Datasets:
Download experiments/gnn_generation_analysis.yaml from timlawrenz/gnn-ruby-code-study: direct link, hf CLI and curl.
- Browser
- Download file 3.48 kB
-
https://huggingface.co/datasets/timlawrenz/gnn-ruby-code-study/resolve/main/experiments/gnn_generation_analysis.yaml
- Command line
-
hf download hf://datasets/timlawrenz/gnn-ruby-code-study/experiments/gnn_generation_analysis.yaml
-
curl -L -o gnn_generation_analysis.yaml https://huggingface.co/datasets/timlawrenz/gnn-ruby-code-study/resolve/main/experiments/gnn_generation_analysis.yaml
3.48 kB
| # Fleet Spec: GNN Code Generation Failure Analysis (Track 2) | |
| # | |
| # Systematic analysis of code generation quality across decoder configurations. | |
| # Tests: loss functions × decoder architectures × hidden dimensions. | |
| # | |
| # Launch: | |
| # ratiocinator fleet run experiments/gnn_generation_analysis.yaml | |
| name: gnn-generation-analysis | |
| description: "Systematic GNN code generation failure analysis" | |
| hardware: | |
| gpu: "RTX 4090" | |
| num_gpus: 1 | |
| min_cpu_ram_gb: 32 | |
| min_inet_down: 1000.0 | |
| min_cuda_version: 12.0 | |
| max_dph: 0.40 | |
| disk_gb: 50.0 | |
| image: pytorch/pytorch:2.7.0-cuda12.8-cudnn9-runtime | |
| repo: | |
| url: https://github.com/timlawrenz/jubilant-palm-tree.git | |
| branch: experiment/ratiocinator-gnn-study | |
| clone_depth: 1 | |
| data: | |
| source: none | |
| deps: | |
| pre_install: | |
| - "apt-get update -qq && apt-get install -y -qq git-lfs > /dev/null 2>&1 || true" | |
| - "cd /workspace/experiment && git lfs install && git lfs pull" | |
| - "pip install torch-geometric torch-scatter torch-sparse -f https://data.pyg.org/whl/torch-2.7.0+cu128.html" | |
| - "pip install pandas tqdm sentence-transformers nltk scikit-learn numpy" | |
| requirements: requirements.txt | |
| exclude_from_requirements: | |
| - torch | |
| - torchvision | |
| - torch_geometric | |
| verify: "python -c \"import torch_geometric; print(f'PyG {torch_geometric.__version__}')\"" | |
| arms: | |
| # ── Loss function comparison (GAT decoder baseline) ── | |
| - name: improved-loss-gat | |
| description: "Improved (cross-entropy) loss, GAT decoder" | |
| command: "bash scripts/run_generation_arm.sh" | |
| env: | |
| DECODER_CONV_TYPE: "GAT" | |
| LOSS_FN: "improved" | |
| HIDDEN_DIM: "256" | |
| NUM_LAYERS: "5" | |
| EPOCHS: "30" | |
| - name: simple-loss-gat | |
| description: "Simple (MSE) loss, GAT decoder" | |
| command: "bash scripts/run_generation_arm.sh" | |
| env: | |
| DECODER_CONV_TYPE: "GAT" | |
| LOSS_FN: "simple" | |
| HIDDEN_DIM: "256" | |
| NUM_LAYERS: "5" | |
| EPOCHS: "30" | |
| - name: comprehensive-loss-gat | |
| description: "Comprehensive (combined) loss, GAT decoder" | |
| command: "bash scripts/run_generation_arm.sh" | |
| env: | |
| DECODER_CONV_TYPE: "GAT" | |
| LOSS_FN: "comprehensive" | |
| HIDDEN_DIM: "256" | |
| NUM_LAYERS: "5" | |
| EPOCHS: "30" | |
| # ── Decoder architecture comparison (improved loss) ── | |
| - name: improved-loss-sage | |
| description: "Improved loss, SAGE decoder" | |
| command: "bash scripts/run_generation_arm.sh" | |
| env: | |
| DECODER_CONV_TYPE: "SAGE" | |
| LOSS_FN: "improved" | |
| HIDDEN_DIM: "256" | |
| NUM_LAYERS: "5" | |
| EPOCHS: "30" | |
| - name: improved-loss-gin | |
| description: "Improved loss, GIN decoder" | |
| command: "bash scripts/run_generation_arm.sh" | |
| env: | |
| DECODER_CONV_TYPE: "GIN" | |
| LOSS_FN: "improved" | |
| HIDDEN_DIM: "256" | |
| NUM_LAYERS: "5" | |
| EPOCHS: "30" | |
| - name: improved-loss-gcn | |
| description: "Improved loss, GCN decoder" | |
| command: "bash scripts/run_generation_arm.sh" | |
| env: | |
| DECODER_CONV_TYPE: "GCN" | |
| LOSS_FN: "improved" | |
| HIDDEN_DIM: "256" | |
| NUM_LAYERS: "5" | |
| EPOCHS: "30" | |
| # ── Wider decoder ── | |
| - name: improved-loss-gat-wide | |
| description: "Improved loss, GAT decoder, hidden_dim=512" | |
| command: "bash scripts/run_generation_arm.sh" | |
| env: | |
| DECODER_CONV_TYPE: "GAT" | |
| LOSS_FN: "improved" | |
| HIDDEN_DIM: "512" | |
| NUM_LAYERS: "5" | |
| EPOCHS: "30" | |
| metrics: | |
| protocol: json_line | |
| json_prefix: "METRICS:" | |
| budget: | |
| max_dollars: 10.00 | |
| train_timeout_s: 3600 | |
| download_timeout_s: 600 | |