Karla Saur
Karla Saur
In working on #226 , I tried to run some of the user-code: ``` onnx_ml_model = convert_xgboost(model, initial_types=initial_types, target_opset=9) onnx_model = convert(onnx_ml_model, "onnx", X) session = ort.InferenceSession(onnx_model.SerializeToString()) ``` which, with...
In #269, the basics are implemented, but to be fully complete we need: (1) degree larger than 2 (2) support for interaction_only (3) more tests
Similar to the [LightGBM converter integration](https://cloudblogs.microsoft.com/opensource/2020/09/29/accelerate-machine-learning-models-gpu-onnx-runtime-hummingbird/) which integrated Hummingbird into onnxmltools, integrate Hummingbird with the [XGBoost converter](https://github.com/onnx/onnxmltools/tree/master/onnxmltools/convert/xgboost). The code will likely look somewhat similar to [the LightGBM](https://github.com/onnx/onnxmltools/commit/8cf85d743024ba17803c910f7b0dc4fece701cda) work (hopefully)... A...
If a user forgets to call `fit` before calling `convert` they will get a confusing error message. We should find a way to print a more helpful error message in...
The existing CountVectorizer code has jit things such as in the forward function ``` doc_ids = torch.jit.annotate(List[Tensor], []) # noqa: F821 ``` which we need to do a bit of...
The test I wrote for OHE strings throws the following error: ``` onnxruntime.capi.onnxruntime_pybind11_state.InvalidGraph: [ONNXRuntimeError] : 10 : INVALID_GRAPH : This is an invalid model. Type Error: Type 'tensor(bool)' of input...
With mulitclass datasets (such as covtype or iris), sometimes we get errors on rounding: ```python import numpy as np import torch from hummingbird import convert_sklearn from sklearn.ensemble import RandomForestClassifier from...
This is a little tricky because we're not going to be able to run the GPU tests. But maybe that there is something we can do to make sure that...
From the new [microsoft/hummingbird/security/dependabot/1](https://github.com/microsoft/hummingbird/security/dependabot/1), we need to update `protobuf>=3.20.2` in `setup.py`. However, this version doesn't work yet with onnx (See PR #635) : ``` × Encountered error while trying to...