datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
mobile-actions
Mobile Actions: A Dataset for On-Device Function Calling
The dataset contains conversational traces designed to train lightweight models (such as FunctionGemma 270M) to translate natural language instructions into executable function calls for Android OS system tools.
Dataset Format
The dataset is provided in JSONL format. Each line represents a data sample. The
dataset is pre-split into training and evaluation sets. This distinction is
denoted by the metadata field… See the full description on the dataset page: https://huggingface.co/datasets/google/mobile-actions.Mobile-Actions-Combined
Mobile Action
This dataset is meade by concatenating these two mobile-actions datasets: Extended Mobile Actions
and Google Mobile Actions.
Dataset Format
Record's fields:
metadata: A flag to determine the sample type assigned among training and evaluation datasets.
tools: Contains a list of available tools (functions) schema that the model literally is able to call and made of the following elements:
function: An object corresponding to a function:
name: name… See the full description on the dataset page: https://huggingface.co/datasets/UGrowAI/Mobile-Actions-Combined.mobile-actions-merged
Mobile Action
This dataset is meade by concatenating these two mobile-actions datasets: Extended Mobile Actions
and Google Mobile Actions.
Dataset Format
Record's fields:
metadata: A flag to determine the sample type assigned among training and evaluation datasets.
tools: Contains a list of available tools (functions) schema that the model literally is able to call and made of the following elements:
function: An object corresponding to a function:
name: name… See the full description on the dataset page: https://huggingface.co/datasets/AliRGHZ/mobile-actions-merged.mobile-actions-ita
Dataset Card: Mobile Actions (Italian Adaptation for Function Calling)
Overview
This dataset is an Italian adaptation of the original Google Mobile Actions dataset, designed to train lightweight models for on-device function calling. It preserves the original tool-calling schema in English while translating user interactions and contextual instructions into Italian.
The goal is to enable models to map natural language instructions in Italian to structured function calls… See the full description on the dataset page: https://huggingface.co/datasets/Mattimax/mobile-actions-ita.Mobile-Actions-Extended
Mobile Action
Mobile Actions is the dataset designed to fine-tune function calling models such as FunctionGemma 270M over mobile functionalities.
The dataset contains conversational traces over current 15 Android OS system capabilities.
Fine-tuned model is able to execute on-device's functions with the following provided tools:
- Turning the flashlight on
- Turning the flashlight off
- Send email
- Check battery status
- Check bluetooth status
- Make a phone call
- Send sms… See the full description on the dataset page: https://huggingface.co/datasets/UGrowAI/Mobile-Actions-Extended.Mobile-Actions
Mobile Action
Mobile Actions is the dataset designed to fine-tune function calling models such as FunctionGemma 270M over mobile functionalities.
The dataset contains conversational traces over current 15 Android OS system capabilities.
Fine-tuned model is able to execute on-device's functions with the following provided tools:
- Turning the flashlight on
- Turning the flashlight off
- Send email
- Check battery status
- Check bluetooth status
- Make a phone call
- Send sms… See the full description on the dataset page: https://huggingface.co/datasets/AliRGHZ/Mobile-Actions.mobile-actions
Mobile Actions: A Dataset for On-Device Function Calling
The dataset contains conversational traces designed to train lightweight models (such as FunctionGemma 270M) to translate natural language instructions into executable function calls for Android OS system tools.
Dataset Format
The dataset is provided in JSONL format. Each line represents a data sample. The
dataset is pre-split into training and evaluation sets. This distinction is
denoted by the metadata… See the full description on the dataset page: https://huggingface.co/datasets/0xmoose0xmoose0xmoose/mobile-actions.mobile-actions
Mobile Actions: A Dataset for On-Device Function Calling
The dataset contains conversational traces designed to train lightweight models (such as FunctionGemma 270M) to translate natural language instructions into executable function calls for Android OS system tools.
Dataset Format
The dataset is provided in JSONL format. Each line represents a data sample. The
dataset is pre-split into training and evaluation sets. This distinction is
denoted by the metadata field… See the full description on the dataset page: https://huggingface.co/datasets/jeffbrian/mobile-actions.arabic-mobile-actions
Arabic Mobile Actions Dataset
Arabic Function Calling dataset reformatted for FunctionGemma fine-tuning.
This dataset converts the Arabic Function Calling Dataset to the Google Mobile Actions format, enabling fine-tuning of FunctionGemma for Arabic on-device function calling.
Dataset Statistics
Metric
Value
Total samples
45,729
Train samples
36,583 (80%)
Eval samples
9,146 (20%)
Positive (with tool call)
41,175
Negative (no tool call)
4,554… See the full description on the dataset page: https://huggingface.co/datasets/Sa74ll/arabic-mobile-actions.mobile-actions
Mobile Actions: A Dataset for On-Device Function Calling
The dataset contains conversational traces designed to train lightweight models (such as FunctionGemma 270M) to translate natural language instructions into executable function calls for Android OS system tools.
