History#

current - 2022-11-22 - 0.00Mb#

  • issue 470: Uses list(graph.node) when using id(node) (2022-11-22)

  • issue 471: Add one unit test to check optimisation is working (2022-11-21)

0.9.1883 - 2022-10-09 - 39.39Mb#

  • issue 469: Implements a converter for a TransformedTargetRegressor (2022-10-09)

  • issue 468: Adds debug functionalities in TfidfVectorizer (2022-10-07)

  • issue 467: Fixes TfIdfVectorizer when input is 1D (2022-10-03)

  • issue 466: Look into sequencemap.py or sequence_map.py to fetch examples (2022-09-15)

  • issue 464: Implements OnnxLoop (2022-08-18)

  • issue 465: Supports for operator DFT, STFT, *windows (2022-08-13)

  • issue 463: Fixes embedded if with XOP API (2022-08-10)

  • issue 462: Increases code coverage, improves ligthgbm converter (2022-08-05)

  • issue 461: Upgrades default supported opset to 17 (2022-08-04)

  • issue 460: Improves C++ implementation, im2col, col2im (2022-08-04)

  • issue 459: Supports OnnxOperator(…) + int or float with CastLike (2022-08-02)

  • issue 458: Changes subgraph separator from :: to :/: in onnx_simple_text_plot (2022-08-01)

  • issue 457: Fix delimiter in extras_require (2022-07-25)

0.8.1863 - 2022-07-23 - 0.79Mb#

  • issue 456: Fixes python runtime for TfIdfVectorizer (2022-07-22)

  • issue 455: Fixes division by zero in Normalizer (2022-07-21)

0.8.1858 - 2022-07-20 - 37.35Mb#

  • issue 454: Fixes compilation issues on windows and python 3.10 (2022-07-20)

  • issue 453: Uses f strings (2022-07-19)

  • issue 452: Extends code coverage (2022-07-18)

  • issue 451: Adds a table for all versions and all operators (2022-07-13)

  • issue 450: Implements node Expression to compress graph (2022-07-12)

  • issue 449: Improves code coverage (2022-07-11)

  • issue 448: Fixes template to export an onnx graph to python (2022-07-06)

  • issue 447: Adds an example to check export issues (2022-07-05)

  • issue 446: Creates an exporter to python (2022-07-05)

  • issue 445: Removes ShapeObject, replaces by OnnxShapeInference (2022-07-02)

  • issue 444: Fixes a bug in to_onnx when as_function=True (2022-07-01)

  • issue 443: Add compiled dynamic libraries to .gitignore (2022-06-30)

  • issue 442: Converts onnx with functions to code based on XOP API (2022-06-30)

  • issue 433: Converts a sklearn model into multiple functions (2022-06-29)

  • issue 441: Renames check_model into check_onnx (2022-06-25)

  • issue 440: Update azure-pipelines.yml (2022-06-22)

  • issue 439: Drops support for python 3.6 (2022-06-22)

  • issue 438: Fixes xgboost converter when base_score is specified (2022-06-22)

  • issue 437: Renders vector attributes in onnx_simple_text_plot (2022-06-21)

  • issue 436: Supports for attributes in onnx functions (2022-06-21)

  • issue 435: Extends documentation to onnxruntime (2022-06-13)

0.8.1826 - 2022-05-29 - 28.43Mb#

  • issue 432: None and [] should be different function get_tensor_shape (2022-05-25)

  • issue 431: Adds functions to change the shape of inputs and outputs (2022-05-25)

  • issue 430: Adds function to rename inputs or outputs (2022-05-23)

  • issue 429: Adds more functions to manipulate graphs (2022-05-20)

  • issue 428: Investigates SVC discrepancies (2022-05-20)

  • issue 427: Adds function to inline function on onnx graph (2022-05-12)

  • issue 426: Adds support for operator RoiAlign for python runtime (2022-04-25)

  • issue 425: Adds support for operator GridSample for python runtime (2022-04-22)

  • issue 424: Adds support for operator GRU in python runtime (2022-04-19)

  • issue 423: Adds support for Momentum for python runtime (2022-04-15)

  • issue 422: Adds support for NonMaxSuppression for python runtime (2022-04-14)

  • issue 421: Adds support for Adagrad, Adam in python runtime (2022-04-12)

  • issue 420: Adds support for operator Resize for python runtime (2022-04-10)

  • issue 419: Adds support for ThresholdedRelu for python runtime (2022-04-09)

