biggpt1/vladikavkaz-train 🖼️🔢❓📝✓ → 🖼️

▶️ 165 runs 📅 Feb 2025 ⚙️ Cog 0.13.7

About

Example Output

Prompt:

TRN glides effortlessly along the weathered cobblestone streets of Vladikavkaz, its gleaming metallic blue and matte black body standing in striking contrast to the cascading magenta and fuchsia bougainvillea flowers that envelop its lower half in a surreal floral embrace. The floral patterns, intricately woven from real flowers, appear almost otherworldly, their delicate petals shimmering under the ethereal glow of the dramatic golden sunset rays piercing through them.

Above the windshield, the LED route display glows warmly, displaying "Abon × Lavar" in rich amber tones, adding a futuristic element to the fantasy-like tram. The tram windows remain clear, revealing an enchanting interior filled with lush flowers, draping over seats and handrails, creating a mesmerizing dreamlike atmosphere. As the tram moves forward, a soft breeze sends slow-motion flower petals dancing across the ancient cobblestones, their gentle motion captured in breathtaking cinematic beauty.

The atmosphere is cinematic and moody, enhanced by an atmospheric haze that softens the edges of the scene. Overhead, intricate electric lines create mesmerizing geometric patterns, stretching across the sky as silhouettes of bare winter trees frame the background. The peach-orange hues of the twilight sky blend with subtle gradients of violet and pink, casting a spellbinding light over the landscape.

The tram windows reflect the mystical glow of the sunset, producing dreamy lens flares that add to the epic cinematic composition. Delicate morning dew clings to the flower petals, catching the last light of day, while the subtle presence of artistic chromatic aberration and film grain adds a tactile realism to the scene.

Shot with ARRI Alexa 65, anamorphic 40mm T/1.4, the image captures hyperdetailed textures, from the weathered patina of the cobblestones to the intricate flower arrangements that seem almost enchanted. The scene is filled with urban poetry, a seamless fusion of modern technology and the delicate, ephemeral beauty of nature.

Every detail, from the floating flower petals to the high-detail wire cables, contributes to a masterpiece of professional cinematography, cinematic lighting, and reflective surfaces, evoking a sense of wonder and magic—an enchanted floral dream unfolding in the heart of the city.

Output

Performance Metrics

13.47s Prediction Time

13.81s Total Time

All Input Parameters

{
  "model": "dev",
  "prompt": "TRN glides effortlessly along the weathered cobblestone streets of Vladikavkaz, its gleaming metallic blue and matte black body standing in striking contrast to the cascading magenta and fuchsia bougainvillea flowers that envelop its lower half in a surreal floral embrace. The floral patterns, intricately woven from real flowers, appear almost otherworldly, their delicate petals shimmering under the ethereal glow of the dramatic golden sunset rays piercing through them.\n\nAbove the windshield, the LED route display glows warmly, displaying \"Abon × Lavar\" in rich amber tones, adding a futuristic element to the fantasy-like tram. The tram windows remain clear, revealing an enchanting interior filled with lush flowers, draping over seats and handrails, creating a mesmerizing dreamlike atmosphere. As the tram moves forward, a soft breeze sends slow-motion flower petals dancing across the ancient cobblestones, their gentle motion captured in breathtaking cinematic beauty.\n\nThe atmosphere is cinematic and moody, enhanced by an atmospheric haze that softens the edges of the scene. Overhead, intricate electric lines create mesmerizing geometric patterns, stretching across the sky as silhouettes of bare winter trees frame the background. The peach-orange hues of the twilight sky blend with subtle gradients of violet and pink, casting a spellbinding light over the landscape.\n\nThe tram windows reflect the mystical glow of the sunset, producing dreamy lens flares that add to the epic cinematic composition. Delicate morning dew clings to the flower petals, catching the last light of day, while the subtle presence of artistic chromatic aberration and film grain adds a tactile realism to the scene.\n\nShot with ARRI Alexa 65, anamorphic 40mm T/1.4, the image captures hyperdetailed textures, from the weathered patina of the cobblestones to the intricate flower arrangements that seem almost enchanted. The scene is filled with urban poetry, a seamless fusion of modern technology and the delicate, ephemeral beauty of nature.\n\nEvery detail, from the floating flower petals to the high-detail wire cables, contributes to a masterpiece of professional cinematography, cinematic lighting, and reflective surfaces, evoking a sense of wonder and magic—an enchanted floral dream unfolding in the heart of the city.\n",
  "go_fast": true,
  "extra_lora": "https://civitai.com/api/download/models/1388131?type=Model&format=SafeTensor&token=ae82acec6ad7050b7f1e3654da2bb265",
  "lora_scale": 0.59,
  "megapixels": "1",
  "num_outputs": 4,
  "aspect_ratio": "9:16",
  "output_format": "jpg",
  "guidance_scale": 2.72,
  "output_quality": 80,
  "prompt_strength": 0.8,
  "extra_lora_scale": 0.57,
  "num_inference_steps": 28
}

