components openai_completions_finetune - Azure/azureml-assets GitHub Wiki
Finetune your own OAI model. Visit https://learn.microsoft.com/en-us/azure/cognitive-services/openai/ for more info.
Version: 0.2.2
View in Studio: https://ml.azure.com/registries/azureml/components/openai_completions_finetune/version/0.2.2
| Name | Description | Type | Default | Optional | Enum |
|---|---|---|---|---|---|
| model | OAI model engine | string | davinci | False | ['ada', 'babbage', 'curie', 'davinci', 'text-davinci-fine-tune-002'] |
| registered_model_name | User-defined registered model name | string | False | ||
| train_dataset | Input dataset (file or folder). If a folder dataset is passed, includes all nested files. | uri_folder | False | ||
| validation_dataset | Input dataset (file or folder). If a folder dataset is passed, includes all nested files. | uri_folder | True | ||
| lora_weights | LoRA weights for continual finetuning. This is optional. | uri_folder | True | ||
| n_epochs | Number of epochs for the training | integer | 4 | True | |
| batch_size | The batch size to use for training. When set to -1, batch_size is calculated as 0.2% of examples in training set and the max is 256. | integer | -1 | True | |
| learning_rate_multiplier | The learning rate multiplier to use for training. Must be between 0.0 and 5.0. | number | 0.1 | True | |
| prompt_loss_weight | The prompt loss weight to use for training | number | 0.1 | True | |
| compute_classification_metrics | If set, we calculate classification-specific metrics such as accuracy and F-1 score using the validation set at the end of every epoch. In order to compute classification metrics, you must provide a validation_file. Additionally, you must specify classification_n_classes for multiclass classification or classification_positive_class for binary classification. | boolean | True | ||
| classification_n_classes | The number of classes in a classification task. This parameter is required for multiclass classification. | integer | True | ||
| classification_positive_class | The positive class in binary classification. This parameter is needed to generate precision, recall, and F1 metrics when doing binary classification. | string | True | ||
| classification_betas | If this is provided, we calculate F-beta scores at the specified beta values. The F-beta score is a generalization of F-1 score. This is only used for binary classification. With a beta of 1 (i.e. the F-1 score), precision and recall are given the same weight. A larger beta score puts more weight on recall and less on precision. A smaller beta score puts more weight on precision and less on recall. The value specified should be a comma separated list of doubles. | string | True | ||
| quota_enforcement_resource_id | Owner subscription id. | string | True |
| Name | Description | Type |
|---|---|---|
| output_model | Dataset with the output model weights (LoRA weights) | uri_folder |