  • issue 418: Adds support for operator OneHot, ScatterND for python runtime (2022-04-09)

  • issue 417: C++ implementation for Im2col and Col2Im in 2D (2022-04-09)

  • issue 416: Adds support for DepthToSpace and SpaceToDepth for python runtime (2022-04-08)

  • issue 415: Implements experimentation im2col (used in image convolution) (2022-04-08)

  • issue 414: Adds support for operator NonZero in python runtime (2022-04-06)

  • issue 413: Adds support for operator Shink for python runtime (2022-04-06)

  • issue 412: Adds support for DynamicQuantizeLinear for python runtime (2022-04-05)

  • issue 411: Adds support for operators Unique, SoftPlus, SoftSign for python runtime (2022-04-05)

  • issue 410: Supports operator GatherND for python runtime (2022-04-05)

  • issue 409: Fixes bug with EyeLike in python runtime (2022-04-04)

  • issue 408: Improves backtest coverage, update documentation (2022-04-03)

  • issue 407: Supports operator Hardmax for python runtime (2022-03-30)

  • issue 406: Supports operator Bernoulli for python runtime (2022-03-30)

  • issue 405: Supports operator PRelu for python runtime (2022-03-30)

  • issue 404: Fixes Trilu (2022-03-30)

  • issue 403: Supports ReduceLogSum for python runtime (2022-03-30)

  • issue 402: Supports operator Xor for python runtime (2022-03-30)

  • issue 401: Removes parameter device, adds parameter provider (2022-03-30)

  • issue 400: Supports local functions calling local functions for python runtime (2022-03-28)

  • issue 399: Supports function SoftmaxCrossEntropyLoss for python runtime (2022-03-28)

  • issue 397: Implements method f in OnnxOperatorItem (2022-03-27)

  • issue 396: Move grammar_sklearn to subfolder. (2022-03-27)

  • issue 395: Supports eager evaluation in XOP API (2022-03-27)

  • issue 394: Enables expression OnnxCos[15](…) (2022-03-26)

  • issue 393: Adds domain in function onnx_simple_text_plot (2022-03-25)

  • issue 392: Supports random operators for python runtime (2022-03-25)

  • issue 391: Adds support for onnx predefined functions for python runtime (2022-03-24)

  • issue 390: Adds support for operator HardSigmoid for python runtime (2022-03-23)

  • issue 389: Adds support for operator Selu for python runtime (2022-03-23)

  • issue 388: Adds support for operator Trilu in python runtime (2022-03-23)

  • issue 387: Supports operator Elu for python runtime (2022-03-23)

  • issue 386: Supports operator BitShift for python runtime (2022-03-23)

  • issue 384: Supports FunctionProto in XOP API. (2022-03-21)

  • issue 383: Improves python runtime for ONNX (2022-03-19)

  • issue 382: Adds one unit test to check lightgbm conversion with opsetml==3 (2022-03-18)

  • issue 381: Documentation, more notebooks on FFT (2022-03-17)

  • issue 380: Removes method get_output in xop API (2022-03-16)

  • issue 379: Improves python runtime coverage (2022-03-14)

  • issue 378: Adds function export2xop, exports onnx graph to XOP API (2022-03-12)

0.8.1762 - 2022-03-10 - 2.01Mb#

  • issue 377: Implements TreeEnsemble* for opsetml==3 (2022-03-10)

  • issue 376: Avoids one circular import. (2022-03-07)

  • issue 375: Adds code to turn onnx example into python unit test (2022-03-05)

  • issue 374: Implements onnx backend with python runtime (2022-03-05)

  • issue 372: Improves importing time (2022-03-05)

  • issue 373: Adds support for Expand in python runtime (2022-03-04)

  • issue 371: Support for ONNX functions (2022-03-04)

  • issue 370: Refactors numpy API to use Xop API (2022-03-03)

  • issue 369: Supports recursive display in onnx_simple_text_plot (2022-02-28)

  • issue 368: Updates requirements, skl2onnx>=1.11 (2022-02-28)

  • issue 367: Refactors results name in Xop API (2022-02-27)

  • issue 366: Adds python runtime for CategoryMapper (2022-02-24)

  • issue 365: Adds command line benchmark_doc (2022-02-24)

  • issue 364: Runs onnx backend test with python runtime (2022-02-23)

  • issue 363: Refactoring, moving files testing.experimental_c (2022-02-23)

