Instructions to use Disty0/Ideogram-4-SDNQ-FP8 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Diffusers
How to use Disty0/Ideogram-4-SDNQ-FP8 with Diffusers:
pip install -U diffusers transformers accelerate
import torch from diffusers import DiffusionPipeline # switch to "mps" for apple devices pipe = DiffusionPipeline.from_pretrained("Disty0/Ideogram-4-SDNQ-FP8", 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
- Local Apps Settings
- Draw Things
- DiffusionBee
This model is a direct conversion of Ideogram-4 FP8 to SDNQ Diffusers format with identical weights from the original FP8 model.
pip install sdnq
import os
import json
import requests
import torch
import diffusers
from sdnq import SDNQConfig # import sdnq to register it into diffusers and transformers
from sdnq.common import use_torch_compile as triton_is_available
from sdnq.loader import apply_sdnq_options_to_model
pipe = diffusers.Ideogram4Pipeline.from_pretrained("Disty0/Ideogram-4-SDNQ-FP8", torch_dtype=torch.bfloat16)
# Enable FP8 MatMul for AMD, Intel ARC and Nvidia GPUs:
if triton_is_available and (torch.cuda.is_available() or torch.xpu.is_available()):
pipe.transformer = apply_sdnq_options_to_model(pipe.transformer, use_quantized_matmul=True)
pipe.unconditional_transformer = apply_sdnq_options_to_model(pipe.unconditional_transformer, use_quantized_matmul=True)
pipe.text_encoder = apply_sdnq_options_to_model(pipe.text_encoder, use_quantized_matmul=True)
# pipe.transformer = torch.compile(pipe.transformer) # optional for faster speeds
# pipe.unconditional_transformer = torch.compile(pipe.unconditional_transformer) # optional for faster speeds
pipe.enable_model_cpu_offload()
# Expand the prompt into a structured JSON caption with Ideogram's free hosted magic-prompt API.
# Get a key at https://developer.ideogram.ai/ (set IDEOGRAM_API_KEY).
resp = requests.post(
"https://api.ideogram.ai/v1/ideogram-v4/magic-prompt",
headers={"Api-Key": "your_ideogram_api_key"},
json={"text_prompt": "a ginger cat wearing a tiny wizard hat reading a spellbook", "aspect_ratio": "1x1"},
).json()
caption = json.dumps(resp["json_prompt"]) # or: token="hf_xxxxxxxxx", token is needed as the repo is gated
# Pass the caption straight to the pipeline (no prompt_upsampling โ it's already upsampled).
image = pipe(
caption,
height=1024, # model supports up to 2048
width=1024, # model supports up to 2048
generator=torch.manual_seed(0),
).images[0]
image.save("ideogram4-sdnq-fp8.png")
- Downloads last month
- -
Model tree for Disty0/Ideogram-4-SDNQ-FP8
Base model
ideogram-ai/ideogram-4-fp8