HiDream-ai/HiDream-I1

Health v2 · maintenance only

Flags: no_releases

Adoption not part of the score

2512 stars · 247 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

HiDream-I1 is an open-source 17B-parameter text-to-image generative foundation model based on a Sparse Diffusion Transformer, with full and distilled (Dev/Fast) variants. The repo provides inference scripts and model weights, and the model is officially supported in Hugging Face diffusers.

Use cases

• generate images from text prompts
• run a state-of-the-art text-to-image model locally
• generate images quickly with a distilled fast model
• integrate a text-to-image model into diffusers pipelines
• compare full vs distilled image generation quality

When to choose

• you need high-quality open-source text-to-image generation with a permissive MIT license
• you have a CUDA GPU with flash-attention and want local inference
• you want diffusers-compatible image generation models

When to avoid

• you lack a high-VRAM GPU, since the model is 17B parameters
• you need image editing rather than text-to-image (use HiDream-E1 instead)
• you need a lightweight model for CPU-only environments

Facets

library · maturity active

machine-learning deep-learning image-processing llm-inference artificial-intelligence machine-learning image-processing deep-learning python text-to-image diffusion-transformer image-generation diffusers generative-ai gpu linux

Member repositories

For agents

markdown · JSON · MCP: product_card(name="HiDream-ai/HiDream-I1")