  • issue 362: Adds command line plot_onnx (2022-02-23)

  • issue 361: Introduces __max_supported_opset__ and refactors the library (2022-02-23)

  • issue 360: Xop API, adds class OnnxSubOnnx to insert ONNX graph (2022-02-22)

  • issue 359: Supports domains in Xop API (2022-02-21)

  • issue 358: Extends supported operator by OnnxShapeInference (2022-02-21)

  • issue 357: Modifies OnnxShapeInference to deal with untyped outputs (2022-02-19)

  • issue 356: Supports multiple affectations (xop) (2022-02-18)

  • issue 355: Fixes for onnx==1.11 (2022-02-18)

  • issue 353: Experimentations with a new API to create ONNX graphs (2022-02-18)

  • issue 352: Supports for shape inference on unary operators (2022-02-14)

0.8.1697 - 2022-02-11 - 1.98Mb#

  • issue 351: Adds name in ShapeResult, fixes zoo links (2022-02-11)

  • issue 350: First version of runtime OnnxShapeInference (2022-02-09)

  • issue 348: Moves OnnxMicroRuntime to onnxrt (2022-02-05)

  • issue 346: Adds runtime for operator CastLike (2022-02-05)

  • issue 347: numpy API for onnx: wrapped function can call other wrapped functions (2022-02-04)

  • issue 345: Improves command line to measure latency for a model (2022-02-03)

  • issue 344: Adds a method to_onnx to easily retrieve the onnx graph from numpy onnx function (2022-02-03)

  • issue 343: Shows links in onnx_simple_text_plot (2022-02-03)

  • issue 342: Displays small arrays in onnx_simple_text_plot (2022-01-22)

0.8.1674 - 2021-12-30 - 23.58Mb#

  • issue 340: Implements tokenizer following scikit-learn’s API using onnxruntime-extensions (2021-12-29)

  • issue 339: op_label_encoder support for keys_strings & values_floats (2) (replaces #335) (2021-12-29)

  • issue 338: Updated to support key_strings and values_floats combo (2021-12-29)

  • issue 335: op_label_encoder support for keys_strings & values_floats (2021-12-29)

  • issue 322: Add tokenizers with onnxruntime-extensions (2021-12-29)

  • issue 337: Supports operator Scan when exporting an onnx graph to onnx code (2021-12-21)

  • issue 336: Enables GPU with OnnxInference and onnxruntime (2021-12-21)

0.7.1672 - 2021-12-19 - 1.95Mb#

  • issue 334: update history (2021-12-19)

  • issue 333: Adds command line latency to measure the latency of a runtime (2021-12-18)

  • issue 332: Improves dot rendering, fixes disconnected subgraphs (2021-12-18)

  • issue 331: Removes measure_time (2021-12-15)

  • issue 330: Reduces verbosity when onnxruntime is used as a runtime for OnnxInference (2021-12-14)

  • issue 329: Fixes type issue in shape inference for operator If (2021-12-14)

  • issue 328: Extends command line onnx_stats (2021-12-14)

  • issue 327: Adds runtime for operator LeakyRelu (2021-12-13)

  • issue 326: Better error messages when name is shared with results and node name in onnx_simple_text_plot (2021-12-10)

0.7.1649 - 2021-12-09 - 1.94Mb#

  • issue 325: Implements a simple text display for ONNX graph (2021-12-08)

  • issue 324: Adds runtime for gradient operators YieldOp, BroadcastGradientArgs (2021-11-30)

  • issue 323: Implements if with numpy API (2021-11-26)

  • issue 320: Fix exporter to tf2onnx (2021-11-13)

  • issue 319: Supports operator SequenceAt in OnnxInference (2021-11-09)

  • issue 318: Disable onnxruntime optimisation on one particular graph (2021-11-04)

  • issue 317: plot_onnx fails when node names contains ‘.’ (2021-10-28)

  • issue 316: failed to use RandomForestRegressor ort in android studio (2021-10-28)

0.7.1626 - 2021-10-21 - 23.49Mb#

  • issue 315: Fixes import issue for python 3.6 (2021-10-21)

0.7.1625 - 2021-10-12 - 0.58Mb#

  • issue 314: Builds mlprodict for python 3.6 on linux (2021-10-11)

  • issue 313: Fix a bug related to shapes when exporting a model to tf2onnx (2021-10-10)

  • issue 312: Add more tests for einsum decomposition (2021-10-08)

