Tuning

This version of ibm-watsonx-ai client introduces support for Tune Experiments.

This section shows how to create a tuning experiment, deploy a model and use the ModelInference module with the created deployment.

• Working with TuneExperiment and PromptTuner
  • Configure PromptTuner
  • Get configuration parameters
  • Run prompt tuning
  • Get run status, get run details
  • Get data connections
  • Summary
  • Plot learning curves
  • Get the model identifier
• Working with TuneExperiment and FineTuner
  • Configure FineTuner
  • Get configuration parameters
  • Run fine tuning
  • Get run status, get run details
  • Get data connections
  • Summary
  • Plot learning curves
  • Get the model identifier
• Tuned Model Inference
  • Working with deployments
  • Creating ModelInference instance
  • Importing data
  • Analyzing satisfaction
  • Generate methods
• Tune Experiment
  • TuneExperiment
    • TuneExperiment
      • TuneExperiment.fine_tuner()
      • TuneExperiment.prompt_tuner()
      • TuneExperiment.runs()
  • Tune Runs
    • TuneRuns
      • TuneRuns.get_run_details()
      • TuneRuns.get_tuner()
      • TuneRuns.list()
  • Prompt Tuner
    • PromptTuner
      • PromptTuner.cancel_run()
      • PromptTuner.get_data_connections()
      • PromptTuner.get_model_id()
      • PromptTuner.get_params()
      • PromptTuner.get_run_details()
      • PromptTuner.get_run_status()
      • PromptTuner.plot_learning_curve()
      • PromptTuner.run()
      • PromptTuner.summary()
  • Enums
    • PromptTuningTypes
      • PromptTuningTypes.PT
    • PromptTuningInitMethods
      • PromptTuningInitMethods.RANDOM
      • PromptTuningInitMethods.TEXT
    • TuneExperimentTasks
      • TuneExperimentTasks.CLASSIFICATION
      • TuneExperimentTasks.CODE_GENERATION_AND_CONVERSION
      • TuneExperimentTasks.EXTRACTION
      • TuneExperimentTasks.GENERATION
      • TuneExperimentTasks.QUESTION_ANSWERING
      • TuneExperimentTasks.RETRIEVAL_AUGMENTED_GENERATION
      • TuneExperimentTasks.SUMMARIZATION
    • PromptTunableModels