inference. For go_fast we apply a 1.5x multiplier to this value; we've generally seen good performance when scaling the base value by that amount. You may still need to experiment to find the best value for your particular lora.
- megapixels
- Approximate number of megapixels for generated image
- num_outputs
- Number of outputs to generate
- aspect_ratio
- Aspect ratio for the generated image. If custom is selected, uses height and width below & will run in bf16 mode
- output_format
- Format of the output images
- guidance_scale
- Guidance scale for the diffusion process. Lower values can give more realistic images. Good values to try are 2, 2.5, 3 and 3.5
- output_quality
- Quality when saving the output images, from 0 to 100. 100 is best quality, 0 is lowest quality. Not relevant for .png outputs
- prompt_strength
- Prompt strength when using img2img. 1.0 corresponds to full destruction of information in image
- extra_lora_scale
- Determines how strongly the extra LoRA should be applied. Sane results between 0 and 1 for base inference. For go_fast we apply a 1.5x multiplier to this value; we've generally seen good performance when scaling the base value by that amount. You may still need to experiment to find the best value for your particular lora.
- replicate_weights
- Load LoRA weights. Supports Replicate models in the format <owner>/<username> or <owner>/<username>/<version>, HuggingFace URLs in the format huggingface.co/<owner>/<model-name>, CivitAI URLs in the format civitai.com/models/<id>[/<model-name>], or arbitrary .safetensors URLs from the Internet. For example, 'fofr/flux-pixar-cars'
- num_inference_steps
- Number of denoising steps. More steps can give more detailed images, but take longer.
- disable_safety_checker
- Disable safety checker for generated images.
Output Schema
Output
Example Execution Logs
Weights already loaded Loaded LoRAs in 0.02s Using seed: 34423 Prompt: TOK, cinematic ultra-detailed photorealistic side-profile portrait of a young woman, shown from the shoulders up. She is nude, her skin smooth and flawless, softly illuminated by cool, blue-toned professional studio lighting. Her profile is elegant and sculptural, and the light subtly defines the line of her neck, collarbone, and cheek. Her hair is styled in voluminous, glossy curls that cascade down and blend into the shadows, catching subtle cool reflections. Large, sharp-edged glass shards are positioned across her upper body and in her hair — dramatic, crystalline pieces with jagged, angular forms, like broken windows and crystals. There are a lot of them - the biggest ones look like a crown. These shards are thick and translucent, refracting the icy blue light with striking clarity. Their placement suggests a surreal fusion of vulnerability and danger: embedded gently but prominently along her shoulder, collarbone, neck, and tangled elegantly within her curls. The sharp points and fractured edges sparkle as they catch the light, creating an intense interplay of glow and shadow. The background is almost completely black, with a deep midnight blue undertone that softly gradients into the darkness. There are no objects or background elements — only the woman, the glass, and the atmosphere. The studio lighting is carefully directed: soft fill from the front, and sharp edge light from behind, all tinted cool blue, making her form glow softly while the glass reflects with dramatic brilliance. The overall mood is haunting, surreal, and visually striking — like a fashion editorial from a dark fantasy world. The sharp, large shards against bare skin add an emotional charge: beauty meets danger, fragility meets defiance. Negative prompt additions: no warm light, no background objects, no clothing, no jewelry, no food, no kitchen, no text, no logos, no clutter, no extra limbs, no distortion, no softness in glass — shards must be large and sharp. [!] txt2img mode 0%| | 0/28 [00:00<?, ?it/s] 4%|▎ | 1/28 [00:00<00:07, 3.83it/s] 7%|▋ | 2/28 [00:00<00:05, 4.34it/s] 11%|█ | 3/28 [00:00<00:06, 4.10it/s] 14%|█▍ | 4/28 [00:00<00:06, 3.98it/s] 18%|█▊ | 5/28 [00:01<00:05, 3.93it/s] 21%|██▏ | 6/28 [00:01<00:05, 3.91it/s] 25%|██▌ | 7/28 [00:01<00:05, 3.89it/s] 29%|██▊ | 8/28 [00:02<00:05, 3.87it/s] 32%|███▏ | 9/28 [00:02<00:04, 3.87it/s] 36%|███▌ | 10/28 [00:02<00:04, 3.86it/s] 39%|███▉ | 11/28 [00:02<00:04, 3.85it/s] 43%|████▎ | 12/28 [00:03<00:04, 3.85it/s] 46%|████▋ | 13/28 [00:03<00:03, 3.85it/s] 50%|█████ | 14/28 [00:03<00:03, 3.85it/s] 54%|█████▎ | 15/28 [00:03<00:03, 3.85it/s] 57%|█████▋ | 16/28 [00:04<00:03, 3.85it/s] 61%|██████ | 17/28 [00:04<00:02, 3.85it/s] 64%|██████▍ | 18/28 [00:04<00:02, 3.85it/s] 68%|██████▊ | 19/28 [00:04<00:02, 3.85it/s] 71%|███████▏ | 20/28 [00:05<00:02, 3.85it/s] 75%|███████▌ | 21/28 [00:05<00:01, 3.85it/s] 79%|███████▊ | 22/28 [00:05<00:01, 3.85it/s] 82%|████████▏ | 23/28 [00:05<00:01, 3.85it/s] 86%|████████▌ | 24/28 [00:06<00:01, 3.85it/s] 89%|████████▉ | 25/28 [00:06<00:00, 3.84it/s] 93%|█████████▎| 26/28 [00:06<00:00, 3.84it/s] 96%|█████████▋| 27/28 [00:06<00:00, 3.84it/s] 100%|██████████| 28/28 [00:07<00:00, 3.84it/s] 100%|██████████| 28/28 [00:07<00:00, 3.87it/s] Total safe images: 1 out of 1
Version Details
- Version ID
412b263ea93ebeb457fb872cf81b34388241bcdf379b93e37fb6d6c459298d0b
- Version Created
- May 1, 2025