Instructions to use nvidia/Cosmos3-Super-Text2Image with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Cosmos
How to use nvidia/Cosmos3-Super-Text2Image with Cosmos:
# No code snippets available yet for this library. # To use this model, check the repository files and the library's documentation. # Want to help? PRs adding snippets are welcome at: # https://github.com/huggingface/huggingface.js
- Diffusers
How to use nvidia/Cosmos3-Super-Text2Image with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("nvidia/Cosmos3-Super-Text2Image", dtype=torch.bfloat16, device_map="cuda") prompt = "Astronaut in a jungle, cold color palette, muted colors, detailed, 8k" image = pipe(prompt).images[0] - Notebooks
- Google Colab
- Kaggle
| """Public defaults for the standalone agentic text-to-image upsampler.""" | |
| from __future__ import annotations | |
| from typing import Any | |
| DEFAULT_OPENAI_ENDPOINT_URL = "https://api.openai.com/v1" | |
| DEFAULT_UPSAMPLER_ENDPOINT_URL = DEFAULT_OPENAI_ENDPOINT_URL | |
| DEFAULT_REWRITER_ENDPOINT_URL = DEFAULT_OPENAI_ENDPOINT_URL | |
| DEFAULT_UPSAMPLER_MODEL = "gpt-5.5" | |
| DEFAULT_REWRITER_MODEL = "gpt-5.5" | |
| DEFAULT_OPENAI_API_KEY_ENV = "OPENAI_API_KEY" | |
| DEFAULT_LLM_EXTRA_BODY: dict[str, Any] = {"reasoning_effort": "low"} | |
| DEFAULT_CRITIC_ENDPOINT_URL = "https://generativelanguage.googleapis.com/v1beta/openai/" | |
| DEFAULT_CRITIC_MODEL = "gemini-3.1-pro-preview" | |
| DEFAULT_GEMINI_API_KEY_ENV = "GEMINI_API_KEY" | |
| DEFAULT_GENERATION_AUTH_KEY_ENV = "AGENTIC_UPSAMPLING_GENERATION_AUTH_KEY" | |
| DEFAULT_GENERATION_MODEL = "nvidia/Cosmos3-Super-Text2Image" | |
| DEFAULT_IMAGE_SIZE = "1024x1024" | |
| DEFAULT_GENERATION_EXTRA_ARGS: dict[str, Any] = {"guardrails": False, "use_resolution_template": False} | |
| DEFAULT_RESOLUTION = "768" | |
| DEFAULT_ASPECT_RATIO = "1,1" | |
| DEFAULT_NUM_STEPS = 50 | |
| DEFAULT_GUIDANCE = 4.0 | |
| DEFAULT_FLOW_SHIFT = 3.0 | |
| DEFAULT_MAX_ITERATIONS = 2 | |
| DEFAULT_SAMPLES_PER_ITERATION = 3 | |
| DEFAULT_JPEG_QUALITY = 99 | |
| STRICT_OVERALL_THRESHOLD = 9.0 | |
| STRICT_PROMPT_THRESHOLD = 9.0 | |