0.7.1624 - 2021-10-02 - 2.69Mb#

  • issue 311: Support opset 15 (onnx>=1.10) (2021-10-02)

  • issue 310: Raise an exception when inplace and intermediate are True (OnnxInference.run) (2021-09-23)

0.7.1602 - 2021-09-21 - 2.69Mb#

  • issue 309: Adds function insert_results_into_onnx to insert results into a graph to debug (2021-09-21)

  • issue 308: Adds function to rename all results in ONNX graphs (2021-09-13)

  • issue 307: Adds runtime for operator SequenceConstruct (2021-09-13)

  • issue 305: Add option to split lightgbm converter into multipule TreeEnsemble (2021-09-10)

  • issue 304: Add tree text visualization for TreeEnsemble (2021-09-01)

  • issue 303: Implements a estimator speeding up the inference using ONNX (2021-08-31)

  • issue 302: Removes unused nodes after changing the outputs. (2021-08-23)

  • issue 298: Remove unused nodes after changing the outputs (2021-08-23)

  • issue 301: Different build for manylinux on python 3.9 (2021-08-18)

  • issue 300: Improves Lightgbm converter design + fix wrong prediction for TreeEnsemble with non contiguous arrays (2021-08-18)

  • issue 297: Adds function to convert ONNX into numpy code. (2021-08-13)

  • issue 296: Lightgbm + add function matmul to numpy API for ONNX (2021-08-07)

  • issue 295: Implements runtime for operator FFT (2021-08-03)

  • issue 291: Fixes infinite loop with operator loop, add support for static variables in Loop (2021-07-31)

  • issue 294: Implements text representation of an ONNX graph (bigraph) (2021-07-30)

  • issue 293: Add a tool to display an ONNX graph into text format (2021-07-30)

  • issue 292: Adds operator AveragePool to the python runtime (2021-07-29)

  • issue 290: Increases code coverage, add infer_size for Loop runtime (2021-07-28)

0.6.1522 - 2021-07-26 - 23.15Mb#

  • issue 289: Avoids raising an exception when an optional parameter is not specified (2021-07-26)

  • issue 288: Extends code coverage (2021-07-25)

  • issue 287: Adds python runtime for operator Loop, SequenceInsert, ConcatFromSequence (2021-07-25)

  • issue 286: Adds runtime for operator Range (2021-07-13)

0.6.1447 - 2021-07-12 - 1.79Mb#

  • issue 285: Adds function cst to create constant with numpy API for ONNX (2021-07-12)

  • issue 283: Commutative property (2021-07-12)

  • issue 281: Infers temporary allocation needed while computing the outputs (2021-07-12)

  • issue 284: Adds function transpose to numpy API for ONNX (2021-07-10)

  • issue 282: Upgrade requirements to skl2onnx>=1.9.0 (2021-07-02)

  • issue 280: More robustness for the python runtime (2021-07-01)

  • issue 279: Implements method infer_types in OnnxInference (2021-06-28)

  • issue 278: Adds operators ReduceSum, Max to OnnxMicroRuntime (2021-06-27)

  • issue 277: Switch to python 3.9 in CI (2021-06-25)

  • issue 276: Use openmp to parallelize QLinearConv (2021-06-25)

  • issue 275: Adds new strategy to pick up the best einsum equation based on ML (2021-06-25)

  • issue 274: Fixes issue raised with scipy 1.7.0 (2021-06-22)

  • issue 273: Adds operator where, improves numpy api (x[x<0]= 2) (2021-06-18)

  • issue 272: Explore custom implementation of operator add (2021-06-18)

  • issue 271: Updates default opset from 13 to 14 (2021-06-17)

  • issue 270: Adds more tests for QLinearConv runtime (2021-06-16)

  • issue 269: Adds runtime for operator QLinearConv (2021-06-04)

  • issue 268: Adds function to prepare data for onnxruntime_perf_test (2021-05-17)

  • issue 267: Moves onnxruntime code inside a wrapper to reduce logs (2021-05-14)

  • issue 266: Optimizes einsum even if not decomposed (2021-05-13)

  • issue 265: Refactoring, moves files to onnx_tools (2021-05-12)

  • issue 264: Support SessionOptions for runtime onnxruntime2 (2021-05-12)

  • issue 263: Refactor einsum files (2021-05-06)

  • issue 262: Refactoring, moving files into onnx_tools (2021-05-06)

