Datasets:

FranckAbgrall
/

glm53-flash-function-calling

messages listlengths 3 3	tools listlengths 4 4	domain stringclasses 25 values	kind stringclasses 3 values
[ { "role": "system", "content": "You are a helpful assistant with access to functions. Use them if they are relevant to the user's request.", "tool_calls": null }, { "role": "user", "content": "Can you email the parents of my third-period class about the field trip to the Boston Museum of Sci...	[ { "type": "function", "function": { "name": "create_task", "description": "Creates a new task within a specified project and assigns it to a team member.", "parameters": { "type": "object", "properties": { "project_id": { "type": "string", ...	task and project management	no_tool
[ { "role": "system", "content": "You are a helpful assistant with access to functions. Use them if they are relevant to the user's request.", "tool_calls": null }, { "role": "user", "content": "Please create a note for tomorrow's parent-teacher conference covering Marcus Delgado's reading pro...	[ { "type": "function", "function": { "name": "create_note", "description": "Creates a new note in the user's workspace and returns the created note object including its ID.", "parameters": { "type": "object", "properties": { "project_id": null, "title": {...	note-taking and documents	single
[ { "role": "system", "content": "You are a helpful assistant with access to functions. Use them if they are relevant to the user's request.", "tool_calls": null }, { "role": "user", "content": "Night shift just ended and I need to fill up the tank and grab my prescription on the way home. Aro...	[ { "type": "function", "function": { "name": "search_nearby_places", "description": "Searches for places (restaurants, gas stations, ATMs, etc.) near a given coordinate, with optional category and radius filters.", "parameters": { "type": "object", "properties": { ...	maps and navigation	parallel
[ { "role": "system", "content": "You are a helpful assistant with access to functions. Use them if they are relevant to the user's request.", "tool_calls": null }, { "role": "user", "content": "hey can you play my lo-fi study playlist on the speaker in my dorm room? i need to cram for my midt...	[ { "type": "function", "function": { "name": "create_automation", "description": "Creates a new automation rule that triggers smart home actions based on a schedule, sensor event, or device state change.", "parameters": { "type": "object", "properties": { "project_...	smart home control	no_tool
[ { "role": "system", "content": "You are a helpful assistant with access to functions. Use them if they are relevant to the user's request.", "tool_calls": null }, { "role": "user", "content": "Set up a new repo for the regional pipeline tracker: name it 'northeast-pipeline-tracker', owner 'v...	[ { "type": "function", "function": { "name": "create_repository", "description": "Creates a new code repository under a specified organization or user account.", "parameters": { "type": "object", "properties": { "project_id": null, "title": null, ...	git and code repository management	single
[ { "role": "system", "content": "You are a helpful assistant with access to functions. Use them if they are relevant to the user's request.", "tool_calls": null }, { "role": "user", "content": "Translate 'The doctor said my appointment is next Tuesday at the clinic' into Japanese.", "tool...	[ { "type": "function", "function": { "name": "translate_text", "description": "Translates a text string from a source language to a target language, with optional formality level and automatic source-language detection.", "parameters": { "type": "object", "properties": { ...	language translation	single
[ { "role": "system", "content": "You are a helpful assistant with access to functions. Use them if they are relevant to the user's request.", "tool_calls": null }, { "role": "user", "content": "Delete the old staff schedule file 'schedules/staff-roster-sept.xlsx' from storage — move it to tra...	[ { "type": "function", "function": { "name": "upload_file_to_storage", "description": "Uploads a file to a specified cloud storage bucket or folder, optionally overwriting existing files.", "parameters": { "type": "object", "properties": { "project_id": null, ...	file and cloud storage management	single
[{"role":"system","content":"You are a helpful assistant with access to functions. Use them if they (...TRUNCATED)	[{"type":"function","function":{"name":"execute_sql_query","description":"Executes a read-only SQL S(...TRUNCATED)	database querying	single
[{"role":"system","content":"You are a helpful assistant with access to functions. Use them if they (...TRUNCATED)	[{"type":"function","function":{"name":"create_task","description":"Creates a new task within a spec(...TRUNCATED)	task and project management	single
[{"role":"system","content":"You are a helpful assistant with access to functions. Use them if they (...TRUNCATED)	[{"type":"function","function":{"name":"create_repository","description":"Creates a new code reposit(...TRUNCATED)	git and code repository management	parallel

End of preview. Expand in Data Studio

GLM-5.3-Flash Synthetic Function-Calling Dataset

298 synthetic function-calling examples generated by zai-org/GLM-5.3-Flash via Hugging Face Inference Providers, in OpenAI chat format (messages + tools, assistant turns use tool_calls). Ready for SFT with TRL / transformers chat templates.

Composition

kind	count	meaning
`single`	175	user request fulfilled by exactly one tool call
`parallel`	73	request requires 2+ tool calls
`no_tool`	50	no relevant tool exists; assistant politely declines

Examples span 25 domains (calendar, email, weather, maps, smart home, finance, travel, e-commerce, git, databases, CRM, HR, healthcare, IoT, …). Each record carries its domain and kind for filtering.

How it was generated

Two-stage, fully synthetic (no human-written tools or examples):

Tool design — for each domain, the model designed 4 tools in OpenAI function format (name, description, JSON-schema parameters). Schemas were checked with jsonschema.Draft7Validator.check_schema.
Example writing — for each domain × kind, the model wrote user requests plus tool calls, seeded with a random persona hint for diversity.

Every example was then programmatically validated: JSON parses, called function names exist in the record's tools, and arguments validate against the tool's JSON schema. Invalid items were dropped (2 of 300 target). Near-duplicate user requests were removed.

The generator script is included in this repo: gen_fc_dataset.py. Run it with HF_TOKEN=... python gen_fc_dataset.py (SMOKE=1 for a quick test).

Format

{
  "messages": [
    {"role": "system", "content": "You are a helpful assistant with access to functions. ..."},
    {"role": "user", "content": "Email the three caterers about Friday's lunch..."},
    {"role": "assistant", "content": null, "tool_calls": [
      {"id": "call_1a2b3c4d", "type": "function",
       "function": {"name": "send_email", "arguments": "{\"to\": [...], ...}"}}
    ]}
  ],
  "tools": [{"type": "function", "function": {"name": "send_email", "description": "...", "parameters": {...}}}],
  "domain": "email management",
  "kind": "single"
}

Note: tool_calls[].function.arguments is a JSON-encoded string (OpenAI convention); no_tool records have a plain assistant content reply and no tool_calls.

Usage

from datasets import load_dataset

ds = load_dataset("FranckAbgrall/glm53-flash-function-calling", split="train")
single_only = ds.filter(lambda r: r["kind"] == "single")

Limitations

Fully synthetic: tool schemas and argument values are plausible but not tied to real APIs; dates center on October 2026 (generation time).
English only; ~300 rows is a pipeline-validation / small-SFT scale, not a full training corpus.
Generated with temperature 1.0 for diversity — expect occasional stylistic quirks; validation guarantees schema correctness, not real-world executability.

Downloads last month: 107