tharunpranavsakthivel/tinyshell-falcon-h1-90m

Falcon-H1 TinyShell

Fine-tuned compact language model for the TinyShell ShellIntent natural-language-to-structured-IR task.

Base model

tiiuae/Falcon-H1-Tiny-Tool-Calling-90M

This is a supervised fine-tune of the upstream Falcon-H1 Tiny tool-calling model. The derivative is distributed under the upstream Falcon-LLM License; see LICENSE for the authoritative license link.

Task

The model converts natural-language instructions into structured TinyShell ShellIntent JSON.

Supported high-level decisions include:

• —compile
• —clarify
• —unsupported

Held-out evaluation

Generation policy

JSON-completion stopping criterion

FunctionGemma and Falcon-H1 initially produced a valid first JSON object but frequently continued generating additional content. Their corrected final evaluation uses a generation-time stopping criterion that terminates once the first complete top-level JSON object is generated. This is generation control, not post-hoc JSON repair.

LFM2.5 terminated correctly under the original inference configuration.

Training

The model was fine-tuned with supervised causal language modeling.

• —Seed: 42
• —Best validation loss: 0.11050145771034295
• —Training time: 2314.405442636 seconds
• —Peak GPU memory: 2.0683422088623047 GB

Prompt tokens were masked from the language-model loss and the assistant JSON response was used as the supervised target.

Training used 1,600 examples, with 200 validation examples and 200 held-out test examples. The random seed was 42. The frozen source hashes and complete training metadata are included in evaluation/training_result.json.

Included files

• —Fine-tuned model weights
• —Model configuration
• —Tokenizer / processor files
• —Chat template when saved
• —Generation configuration when saved
• —evaluation/final_metrics.json
• —evaluation/test_predictions.jsonl
• —evaluation/training_result.json
• —inference_example.py
• —requirements.txt
• —LICENSE and NOTICE
• —SHA256SUMS.txt

Limitations

This pilot used one training seed. Test-set bootstrap intervals quantify held-out sample uncertainty but do not replace independent repeated training.

Exact ShellIntent matching is intentionally strict: one incorrect operation, argument, or structured field makes the complete IR prediction incorrect.

This model emits untrusted structured intent. Do not execute model output directly. Validate the JSON against the TinyShell schema, compile it through a deterministic platform-aware compiler, apply safety checks, and require user confirmation where appropriate.

License

The model weights are a derivative of tiiuae/Falcon-H1-Tiny-Tool-Calling-90M and are subject to the Falcon-LLM License. The license link is also recorded in LICENSE.

The TinyShell training data contribution is attributed under CC BY 4.0. Upstream source material may have separate terms; see the TinyShell dataset documentation for details.