Skip to content

feat: add ONNX ML inference support for Render deployment - #25

Merged
SamoraDC merged 2 commits into
mainfrom
fix-rust
Jan 22, 2026
Merged

feat: add ONNX ML inference support for Render deployment#25
SamoraDC merged 2 commits into
mainfrom
fix-rust

Conversation

@SamoraDC

Copy link
Copy Markdown
Owner
  • Export LightGBM and XGBoost models to ONNX format (332 features each)
  • Update Dockerfile with ONNX Runtime 1.19.2 for hot-path inference
  • Add ML_ENABLED and ONNX_MODEL_DIR env vars to render.yaml
  • Create export_onnx.py script for model conversion
  • Fix XGBoost feature naming for onnxmltools compatibility

Models exported:

  • lightgbm_model.onnx (2.3MB)
  • xgboost_model.onnx (2.1MB)

SamoraDC and others added 2 commits January 22, 2026 13:06
- Export LightGBM and XGBoost models to ONNX format (332 features each)
- Update Dockerfile with ONNX Runtime 1.19.2 for hot-path inference
- Add ML_ENABLED and ONNX_MODEL_DIR env vars to render.yaml
- Create export_onnx.py script for model conversion
- Fix XGBoost feature naming for onnxmltools compatibility

Models exported:
- lightgbm_model.onnx (2.3MB)
- xgboost_model.onnx (2.1MB)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- nsmi.rs: Use .clamp(2, 5) instead of .max(2).min(5)
- nsmi.rs: Use iterator zip pattern instead of index loops
- ml_inference.rs: Rename ModelType variants (LSTM->Lstm, CNN->Cnn, etc.)
- Add #[allow(dead_code)] for modules not yet integrated
- Add #[allow(unused_imports)] for conditional exports

All fixes maintain identical runtime behavior while satisfying
Clippy warnings-as-errors (-D warnings) in CI.

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
@SamoraDC
SamoraDC merged commit 664166f into main Jan 22, 2026
6 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant