$ loading_
Manage ML experiments, training runs, and reports across GPU environments.
Copy the install command and let the AI configure it · recommended for beginners
No copy-paste install info for "projtool" yet — see the docs or source repo.
Use projtool to create a new experiment for an image classification project named resnet50-baseline on a remote AutoDL GPU instance, set the dataset path, learning rate to 0.001, batch size to 64, and launch training immediately.
Returns the experiment creation result, training job launch status, and a summary of key configurations.
Use projtool to check the latest run status of the resnet50-baseline experiment, poll it periodically, and summarize the current epoch, loss, accuracy, whether any errors occurred, and the GPU instance status.
Outputs the latest experiment progress, key metric changes, error information, and a summary of the runtime environment status.
Use projtool to generate an experiment report from the past week's training runs, compare results across different hyperparameter settings, and summarize the best model performance, failure reasons, and next optimization suggestions.
Generates a structured experiment report with result comparisons, the best setup, issue analysis, and next-step recommendations.
Profile a codebase to generate guides, context pages, and a searchable vector store.
Read, analyze, create, and edit ProjectLibre project files.
Manage GPU training jobs end-to-end with natural language.
Turn SSH training-server operations into AI-callable tools for lab ops.
Build persistent project memory for AI coding assistants across logs, rules, TODOs, and handoffs.
Make project docs instantly accessible so Claude Code understands architecture and conventions.