Dataset Format
The dataset is provided in JSONL format. Each line represents a data sample. The
dataset is pre-split into training and evaluation sets. This distinction is
denoted by the metadata field… See the full description on the dataset page: https://huggingface.co/datasets/lanneret/mobile-actions.mobile-actions
Mobile Actions: A Dataset for On-Device Function Calling
The dataset contains conversational traces designed to train lightweight models (such as FunctionGemma 270M) to translate natural language instructions into executable function calls for Android OS system tools.
Dataset Format
The dataset is provided in JSONL format. Each line represents a data sample. The
dataset is pre-split into training and evaluation sets. This distinction is
denoted by the metadata field… See the full description on the dataset page: https://huggingface.co/datasets/burancel/mobile-actions.tool-reasoning-sft-TOOLS-mobile-actions-data-cleaned-rectified
Mobile Actions — Cleaned & Rectified
8.7K on-device function calling conversations converted into a strict reasoning + tool-call format. Covers 7 Android mobile actions including calendar events, emails, contacts, maps, flashlight, and Wi-Fi settings.
Format
Each row contains a structured conversation with explicit reasoning traces and validated tool calls.
Message Roles
Role
Content
system
Tool-use protocol + cleaned JSON tool schemas +… See the full description on the dataset page: https://huggingface.co/datasets/AmanPriyanshu/tool-reasoning-sft-TOOLS-mobile-actions-data-cleaned-rectified.mobile-actions
Mobile Actions: A Dataset for On-Device Function Calling
The dataset contains conversational traces designed to train lightweight models (such as FunctionGemma 270M) to translate natural language instructions into executable function calls for Android OS system tools.
Dataset Format
The dataset is provided in JSONL format. Each line represents a data sample. The
dataset is pre-split into training and evaluation sets. This distinction is
denoted by the metadata field… See the full description on the dataset page: https://huggingface.co/datasets/Macarao26722/mobile-actions.mobile-actions
Mobile Actions: A Dataset for On-Device Function Calling
The dataset contains conversational traces designed to train lightweight models (such as FunctionGemma 270M) to translate natural language instructions into executable function calls for Android OS system tools.
Dataset Format
The dataset is provided in JSONL format. Each line represents a data sample. The
dataset is pre-split into training and evaluation sets. This distinction is
denoted by the metadata field… See the full description on the dataset page: https://huggingface.co/datasets/mindchain/mobile-actions.mobile-actions
Mobile Actions: A Dataset for On-Device Function Calling
The dataset contains conversational traces designed to train lightweight models (such as FunctionGemma 270M) to translate natural language instructions into executable function calls for Android OS system tools.
Dataset Format
The dataset is provided in JSONL format. Each line represents a data sample. The
dataset is pre-split into training and evaluation sets. This distinction is
denoted by the metadata field… See the full description on the dataset page: https://huggingface.co/datasets/HexQuant/mobile-actions.mobile-actions
Mobile Actions: A Dataset for On-Device Function Calling
The dataset contains conversational traces designed to train lightweight models (such as FunctionGemma 270M) to translate natural language instructions into executable function calls for Android OS system tools.
Dataset Format
The dataset is provided in JSONL format. Each line represents a data sample. The
dataset is pre-split into training and evaluation sets. This distinction is
denoted by the metadata field… See the full description on the dataset page: https://huggingface.co/datasets/amine-khelif/mobile-actions.mobile-actions
Mobile Actions: A Dataset for On-Device Function Calling
The dataset contains conversational traces designed to train lightweight models (such as FunctionGemma 270M) to translate natural language instructions into executable function calls for Android OS system tools.
Dataset Format
The dataset is provided in JSONL format. Each line represents a data sample. The
dataset is pre-split into training and evaluation sets. This distinction is
denoted by the metadata field… See the full description on the dataset page: https://huggingface.co/datasets/muralcode/mobile-actions.mobile-actions
Mobile Actions: A Dataset for On-Device Function Calling
The dataset contains conversational traces designed to train lightweight models (such as FunctionGemma 270M) to translate natural language instructions into executable function calls for Android OS system tools.
Dataset Format
The dataset is provided in JSONL format. Each line represents a data sample. The
dataset is pre-split into training and evaluation sets. This distinction is
denoted by the metadata field… See the full description on the dataset page: https://huggingface.co/datasets/tatan2/mobile-actions.mobile-actions
Mobile Actions: A Dataset for On-Device Function Calling
The dataset contains conversational traces designed to train lightweight models (such as FunctionGemma 270M) to translate natural language instructions into executable function calls for Android OS system tools.
Dataset Format
The dataset is provided in JSONL format. Each line represents a data sample. The
dataset is pre-split into training and evaluation sets. This distinction is
denoted by the metadata field… See the full description on the dataset page: https://huggingface.co/datasets/JesseJelinek/mobile-actions.mobile-actions
Mobile Actions: A Dataset for On-Device Function Calling
The dataset contains conversational traces designed to train lightweight models (such as FunctionGemma 270M) to translate natural language instructions into executable function calls for Android OS system tools.
