LocalAgent โ€” tiny 30M byte-level tool-calling agent

A 28M-parameter, pretrained-from-scratch, byte-level agent for tool dispatch. Selection is generable (no fixed-N classifier): a 5-way route head gates modality, a dense two-tower selector scores any tool by its description embedding, and a pointer head copies argument values from the prompt. 50-tool surface.

  • Free-form out-of-distribution call-name ~57%; multi-turn next-tool selection ~74%.
  • Pure PyTorch (no transformers). Load with this repo's LocalAgentLM / ModelConfig.

Files

  • model.pt โ€” PyTorch checkpoint (backbone + ptr_head + dense_selector + route_head + examples).
  • model.fp16.onnx โ€” ONNX graph (logits, hidden) for the in-browser demo.
  • dispatch_heads.json / heads.json / meta.json โ€” head weights + tokenizer/tool contract.

Demo: https://huggingface.co/spaces/danelcsb/localagent-webgpu ยท Code: https://github.com/sangbumchoi/localagent

Downloads last month
56
Safetensors
Model size
28.3M params
Tensor type
F32
ยท
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Space using danelcsb/localagent-tiny-30m-byte 1