Important

You are viewing the NeMo 2.0 documentation. This release introduces significant changes to the API and a new library, NeMo Run. We are currently porting all features from NeMo 1.0 to 2.0. For documentation on previous versions or features not yet available in 2.0, please refer to the NeMo 24.07 documentation.

Run NeMo Framework Inference#

For text-to-image models, the inference script generates images from text prompts defined in the config file.

To enable the inference stage with Stable Diffusion, change the settings in the configuration files:

  1. In the defaults section of conf/config.yaml, update the fw_inference field to point to the desired Stable Diffusion inference configuration file. For example, if you want to use the stable_diffusion/text2img.yaml configuration, change the fw_inference field to stable_diffusion/text2img. For sdxl inference, change the field to sdxl/sdxl_infer.

    defaults:
      - fw_inference: stable_diffusion/text2img
      ...
    
  2. In the stages field of conf/config.yaml, make sure the fw_inference stage is included. For example,

    stages:
      - fw_inference
      ...
    
  3. Configure prompts and num_images_per_prompt fields of conf/fw_inference/stable_diffusion/text2img.yaml. Set model.restore_from_path to the .nemo ckpt you want generate images with.

Remarks:

We support three types of inference samplers, ‘DDIM’, ‘PLMS’ and ‘DPM’, which can be changed in the config files. The ‘DPM’ sampler is added in recent updates. It is able to achieve similar image quality with half the number of steps needed for inference.