  • issue 261: Improves einsum decomposition by using gemm and removing a transpose (2021-05-05)

  • issue 260: New command line to benchmark einsum decomposition (2021-05-03)

  • issue 259: Minor changes to Einsum decomposition (2021-05-02)

  • issue 258: Decomposes Einsum into simple matrix operations (2021-04-30)

  • issue 257: Fixes #256, add method to validate input data in numpy API for ONNX (2021-04-20)

  • issue 256: Add virtual method to validate input before predictions in numpy API for ONNX (2021-04-20)

0.5.1447 - 2021-04-17 - 1.54Mb#

  • issue 255: Supports any embedded estimator with numpy API (2021-04-17)

  • issue 254: Adds python runtime for operator ReduceL1 (2021-04-16)

  • issue 253: Adds runtime for operator ReduceL2 (2021-04-14)

  • issue 252: Implements an experimental version of reducesum for the case RK (2021-04-07)

  • issue 251: Increases code coverage (2021-04-07)

  • issue 250: Increases code coverage of unit tests (2021-04-03)

  • issue 248: Adds implementation of BatchNormalization opset 14 (2021-03-29)

  • issue 247: Introduces FctVersion to fix issue with optional arguments (2021-03-29)

  • issue 246: Extends example on ReduceSum benchmark (2021-03-26)

  • issue 244: Supports embedded models, complete tutorial on numpy API for ONNX (2021-03-26)

  • issue 243: Add decorator to wrap converter for clustering (numpy API) (2021-03-17)

  • issue 242: Add decorator to wrap converter for classifier (numpy API) (2021-03-17)

  • issue 241: Add decorator to register scikit-learn classes with numpy API for ONNX (2021-03-14)

  • issue 240: Add decorator to wrap converter for regressor (numpy API) (2021-03-14)

  • issue 239: Add runtime empty (2021-03-13)

  • issue 238: Use numpy API for ONNX to write custom converters (2021-03-13)

  • issue 237: Add a unit test to check an exception (2021-03-10)

  • issue 236: Implements __setitem__ for one dimension array (2021-03-08)

  • issue 235: Supports profiling for runtime onnxruntime1 (2021-03-04)

  • issue 233: Extend documentation about numpy API for ONNX (2021-03-04)

  • issue 234: Add parameter overwrite to select_model_inputs_outputs (2021-03-03)

  • issue 232: Implements pickling for functions used in numpy API for ONNX (2021-03-03)

  • issue 231: Supports different inputs in select_model_inputs_outputs (2021-03-03)

  • issue 230: Add unsqueeze, squeeze, expand_dims to numpy API for ONNX (2021-03-02)

  • issue 229: Add method flatten, function pad to numpy API for ONNX (2021-03-01)

  • issue 228: Improves numpy API for ONNX: type constraints (2021-03-01)

  • issue 227: Add functions arange, cumsum, compress to numpy API for ONNX (2021-03-01)

  • issue 226: Add function Einsum to numpy API for ONNX (2021-02-28)

  • issue 225: Adds function Clip to numpy API for ONNX (2021-02-28)

  • issue 224: Adds functions ceil, round to numpy API for onnx (2021-02-27)

  • issue 223: Test numpy API against onnxruntime (2021-02-27)

  • issue 222: Add hyperbolic function, prod, mean, argmin, argmax (2021-02-26)

  • issue 221: Add many simple functions to numpy API for ONNX (2021-02-26)

  • issue 220: Tutorial on numpy API for ONNX (2021-02-26)

  • issue 219: Simplifies onnxfication of FunctionTransformer (2021-02-23)

  • issue 218: Implements __setitem__ for class OnnxVar (2021-02-21)

  • issue 217: Move custom operator to a specific method easier to maintain (2021-02-21)

  • issue 216: Fix crash with Gather, TopK when k=0 or indices is empty. (2021-02-20)

  • issue 215: Implements __getitem__ for OnnxVar (onnxnumpy) (2021-02-20)

  • issue 214: Implements numpy functions with onnx (2021-02-19)

  • issue 213: Add parameter show to plot_onnx. (2021-02-11)

  • issue 212: Fixes #210, check first models from zoo, fix operator conv when B is not null (2021-02-05)

  • issue 210: Investigate models from ONNX zoo (2021-02-05)

  • issue 211: numpy 1.20 does not allow nan values in int64 arrays any more, fix a unit test about imputer (2021-02-02)