Input Parameters

mask Type: string: Image mask for image inpainting mode. If provided, aspect_ratio, width, and height inputs are ignored.
seed Type: integer: Random seed. Set for reproducible generation
image Type: string: Input image for image to image or inpainting mode. If provided, aspect_ratio, width, and height inputs are ignored.
model Default: dev: 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 Type: integerRange: 256 - 1440: 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 Type: integerRange: 256 - 1440: 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) Type: string: 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 Type: booleanDefault: false: Run faster predictions with model optimized for speed (currently fp8 quantized); disable to run in original bf16
extra_lora Type: string: 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 Type: numberDefault: 1Range: -1 - 3: 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 Default: 1: Approximate number of megapixels for generated image
num_outputs Type: integerDefault: 1Range: 1 - 4: Number of outputs to generate
aspect_ratio Default: 1:1: Aspect ratio for the generated image. If custom is selected, uses height and width below & will run in bf16 mode
output_format Default: webp: Format of the output images
guidance_scale Type: numberDefault: 3Range: 0 - 10: 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 Type: integerDefault: 80Range: 0 - 100: 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 Type: numberDefault: 0.8Range: 0 - 1: Prompt strength when using img2img. 1.0 corresponds to full destruction of information in image
extra_lora_scale Type: numberDefault: 1Range: -1 - 3: 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 Type: string: 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 Type: integerDefault: 28Range: 1 - 50: Number of denoising steps. More steps can give more detailed images, but take longer.
disable_safety_checker Type: booleanDefault: false: Disable safety checker for generated images.

