/*
 * Code generated by Speakeasy (https://speakeasy.com). DO NOT EDIT.
 */

import * as z from "zod/v3";
import { remap as remap$ } from "../../lib/primitives.js";

/**
 * The fine-tuning hyperparameter settings used in a classifier fine-tune job.
 */
export type ClassifierTrainingParametersIn = {
  /**
   * The number of training steps to perform. A training step refers to a single update of the model weights during the fine-tuning process. This update is typically calculated using a batch of samples from the training dataset.
   */
  trainingSteps?: number | null | undefined;
  /**
   * A parameter describing how much to adjust the pre-trained model's weights in response to the estimated error each time the weights are updated during the fine-tuning process.
   */
  learningRate?: number | undefined;
  /**
   * (Advanced Usage) Weight decay adds a term to the loss function that is proportional to the sum of the squared weights. This term reduces the magnitude of the weights and prevents them from growing too large.
   */
  weightDecay?: number | null | undefined;
  /**
   * (Advanced Usage) A parameter that specifies the percentage of the total training steps at which the learning rate warm-up phase ends. During this phase, the learning rate gradually increases from a small value to the initial learning rate, helping to stabilize the training process and improve convergence. Similar to `pct_start` in [mistral-finetune](https://github.com/mistralai/mistral-finetune)
   */
  warmupFraction?: number | null | undefined;
  epochs?: number | null | undefined;
  seqLen?: number | null | undefined;
};

/** @internal */
export type ClassifierTrainingParametersIn$Outbound = {
  training_steps?: number | null | undefined;
  learning_rate: number;
  weight_decay?: number | null | undefined;
  warmup_fraction?: number | null | undefined;
  epochs?: number | null | undefined;
  seq_len?: number | null | undefined;
};

/** @internal */
export const ClassifierTrainingParametersIn$outboundSchema: z.ZodType<
  ClassifierTrainingParametersIn$Outbound,
  z.ZodTypeDef,
  ClassifierTrainingParametersIn
> = z.object({
  trainingSteps: z.nullable(z.number().int()).optional(),
  learningRate: z.number().default(0.0001),
  weightDecay: z.nullable(z.number()).optional(),
  warmupFraction: z.nullable(z.number()).optional(),
  epochs: z.nullable(z.number()).optional(),
  seqLen: z.nullable(z.number().int()).optional(),
}).transform((v) => {
  return remap$(v, {
    trainingSteps: "training_steps",
    learningRate: "learning_rate",
    weightDecay: "weight_decay",
    warmupFraction: "warmup_fraction",
    seqLen: "seq_len",
  });
});

export function classifierTrainingParametersInToJSON(
  classifierTrainingParametersIn: ClassifierTrainingParametersIn,
): string {
  return JSON.stringify(
    ClassifierTrainingParametersIn$outboundSchema.parse(
      classifierTrainingParametersIn,
    ),
  );
}