  • issue 208: Add try catch around import in asv benchmark (2021-01-30)

  • issue 207: Reduces greater batch size to 10.000 instead of 100.000. (2021-01-29)

  • issue 205: Fixes asv configuration (2021-01-18)

  • issue 206: Build wheel for all many platforms in CI (2021-01-17)

0.5.1360 - 2021-01-04 - 1.44Mb#

  • issue 203: Enable Python 3.9, enable opset 13, upgrade version number (2021-01-04)

  • issue 202: Enable opset 13 (ONNX) (2021-01-04)

  • issue 201: Fixes #200, add support for float16 (2020-12-30)

  • issue 200: Add support for bfloat16 (2020-12-30)

  • issue 199: Fix unit tests recently failing due to onnxruntime update. (2020-12-15)

0.4.1352 - 2020-12-11 - 0.34Mb#

  • issue 196: Fixes operator Slice for opset 9 (2020-12-11)

  • issue 198: Fixes #197, add function to plot onnx graph with matplotlib (2020-12-09)

  • issue 197: Add a function to plot an onnx graph into matplotlib (2020-12-09)

  • issue 195: Fixes #194, add function to add an operator in the graph (2020-12-08)

  • issue 194: Add a function to insert a cast operator between two nodes (2020-12-08)

  • issue 193: Improves notebook coverage, update CI (2020-11-29)

  • issue 192: Fixes #191, improves performance of TreeEnsemble (2020-11-28)

  • issue 191: Improves performance of TreeEnsemble (2020-11-28)

  • issue 190: Fixes #189, parallelization of Einsum (2020-11-17)

  • issue 189: Introduce parallelization in experimental einsum implementation (2020-11-17)

  • issue 188: Fixes #187, custom implementation for operator Einsum (2020-11-15)

  • issue 187: Custom implementation for operator Einsum (2020-11-15)

  • issue 186: Fixes #185, add operator LessOrEqual (2020-11-15)

  • issue 185: Add operator LessOrEqual (2020-11-15)

  • issue 181: Fix converter xgboost when ntree_limit is set up (2020-11-14)

  • issue 184: Fixes #183, fix missing parameter black_op in OnnxPipeline (2020-11-07)

  • issue 183: Fix error in OnnxPipeline, parameter black_op not found (2020-11-07)

  • issue 182: Fixes #178, fix xgboost issue with ntree_limit (2020-11-07)

  • issue 178: Fixes unit test testing OnnxConv (issue with shapes) (2020-11-07)

  • issue 180: Fixes #179, fix guess_schema_from_data for categories (2020-11-03)

  • issue 179: guess_schema_data_type fails with category in dataframe (2020-11-03)

  • issue 176: Fixes #175, add operator dropout (2020-09-29)

  • issue 175: Add operator Dropout (2020-09-29)

  • issue 174: Add support for ReduceSum >= 13 (2020-09-21)

  • issue 173: Fixes #172, add runtime for operator MaxPool (2020-09-16)

  • issue 172: Add runtime for operator MaxPool (2020-09-16)

  • issue 171: Fixes #170, add operator Pad (2020-09-10)

  • issue 170: Add runtime for operator Pad (2020-09-10)

  • issue 169: fix compiling issue with ubuntu 16.04 (2020-09-03)

  • issue 167: Add runtime for Operator Or (2020-08-25)

  • issue 166: Add runtime for operator And (2020-08-25)

  • issue 165: Add runtime for operator GreaterOrEqual (2020-08-25)

  • issue 164: Add runtime for operator If (2020-08-25)

  • issue 163: Add runtime for operator Unsqueeze (2020-08-25)

  • issue 162: Add runtime for operator Split (2020-08-25)

  • issue 161: Add support for disable_optimisation (2020-08-12)

  • issue 160: Fixes #159, add operator ConvTranspose, refactoring. (2020-08-07)

  • issue 159: Implements runtime for ConvTranspose (2020-08-07)

  • issue 158: Fixes benchmark import issues (2020-08-03)

  • issue 157: Simplify scenarios, reduce time for benchmark. (2020-08-02)

  • issue 156: Fixes #155, improves documentation (2020-08-02)

  • issue 155: Fixes API on documentation (2020-08-02)

  • issue 154: Fixes y_train dtype for most of the problems. Fixes subproblems with GridSearchCV (2020-07-31)

  • issue 153: Fixes #152, set set n_jobs to the number of CPU (2020-07-31)

