How to generate text with KerasHub

Pretrained language-model presets in KerasHub let a Keras 3 project turn a prompt string into generated text without training a model first. A short generation run checks the package install, backend selection, preset download, tokenizer/preprocessor path, and model call before the code moves into a notebook, API, or batch job.

The GPT2CausalLM task exposes generate() for causal language modeling. Loading from gpt2_base_en attaches the matching preprocessor, and a shorter preprocessor sequence_length keeps a CPU smoke test bounded while still exercising the preset assets.

The script uses the TensorFlow backend because keras-hub installs TensorFlow for preprocessing and validates without extra backend wheels. Generated wording can vary when sampling is enabled; compiling with the greedy sampler keeps the smoke-test output tied to one completion path.

1. Install KerasHub and the TensorFlow backend in the project environment.

   $ python -m pip install --upgrade keras-hub tensorflow

   Use an isolated virtual environment, notebook kernel, or container for model dependencies. The first preset load also downloads model assets into the Keras/Kaggle cache for that environment.

2. Create generate_kerashub_text.py with a backend choice, GPT-2 preset, greedy sampler, and prompt.

   import os
    
   os.environ.setdefault("KERAS_BACKEND", "tensorflow")
    
   import keras
   import keras_hub
    
    
   PRESET = "gpt2_base_en"
   PROMPT = "KerasHub lets developers"
    
    
   preprocessor = keras_hub.models.GPT2CausalLMPreprocessor.from_preset(
       PRESET,
       sequence_length=64,
   )
   model = keras_hub.models.GPT2CausalLM.from_preset(
       PRESET,
       preprocessor=preprocessor,
   )
   model.compile(sampler="greedy")
    
   completion = model.generate(PROMPT, max_length=40, strip_prompt=True)
    
   print(f"backend: {keras.config.backend()}")
   print(f"preset: {PRESET}")
   print(f"prompt: {PROMPT}")
   print(f"completion: {completion.strip()}")

   Set KERAS_BACKEND before importing keras, keras_hub, or any project module that imports Keras. A backend change after import does not move the running process to another framework.

3. Run the script.

   $ python generate_kerashub_text.py
   backend: tensorflow
   preset: gpt2_base_en
   prompt: KerasHub lets developers
   completion: create and share their own content with the community.
   
   KerasHub is a free, open source, open source, and open source software platform for developers

   The first run downloads gpt2_base_en weights and tokenizer files. Use a larger preset or a GPU-backed environment only when the project needs higher-quality generation than a local smoke test.

4. Confirm the printed fields match the selected backend, preset, prompt, and completion.

   strip_prompt=True prints only the generated continuation. Remove it when the downstream code needs the prompt and completion in one returned string.