Model reference · open weights
HiDream-I1-Fast is an open-weight image model from HiDream-ai. HiDream-I1-Fast (BF16) weighs 47.2 GB; the smallest configuration that runs it is L40S 48 GB.
Summary of the HiDream-ai/HiDream-I1-Fast model card, 2026-10-01
What it is
| Released by | HiDream-ai |
|---|---|
| Released | 2025-04-06 |
| Parameters | 17.1B |
| VRAM | 47.2 GB for the weights |
What it runs on
| Card | One image | Memory |
|---|---|---|
| RTX 3060 12 GB … RTX 5090 32 GB | does not fit | |
| L40S 48 GB | tight (encoders offloaded) | 44.0 GB |
| A100 80 GB | fits | 78.2 GB |
| H100 80 GB | fits | 78.1 GB |
| RTX PRO 6000 Blackwell 96 GB | fits | 93.8 GB |
| DGX Spark (GB10) 128 GB unified | fits | 107 GB |
| H200 141 GB | fits | 138 GB |
| B200 180 GB | fits | 176 GB |
From the model card
HiDream-I1 is a new open-source image generative foundation model with 17B parameters that achieves state-of-the-art image generation quality within seconds.
Please make sure you have installed Flash Attention. We recommend CUDA version 12.4 for the manual installation.
pip install -r requirements.txt
Clone the GitHub repo:
git clone https://github.com/HiDream-ai/HiDream-I1
Then you can run the inference scripts to generate images:
# For full model inference
python ./inference.py --model_type full
# For distilled dev model inference
python ./inference.py --model_type dev
# For distilled fast model inference
python ./inference.py --model_type fast
Note: The inference script will automatically download
meta-llama/Meta-Llama-3.1-8B-Instructmodel files. If you encounter network issues, you can download these files ahead of time and place them in the appropriate cache directory to avoid download failures during inference.
We also provide a Gradio demo for interactive image generation. You can run the demo with:
python gradio_demo.py
| Model | Overall | Global | Entity | Attribute | Relation | Other |
|---|---|---|---|---|---|---|
| PixArt-alpha | 71.11 | 74.97 | 79.32 | 78.60 | 82.57 | 76.96 |
| SDXL | 74.65 | 83.27 | 82.43 | 80.91 | 86.76 | 80.41 |
| DALL-E 3 | 83.50 | 90.97 | 89.61 | 88.39 | 90.58 | 89.83 |
| Flux.1-dev | 83.79 | 85.80 | 86.79 | 89.98 | 90.04 | 89.90 |
| SD3-Medium | 84.08 | 87.90 | 91.01 | 88.83 | 80.70 | 88.68 |
| Janus-Pro-7B | 84.19 | 86.90 | 88.90 | 89.40 | 89.32 | 89.48 |
| CogView4-6B | 85.13 | 83.85 | 90.35 | 91.17 | 91.14 | 87.29 |
| HiDream-I1 | 85.89 | 76.44 | 90.22 | 89.48 | 93.74 | 91.83 |
| Model | Overall | Single Obj. | Two Obj. | Counting | Colors | Position | Color attribution |
|---|---|---|---|---|---|---|---|
| SDXL | 0.55 | 0.98 | 0.74 | 0.39 | 0.85 | 0.15 | 0.23 |
| PixArt-alpha | 0.48 | 0.98 | 0.50 | 0.44 | 0.80 | 0.08 | 0.07 |
| Flux.1-dev | 0.66 | 0.98 | 0.79 | 0.73 | 0.77 | 0.22 | 0.45 |
| DALL-E 3 | 0.67 | 0.96 | 0.87 | 0.47 | 0.83 | 0.43 | 0.45 |
| CogView4-6B | 0.73 | 0.99 | 0.86 | 0.66 | 0.79 | 0.48 | 0.58 |
| SD3-Medium | 0.74 | 0.99 | 0.94 | 0.72 | 0.89 | 0.33 | 0.60 |
| Janus-Pro-7B | 0.80 | 0.99 | 0.89 | 0.59 | 0.90 | 0.79 | 0.66 |
| HiDream-I1 | 0.83 | 1.00 | 0.98 | 0.79 | 0.91 | 0.60 | 0.72 |
| Model | Averaged | Animation | Concept-art | Painting | Photo |
|---|---|---|---|---|---|
| Stable Diffusion v2.0 | 26.38 | 27.09 | 26.02 | 25.68 | 26.73 |
| Midjourney V6 | 30.29 | 32.02 | 30.29 | 29.74 | 29.10 |
| SDXL | 30.64 | 32.84 | 31.36 | 30.86 | 27.48 |
| Dall-E3 | 31.44 | 32.39 | 31.09 | 31.18 | 31.09 |
| SD3 | 31.53 | 32.60 | 31.82 | 32.06 | 29.62 |
| Midjourney V5 | 32.33 | 34.05 | 32.47 | 32.24 | 30.56 |
| CogView4-6B | 32.31 | 33.23 | 32.60 | 32.89 | 30.52 |
| Flux.1-dev | 32.47 | 33.87 | 32.27 | 32.62 | 31.11 |
| stable cascade | 32.95 | 34.58 | 33.13 | 33.29 | 30.78 |
| HiDream-I1 | 33.82 | 35.05 | 33.74 | 33.88 | 32.61 |
The Transformer models in this repository are licensed under the MIT License. The VAE is from FLUX.1 [schnell], and the text encoders from google/t5-v1_1-xxl and meta-llama/Meta-Llama-3.1-8B-Instruct. Please follow the license terms specified for these components. You own all content you create with this model. You can use your generated content fre
Quoted from the model card on Hugging Face. The full card is behind the Hugging Face link above.
Running it yourself
Rent a machine by the hour. ComfyUI is installed on it. Open ComfyUI through the tunnel, then Templates → HiDream I1 Fast: it loads ComfyUI's own build of this model, the files above.
# on your rented machine: pip install diffusers transformers accelerate
import torch
from diffusers import DiffusionPipeline
pipe = DiffusionPipeline.from_pretrained("HiDream-ai/HiDream-I1-Fast", torch_dtype=torch.bfloat16).to("cuda")
image = pipe(prompt="a red bicycle on a cobbled street").images[0]
image.save("/workspace/out.png")# on your rented machine (the ssh line is on its page in the console)
# get REPO FILE FOLDER: one file into /workspace/models/FOLDER, where ComfyUI loads it from
get() { hf download "$1" "$2" --local-dir /workspace/hf-files && mkdir -p "/workspace/models/$3" && mv "/workspace/hf-files/$2" "/workspace/models/$3/$4"; }
# the files of the template “HiDream I1 Fast” (31.0 GB)
get Comfy-Org/HiDream-I1_ComfyUI split_files/vae/ae.safetensors vae
get Comfy-Org/HiDream-I1_ComfyUI split_files/text_encoders/clip_l_hidream.safetensors text_encoders
get Comfy-Org/HiDream-I1_ComfyUI split_files/text_encoders/clip_g_hidream.safetensors text_encoders
get Comfy-Org/HiDream-I1_ComfyUI split_files/text_encoders/t5xxl_fp8_e4m3fn_scaled.safetensors text_encoders
get Comfy-Org/HiDream-I1_ComfyUI split_files/text_encoders/llama_3.1_8b_instruct_fp8_scaled.safetensors text_encoders
get Comfy-Org/HiDream-I1_ComfyUI split_files/diffusion_models/hidream_i1_fast_fp8.safetensors diffusion_models
start-comfyui# on your computer, in a second terminal: ComfyUI in your browser at http://localhost:8188
# HOST and PORT are your machine's, from its page in the console
ssh -L 8188:localhost:8188 dev@HOST -p PORT