Output Schema

Output

Type: array • Items Type: string • Items Format: uri

Example Execution Logs

2025-02-20 22:17:30.310 | INFO     | fp8.lora_loading:convert_lora_weights:502 - Loading LoRA weights for /src/weights-cache/75d9ad6083eb1ffd
2025-02-20 22:17:30.416 | INFO     | fp8.lora_loading:convert_lora_weights:523 - LoRA weights loaded
2025-02-20 22:17:30.416 | DEBUG    | fp8.lora_loading:apply_lora_to_model_and_optionally_store_clones:610 - Extracting keys
2025-02-20 22:17:30.417 | DEBUG    | fp8.lora_loading:apply_lora_to_model_and_optionally_store_clones:617 - Keys extracted
Applying LoRA:   0%|          | 0/304 [00:00<?, ?it/s]
Applying LoRA:  40%|███▉      | 121/304 [00:00<00:00, 1200.91it/s]
Applying LoRA:  80%|███████▉  | 242/304 [00:00<00:00, 948.22it/s] 
Applying LoRA: 100%|██████████| 304/304 [00:00<00:00, 956.49it/s]
2025-02-20 22:17:30.735 | INFO     | fp8.lora_loading:apply_lora_to_model_and_optionally_store_clones:669 - Loading LoRA in fp8
2025-02-20 22:17:30.735 | SUCCESS  | fp8.lora_loading:load_lora:547 - LoRA applied in 0.43s
2025-02-20 22:17:30.735 | INFO     | fp8.lora_loading:convert_lora_weights:502 - Loading LoRA weights for /src/weights-cache/c4ed393014c3c290
2025-02-20 22:17:30.822 | INFO     | fp8.lora_loading:convert_lora_weights:523 - LoRA weights loaded
2025-02-20 22:17:30.822 | DEBUG    | fp8.lora_loading:apply_lora_to_model_and_optionally_store_clones:610 - Extracting keys
2025-02-20 22:17:30.823 | DEBUG    | fp8.lora_loading:apply_lora_to_model_and_optionally_store_clones:617 - Keys extracted
Applying LoRA:   0%|          | 0/304 [00:00<?, ?it/s]
Applying LoRA:  40%|███▉      | 121/304 [00:00<00:00, 1199.91it/s]
Applying LoRA:  79%|███████▉  | 241/304 [00:00<00:00, 948.23it/s] 
Applying LoRA: 100%|██████████| 304/304 [00:00<00:00, 952.11it/s]
2025-02-20 22:17:31.142 | INFO     | fp8.lora_loading:apply_lora_to_model_and_optionally_store_clones:669 - Loading LoRA in fp8
2025-02-20 22:17:31.142 | SUCCESS  | fp8.lora_loading:load_lora:547 - LoRA applied in 0.41s
running quantized prediction
Using seed: 3089291403
  0%|          | 0/28 [00:00<?, ?it/s]
  7%|▋         | 2/28 [00:00<00:01, 16.86it/s]
 14%|█▍        | 4/28 [00:00<00:01, 12.43it/s]
 21%|██▏       | 6/28 [00:00<00:01, 11.47it/s]
 29%|██▊       | 8/28 [00:00<00:01, 11.06it/s]
 36%|███▌      | 10/28 [00:00<00:01, 10.75it/s]
 43%|████▎     | 12/28 [00:01<00:01, 10.52it/s]
 50%|█████     | 14/28 [00:01<00:01, 10.50it/s]
 57%|█████▋    | 16/28 [00:01<00:01, 10.50it/s]
 64%|██████▍   | 18/28 [00:01<00:00, 10.48it/s]
 71%|███████▏  | 20/28 [00:01<00:00, 10.42it/s]
 79%|███████▊  | 22/28 [00:02<00:00, 10.33it/s]
 86%|████████▌ | 24/28 [00:02<00:00, 10.31it/s]
 93%|█████████▎| 26/28 [00:02<00:00, 10.36it/s]
100%|██████████| 28/28 [00:02<00:00, 10.37it/s]
100%|██████████| 28/28 [00:02<00:00, 10.66it/s]
  0%|          | 0/28 [00:00<?, ?it/s]
  7%|▋         | 2/28 [00:00<00:02, 10.37it/s]
 14%|█▍        | 4/28 [00:00<00:02, 10.25it/s]
 21%|██▏       | 6/28 [00:00<00:02, 10.30it/s]
 29%|██▊       | 8/28 [00:00<00:01, 10.37it/s]
 36%|███▌      | 10/28 [00:00<00:01, 10.39it/s]
 43%|████▎     | 12/28 [00:01<00:01, 10.35it/s]
 50%|█████     | 14/28 [00:01<00:01, 10.29it/s]
 57%|█████▋    | 16/28 [00:01<00:01, 10.32it/s]
 64%|██████▍   | 18/28 [00:01<00:00, 10.37it/s]
 71%|███████▏  | 20/28 [00:01<00:00, 10.40it/s]
 79%|███████▊  | 22/28 [00:02<00:00, 10.41it/s]
 86%|████████▌ | 24/28 [00:02<00:00, 10.34it/s]
 93%|█████████▎| 26/28 [00:02<00:00, 10.32it/s]
100%|██████████| 28/28 [00:02<00:00, 10.35it/s]
100%|██████████| 28/28 [00:02<00:00, 10.35it/s]
  0%|          | 0/28 [00:00<?, ?it/s]
  7%|▋         | 2/28 [00:00<00:02, 10.38it/s]
 14%|█▍        | 4/28 [00:00<00:02, 10.39it/s]
 21%|██▏       | 6/28 [00:00<00:02, 10.34it/s]
 29%|██▊       | 8/28 [00:00<00:01, 10.29it/s]
 36%|███▌      | 10/28 [00:00<00:01, 10.33it/s]
 43%|████▎     | 12/28 [00:01<00:01, 10.33it/s]
 50%|█████     | 14/28 [00:01<00:01, 10.34it/s]
 57%|█████▋    | 16/28 [00:01<00:01, 10.32it/s]
 64%|██████▍   | 18/28 [00:01<00:00, 10.32it/s]
 71%|███████▏  | 20/28 [00:01<00:00, 10.33it/s]
 79%|███████▊  | 22/28 [00:02<00:00, 10.35it/s]
 86%|████████▌ | 24/28 [00:02<00:00, 10.32it/s]
 93%|█████████▎| 26/28 [00:02<00:00, 10.35it/s]
100%|██████████| 28/28 [00:02<00:00, 10.35it/s]
100%|██████████| 28/28 [00:02<00:00, 10.34it/s]
  0%|          | 0/28 [00:00<?, ?it/s]
  7%|▋         | 2/28 [00:00<00:02, 10.31it/s]
 14%|█▍        | 4/28 [00:00<00:02, 10.29it/s]
 21%|██▏       | 6/28 [00:00<00:02, 10.33it/s]
 29%|██▊       | 8/28 [00:00<00:01, 10.33it/s]
 36%|███▌      | 10/28 [00:00<00:01, 10.37it/s]
 43%|████▎     | 12/28 [00:01<00:01, 10.35it/s]
 50%|█████     | 14/28 [00:01<00:01, 10.32it/s]
 57%|█████▋    | 16/28 [00:01<00:01, 10.32it/s]
 64%|██████▍   | 18/28 [00:01<00:00, 10.30it/s]
 71%|███████▏  | 20/28 [00:01<00:00, 10.26it/s]
 79%|███████▊  | 22/28 [00:02<00:00, 10.31it/s]
 86%|████████▌ | 24/28 [00:02<00:00, 10.34it/s]
 93%|█████████▎| 26/28 [00:02<00:00, 10.35it/s]
100%|██████████| 28/28 [00:02<00:00, 10.33it/s]
100%|██████████| 28/28 [00:02<00:00, 10.32it/s]
Total safe images: 4 out of 4

Version Details

Version ID: eb8a05238a7228257baaec309346827c7e09cf6f47c884c094b30d049590e0c9
Version Created: February 12, 2025

Run on Replicate →