LLM¶

Large-language model preview components.

Pipelines:

preview.llm.infer_pipeline(large_model_reference: str, prompt_dataset: str, model_checkpoint: str | None = None, prompt_sequence_length: int = 512, target_sequence_length: int = 64, sampling_strategy: str = 'greedy', instruction: str | None = None, project: str = '{{$.pipeline_google_cloud_project_id}}', location: str = '{{$.pipeline_google_cloud_location}}') → Outputs[source]¶

Uses a large-language model to perform bulk inference on a prompt dataset.

Parameters¶

large_model_reference: str¶

Name of the base model. Supported values are text-bison@001, t5-small, t5-large, t5-xl and t5-xxl. text-bison@001 and t5-small are supported in us-central1 and europe-west4. t5-large, t5-xl and t5-xxl are only supported in europe-west4.

model_checkpoint: str | None = None¶

Optional Cloud storage path to the model checkpoint. If not provided, the default checkpoint for the large_model_reference will be used.

prompt_dataset: str¶

Cloud storage path to an unlabled JSONL dataset that contains prompts. Text datasets must contain an input_text field that contains the prompt. Chat datasets must contain at least 1 message in a messages field. Each message must be valid JSON that contains author and content fields, where valid author values are user and assistant and content must be non-empty. Each row may contain multiple messages, but the first and last author must be the user. An optional context field may be provided for each example in a chat dataset. If provided, the context will preprended to the message content. The instruction serves as the default context. (Useful if most messages use the same system-level context.) Any context provided in the example will override the default value.

prompt_sequence_length: int = 512¶

Maximum tokenized sequence length for input text. Higher values increase memory overhead. This value should be at most 8192. Default value is 512.

target_sequence_length: int = 64¶

Maximum tokenized sequence length for target text. Higher values increase memory overhead. This value should be at most 1024. Default value is 64.

sampling_strategy: str = 'greedy'¶

This field specifies the sampling strategy. The valid options are ‘greedy’ and ‘temperature_sampling’.

instruction: str | None = None¶

This field lets the model know what task it needs to perform. Base models have been trained over a large set of varied instructions. You can give a simple and intuitive description of the task and the model will follow it, e.g. “Classify this movie review as positive or negative” or “Translate this sentence to Danish”. Do not specify this if your dataset already prepends the instruction to the inputs field.

project: str = '{{$.pipeline_google_cloud_project_id}}'¶

Project used to run custom jobs. If not specified the project used to run the pipeline will be used.

location: str = '{{$.pipeline_google_cloud_location}}'¶

Location used to run custom jobs. If not specified the location used to run the pipeline will be used.

Returns¶

Cloud storage path to output predictions.

preview.llm.rlhf_pipeline(prompt_dataset: str, preference_dataset: str, large_model_reference: str, model_display_name: str | None = None, prompt_sequence_length: int = 512, target_sequence_length: int = 64, reward_model_learning_rate_multiplier: float = 1.0, reinforcement_learning_rate_multiplier: float = 1.0, reward_model_train_steps: int = 1000, reinforcement_learning_train_steps: int = 1000, kl_coeff: float = 0.1, instruction: str | None = None, deploy_model: bool = True, eval_dataset: str | None = None, project: str = '{{$.pipeline_google_cloud_project_id}}', location: str = '{{$.pipeline_google_cloud_location}}', tensorboard_resource_id: str | None = None) → Outputs[source]¶

Performs reinforcement learning from human feedback.