  • issue 152: Set n_jobs to the number of core - 1 when doing benchmark (2020-07-31)

  • issue 151: Force operator Conv to use continuous array (2020-07-30)

  • issue 150: Fixes nan issue in operator conv (2020-07-29)

  • issue 147: Fixes #145, #150, shape inference for operator Conv (2020-07-29)

  • issue 145: Fixes missing shape inference for operator conv (2020-07-29)

  • issue 149: Fixes #148, add operator Atan (2020-07-22)

  • issue 148: Add operator atan (2020-07-22)

  • issue 146: Fixes #144, add operator GlobalAveragePool (2020-07-21)

  • issue 144: Implements operator GlobalAveragePool (2020-07-21)

  • issue 143: Fixes #142, add operator BatchNormalization (2020-07-21)

  • issue 142: Implement python runtime for operator BatchNormalization (2020-07-21)

  • issue 141: Fixes #140, add runtime for QuantizeLinear, DequantizeLinear (2020-07-20)

  • issue 140: Implement runtime for QuantizeLinear, DequantizeLinear (2020-07-20)

  • issue 139: Add runtime for operator EyeLike (2020-07-08)

  • issue 138: Add code to register custom python operator (2020-07-08)

  • issue 137: Remove parameter dtype (onnx conversion) (2020-07-08)

  • issue 136: Add parameter reshape to OnnxTransformer (2020-07-03)

  • issue 135: Add a function to change the first dimension output (ONNX). (2020-07-03)

  • issue 133: Implements runtime for operator Gather (ONNX) (2020-06-18)

  • issue 132: Add operator StringNormalizer, Tokenizer, TfidfVectorizer (ONNX) (2020-06-15)

  • issue 131: Add custom operator solve (2020-06-12)

  • issue 130: Add operator Erf (ONNX) (2020-06-11)

  • issue 129: Add operator Einsum (ONNX) (2020-06-11)

  • issue 128: Fixes #127, implements OnnxPipeline, train, convert at each step (2020-06-08)

  • issue 127: Implements a pipeline which replaces early stages by onnx (2020-06-08)

  • issue 123: Enables opset 12 (ONNX) (2020-06-04)

  • issue 117: Support for op_version in onnx grammar (2020-06-04)

  • issue 126: Fix xgboost converter for xgboost >= 1.0 (2020-05-18)

  • issue 125: Refactor rewritten sklearn operators (2020-05-18)

  • issue 124: Fixes #122, capture standard C ouptput with dump_data_model, first step for #123 (2020-05-16)

  • issue 122: Captures C output when calling dump_data_and_model (2020-05-16)

  • issue 121: Add function to convert array to bytes and bytes to array (onnx tensor) (2020-04-30)

  • issue 120: Fix discrepencies for SVM classifier (ONNX) (2020-04-30)

  • issue 119: Keep order in topk implementation (2020-04-17)

  • issue 118: opset is not propagated in OnnxTransformer (2020-04-09)

  • issue 115: Add a function to replay a benchmark when this one was dumped (more accurate) (2020-04-06)

  • issue 116: Makes ZipMapDictionary picklable (2020-03-30)

  • issue 114: Add more parameters to specify benchmark time (2020-03-30)

  • issue 113: Add operators for opset 12 (2020-03-26)

  • issue 112: Number of feature is wrong for problem num-tr-clus (2020-03-20)

  • issue 111: Reduce the number of allocation in TreeEnsemble when it is parallelized (cache) (2020-03-13)

  • issue 110: Implements runtime for operator Constant-12 (2020-03-06)

  • issue 109: Generate a benchmark with asv to compare different runtime. Update modules in asv. (2020-03-06)

  • issue 108: Add a function to reduce the memory footprint (2020-02-25)

  • issue 106: Add operator Neg (2020-02-25)

  • issue 101: Fix DecisionTreeClassifier disappearance on the benchmark graph (2020-02-25)

  • issue 107: Add operator IsNaN (2020-02-24)

  • issue 105: Support string labels for Linear, TreeEnsemble, SVM classifiers. (2020-02-24)

  • issue 104: Enable / disable parallelisation in topk (2020-02-23)

  • issue 103: Implements plot benchmark ratio depending on two parameters (2020-02-22)

  • issue 102: Fix conversion for xgboost 1.0 (2020-02-21)

  • issue 100: add notebook on TreeEnsemble (2020-02-19)

