microsoft--agent-lightning/examples/spider

Spider Example

This example requires a single node with one GPU of at least 40GB memory.

1. Download Spider 1.0 dataset from here and unzip it to the data folder.
2. Use python spider_eval/convert_dataset.py to convert the dataset to the parquet format.
3. Start ray: bash ../../scripts/restart_ray.sh. To use Wandb, you need to set the WANDB_API_KEY environment variable before starting ray.
4. Run the agent: VERL_API_BASE=http://localhost:9999/ python sql_agent.py. Use python sql_agent.py --help to see options like running multiple agents.
5. In another terminal, launch the training server: bash train.sh.

Evaluation

Results are coming soon.