Spaces:
Running
Running
| # Tools | |
| Reusable scripts that are meant to be run from the repo root. | |
| ## `sync-readme-stats` | |
| Updates `README.md` with live test and benchmark totals from the current repo. | |
| Usage: | |
| ```bash | |
| ./tools/sync-readme-stats | |
| ``` | |
| ## `model_lab_frontier.py` | |
| Runs the ShopStack frontier planner sweep against current Hugging Face models and writes: | |
| - a JSONL results file | |
| - a markdown summary report | |
| Default models: | |
| - `qwen3.5-9b` -> `Qwen/Qwen3.5-9B` | |
| - `qwen3.6-35b-a3b` -> `Qwen/Qwen3.6-35B-A3B` | |
| Usage: | |
| ```bash | |
| uv run python tools/model_lab_frontier.py | |
| ``` | |
| Override the output directory: | |
| ```bash | |
| SHOPSTACK_MODEL_LAB_OUTPUT_DIR=benchmarks/model_lab/results uv run python tools/model_lab_frontier.py | |
| ``` | |
| Override the models: | |
| ```bash | |
| uv run python tools/model_lab_frontier.py --model qwen3.5-9b --model qwen3.6-35b-a3b | |
| ``` | |