Dataset Format
The dataset is provided in JSONL format. Each line represents a data sample. The
dataset is pre-split into training and evaluation sets. This distinction is
denoted by the metadata field… See the full description on the dataset page: https://huggingface.co/datasets/deougege/mobile-actions.mobile-actions-language-modeling
Mobile Actions SFT Dataset
A converted version of the google/mobile-actions dataset for supervised fine-tuning (SFT) of Qwen models with tool calling capabilities.
Dataset Description
This dataset is derived from the google/mobile-actions dataset, which contains human-AI conversations about performing actions on mobile devices. The original dataset has been converted to the Qwen chat template format for efficient training of Qwen models.
Conversion Process
The… See the full description on the dataset page: https://huggingface.co/datasets/niwang66/mobile-actions-language-modeling.mobile-actions
Mobile Actions: A Dataset for On-Device Function Calling
The dataset contains conversational traces designed to train lightweight models (such as FunctionGemma 270M) to translate natural language instructions into executable function calls for Android OS system tools.
Dataset Format
The dataset is provided in JSONL format. Each line represents a data sample. The
dataset is pre-split into training and evaluation sets. This distinction is
denoted by the metadata field… See the full description on the dataset page: https://huggingface.co/datasets/Rendy45/mobile-actions.liquidchat-mobile-actions-2026-02-24
liquidchat-mobile-actions-2026-02-24
Chat history dataset exported from LiquidChat mobile app.
Dataset Description
This dataset contains conversations with a mobile-actions fine-tuned LLM, including tool calls (flashlight, calendar, email, maps, contacts, wifi) and their results.
Usage
from datasets import load_dataset
dataset = load_dataset("kshitijthakkar/liquidchat-mobile-actions-2026-02-24")
Created by: kshitijthakkar
Exported from: LiquidChat
License:… See the full description on the dataset page: https://huggingface.co/datasets/kshitijthakkar/liquidchat-mobile-actions-2026-02-24.mobile-actions
Mobile Actions: A Dataset for On-Device Function Calling
The dataset contains conversational traces designed to train lightweight models (such as FunctionGemma 270M) to translate natural language instructions into executable function calls for Android OS system tools.
Dataset Format
The dataset is provided in JSONL format. Each line represents a data sample. The
dataset is pre-split into training and evaluation sets. This distinction is
denoted by the metadata field… See the full description on the dataset page: https://huggingface.co/datasets/ettilapse/mobile-actions.mobile-actions
Mobile Actions: A Dataset for On-Device Function Calling
The dataset contains conversational traces designed to train lightweight models (such as FunctionGemma 270M) to translate natural language instructions into executable function calls for Android OS system tools.
Dataset Format
The dataset is provided in JSONL format. Each line represents a data sample. The
dataset is pre-split into training and evaluation sets. This distinction is
denoted by the metadata field… See the full description on the dataset page: https://huggingface.co/datasets/Rendra86318/mobile-actions.mobile-actions2
Mobile Actions: A Dataset for On-Device Function Calling
The dataset contains conversational traces designed to train lightweight models (such as FunctionGemma 270M) to translate natural language instructions into executable function calls for Android OS system tools.
Dataset Format
The dataset is provided in JSONL format. Each line represents a data sample. The
dataset is pre-split into training and evaluation sets. This distinction is
denoted by the metadata… See the full description on the dataset page: https://huggingface.co/datasets/0xmoose0xmoose0xmoose/mobile-actions2.liquidchat-mobile-actions-2026-02-23
liquidchat-mobile-actions-2026-02-23
Chat history dataset exported from LiquidChat mobile app.
Dataset Description
This dataset contains conversations with a mobile-actions fine-tuned LLM, including tool calls (flashlight, calendar, email, maps, contacts, wifi) and their results.
Usage
from datasets import load_dataset
dataset = load_dataset("kshitijthakkar/liquidchat-mobile-actions-2026-02-23")
Created by: kshitijthakkar
Exported from: LiquidChat
License:… See the full description on the dataset page: https://huggingface.co/datasets/kshitijthakkar/liquidchat-mobile-actions-2026-02-23.mobile-actions
Mobile Actions: A Dataset for On-Device Function Calling
The dataset contains conversational traces designed to train lightweight models (such as FunctionGemma 270M) to translate natural language instructions into executable function calls for Android OS system tools.
Dataset Format
The dataset is provided in JSONL format. Each line represents a data sample. The
dataset is pre-split into training and evaluation sets. This distinction is
denoted by the metadata field… See the full description on the dataset page: https://huggingface.co/datasets/R1Sh111/mobile-actions.mobile-actions-customsmarthome-mobile-actions