phospho-app/act-policy – API reference

phospho-app / act-policy

Let's you train an ACT policy from a LeRobot dataset

Public
11 runs

Run phospho-app/act-policy with an API

Use one of our client libraries to get started quickly. Clicking on a library will take you to the Playground tab where you can tweak different inputs, see the results, and copy the corresponding code to use in your own project.

Input schema

The fields you can use to run this model with an API. If you don't give a value for a field its default value will be used.

Field	Type	Default value	Description
dataset_repo_id	string		Hugging Face dataset ID to train on, LeRobot format > v2.0 expected, i.e. 'LegrandFrederic/dual-setup'
hugging_face_token	string		Hugging Face API token (used to upload your trained model to your HF profile), find yours here: https://huggingface.co/settings/token
hf_model_name	string		Hugging Face model name, the name of the model to be created, i.e. 'trained-act-replicate'
wandb_api_key	string		Weights & Biases API key (optional, to track the online training), find yours here: https://wandb.ai/authorize
wandb_project	string	ACT-Replicate	Weights & Biases project name
job_name	string	replicate_training	Name for this training job
steps	integer	100000 Min: 1 Max: 200000	Number of steps to train for
batch_size	integer	32 Min: 1 Max: 128	Batch size for training

{
  "type": "object",
  "title": "Input",
  "required": [
    "dataset_repo_id",
    "hugging_face_token"
  ],
  "properties": {
    "steps": {
      "type": "integer",
      "title": "Steps",
      "default": 100000,
      "maximum": 200000,
      "minimum": 1,
      "x-order": 6,
      "description": "Number of steps to train for"
    },
    "job_name": {
      "type": "string",
      "title": "Job Name",
      "default": "replicate_training",
      "x-order": 5,
      "description": "Name for this training job"
    },
    "batch_size": {
      "type": "integer",
      "title": "Batch Size",
      "default": 32,
      "maximum": 128,
      "minimum": 1,
      "x-order": 7,
      "description": "Batch size for training"
    },
    "hf_model_name": {
      "type": "string",
      "title": "Hf Model Name",
      "default": "",
      "x-order": 2,
      "description": "Hugging Face model name, the name of the model to be created, i.e. 'trained-act-replicate'"
    },
    "wandb_api_key": {
      "type": "string",
      "title": "Wandb Api Key",
      "format": "password",
      "default": "",
      "x-order": 3,
      "description": "Weights & Biases API key (optional, to track the online training), find yours here: https://wandb.ai/authorize",
      "x-cog-secret": true
    },
    "wandb_project": {
      "type": "string",
      "title": "Wandb Project",
      "default": "ACT-Replicate",
      "x-order": 4,
      "description": "Weights & Biases project name"
    },
    "dataset_repo_id": {
      "type": "string",
      "title": "Dataset Repo Id",
      "x-order": 0,
      "description": "Hugging Face dataset ID to train on, LeRobot format > v2.0 expected, i.e. 'LegrandFrederic/dual-setup'"
    },
    "hugging_face_token": {
      "type": "string",
      "title": "Hugging Face Token",
      "format": "password",
      "x-order": 1,
      "description": "Hugging Face API token (used to upload your trained model to your HF profile), find yours here: https://huggingface.co/settings/token",
      "x-cog-secret": true
    }
  }
}

Output schema

The shape of the response you’ll get when you run this model with an API.

Schema

{
  "type": "string",
  "title": "Output",
  "format": "uri"
}