Parameters¶

prompt_dataset: str¶

Cloud storage path to an unlabled JSONL dataset that contains prompts. Text datasets must contain an input_text field that contains the prompt. Chat datasets must contain at least 1 message in a messages field. Each message must be valid JSON that contains author and content fields, where valid author values are user and assistant and content must be non-empty. Each row may contain multiple messages, but the first and last author must be the user. An optional context field may be provided for each example in a chat dataset. If provided, the context will preprended to the message content. The instruction serves as the default context. (Useful if most messages use the same system-level context.) Any context provided in the example will override the default value.

preference_dataset: str¶

Cloud storage path to a human preference JSONL dataset used to train a reward model. Each example in a preference dataset must contain candidate_0 and candidate_1 fields that contain candidate responses, choice that specifies the preferred candidate and either input_text (if tuning a text model) or messages (if tuning a chat model). Chat datasets must contain at least 1 message in a messages field. Each message must be valid JSON that contains author and content fields, where valid author values are user and assistant and content must be non-empty. Each row may contain multiple messages, but the first and last author must be the user. An optional context field may be provided for each example in a chat dataset. If provided, the context will preprended to the message content. The instruction serves as the default context. (Useful if most messages use the same system-level context.) Any context provided in the example will override the default value.

large_model_reference: str¶

Name of the base model. Supported values are text-bison@001, t5-small, t5-large, t5-xl and t5-xxl. text-bison@001 and t5-small are supported in us-central1 and europe-west4. t5-large, t5-xl and t5-xxl are only supported in europe-west4.

model_display_name: str | None = None¶

Name of the fine-tuned model shown in the Model Registry. If not provided, a default name will be created.

prompt_sequence_length: int = 512¶

Maximum tokenized sequence length for input text. Higher values increase memory overhead. This value should be at most 8192. Default value is 512.

target_sequence_length: int = 64¶

Maximum tokenized sequence length for target text. Higher values increase memory overhead. This value should be at most 1024. Default value is 64.

reward_model_learning_rate_multiplier: float = 1.0¶

Constant used to adjust the base learning rate used when training a reward model. Multiply by a number > 1 to increase the magnitude of updates applied at each training step or multiply by a number < 1 to decrease the magnitude of updates. Default value is 1.0.

reinforcement_learning_rate_multiplier: float = 1.0¶

Constant used to adjust the base learning rate used during reinforcement learning. Multiply by a number > 1 to increase the magnitude of updates applied at each training step or multiply by a number < 1 to decrease the magnitude of updates. Default value is 1.0.

reward_model_train_steps: int = 1000¶

Number of steps to use when training a reward model. Default value is 1000.

reinforcement_learning_train_steps: int = 1000¶

Number of reinforcement learning steps to perform when tuning a base model. Default value is 1000.

kl_coeff: float = 0.1¶

Coefficient for KL penalty. This regularizes the policy model and penalizes if it diverges from its initial distribution. If set to 0, the reference language model is not loaded into memory. Default value is 0.1.

instruction: str | None = None¶

This field lets the model know what task it needs to perform. Base models have been trained over a large set of varied instructions. You can give a simple and intuitive description of the task and the model will follow it, e.g. “Classify this movie review as positive or negative” or “Translate this sentence to Danish”. Do not specify this if your dataset already prepends the instruction to the inputs field.

deploy_model: bool = True¶

Whether to deploy the model to an endpoint in us-central1. Default is True.

eval_dataset: str | None = None¶

Optional Cloud storage path to an evaluation dataset. If provided, inference will be performed on this dataset after training. The dataset format is jsonl. Each example in the dataset must contain a field input_text that contains the prompt.

project: str = '{{$.pipeline_google_cloud_project_id}}'¶

Project used to run custom jobs. If not specified the project used to run the pipeline will be used.

location: str = '{{$.pipeline_google_cloud_location}}'¶

Location used to run custom jobs. If not specified the location used to run the pipeline will be used.

tensorboard_resource_id: str | None = None¶

Optional tensorboard resource id in format projects/{project_number}/locations/{location}/tensorboards/{tensorboard_id}. If provided, tensorboard metrics will be uploaded to this location.

Returns¶

Path to the model uploaded to the Model Registry. This will be an empty string if the model was not deployed.

endpoint_resource_name: Path the Online Prediction Endpoint. This will be an empty string if the model was not deployed. :rtype: model_resource_name