Datasets:
The dataset viewer is not available because its heuristics could not detect any supported data files. You can try uploading some data files, or configuring the data files location manually.
GLM-5.3-Flash Function Calling (synthetic)
Synthetic function-calling dataset generated with zai-org/GLM-5.3-Flash via the Hugging Face Inference Providers router.
Status: generation pending. This card and the repo were created in a session with no
compute budget granted, so no generation job was run yet. Run generate_dataset.py
(also in this repo) with an HF_TOKEN to fill the configs below.
Planned configs
| config | format | contents |
|---|---|---|
train_plus |
full multi-turn trace | system, user, assistant with tool_calls (name + JSON arguments), tool role responses, final assistant answer; plus a tools column with the available tool schemas |
train_basic |
minimal single-turn | system prompt embedding the tool JSON schemas (Hermes-style), user query, assistant reply containing the tool call as a ```json block; plus a tools column |
Tool coverage
Mixed domains, ~55 tools across 12 domains: weather, calendar, e-commerce, database/SQL, code, files, math, travel, finance, communication, web, smart home. Per example, 3–6 tools from one domain (80%) or across domains (20%) are offered, and the model decides which to call — including parallel calls.
Provenance
- Generator:
zai-org/GLM-5.3-Flash(MIT licence) - Method: LLM-written user queries → native tool-call sampling → validated against JSON schema (name known, arguments parse, required params present) → deterministic simulated tool responses → LLM final answer
- Target size: 10,000 validated examples per config
- Downloads last month
- 54