  • issue 99: Fixes #93, use same code for TreeEnsembleClassifier and TreeEnsembleRegression (2020-02-19)

  • issue 93: Use pointer for TreeClassifier (2020-02-19)

  • issue 98: mlprodict i broken after onnxruntime, skl2onnx update (2020-02-15)

  • issue 97: Add runtime for operator Conv (2020-01-24)

  • issue 96: Fixes #97, add runtime for operator Conv (2020-01-24)

  • issue 95: Fix OnnxInference where an output and an operator share the same name (2020-01-15)

  • issue 94: Raw scores are always positive for TreeEnsembleClassifier (binary) (2020-01-13)

  • issue 90: Implements a C++ runtime for topk (2019-12-17)

  • issue 86: Use pointers to replace treeindex in tree ensemble cpp runtime (2019-12-17)

  • issue 92: Implements a C++ version of ArrayFeatureExtractor (2019-12-14)

  • issue 89: Implements a function which extracts some informations on the models (2019-12-14)

  • issue 88: Fix bug in runtime of GatherElements (2019-12-14)

  • issue 87: Add converter for HistGradientBoostRegressor (2019-12-09)

  • issue 85: Implements a precompiled run method in OnnxInference (runtime=’python_compiled’) (2019-12-07)

  • issue 84: Automatically creates files to profile time_predict function in the benchmark with py-spy (2019-12-04)

  • issue 83: ONNX: includes experimental operators in the benchmark (2019-12-04)

  • issue 82: Function translate_fct2onnx: use of opset_version (2019-12-04)

  • issue 81: ONNX benchmark: track_score returns scores equal to 0 or 1 (unexpected) (2019-12-04)

  • issue 80: ONNX: extend benchmark to decision_function for some models (2019-12-03)

  • issue 77: Improves ONNX benchmark to measure zipmap impact. (2019-12-03)

  • issue 76: Implements ArgMax 12, ArgMax 12 (python onnx runtime) (2019-11-27)

  • issue 75: ONNX: fix random_state whevever it is available when running benchmark (2019-11-27)

  • issue 59: ONNX: Investigate kmeans and opset availability. (2019-11-21)

  • issue 66: ONNX: improves speed of python runtime for decision trees (2019-11-19)

  • issue 74: Function _modify_dimension should return the same dataset if called the same parameter (even if it uses random functions) (2019-11-15)

  • issue 73: ONNX: fix links on benchmark page (opset is missing) (2019-11-07)

  • issue 72: ONNX: support of sparse tensor for a unary and binary python operators (2019-11-06)

  • issue 71: ONNX: add operator Constant (2019-11-06)

  • issue 67: ONNX: improves speed of svm regressor (2019-11-06)

  • issue 70: ONNX: write tools to test convervsion for models in scikit-learn examples (2019-10-29)

  • issue 65: ONNX: investigate discrepencies for k-NN (2019-10-28)

  • issue 69: ONNX: side by side should work by name and not by positions (2019-10-23)

  • issue 68: ONNX: improves speed of SGDClassifier (2019-10-23)

  • issue 61: Implements a function to create a benchmark based on asv (ONNX) (2019-10-17)

  • issue 63: Export asv results to csv (ONNX) + command line (2019-10-11)

  • issue 64: Add an example with lightgbm and categorical variables (ONNX) (2019-10-07)

  • issue 62: Implements command line for the asv benchmark (ONNX) (2019-10-04)

  • issue 60: Improve lightgbm converter (ONNX) (2019-09-30)

  • issue 58: Fix table checking model, merge is wrong in documentation (2019-09-20)

  • issue 57: ONNX: handles dataframe when converting a model (2019-09-15)

  • issue 56: ONNX: implements cdist operator (2019-09-12)

  • issue 54: ONNX: fix summary, it produces multiple row when model are different when opset is different (2019-09-12)

  • issue 51: ONNX: measure the time performance obtained by using optimization (2019-09-11)

  • issue 52: ONNC-cli: add a command line to optimize an onnx model (2019-09-10)

  • issue 49: ONNX optimization: remove redundant subparts of a graph (2019-09-09)

  • issue 48: ONNX optimization: reduce the number of Identity nodes (2019-09-09)

  • issue 47: Implements statistics on onnx graph and sklearn models, add them to the documentation (2019-09-06)

  • issue 46: Implements KNearestNeibhorsRegressor supporting batch mode (ONNX) (2019-08-31)

  • issue 45: KNearestNeighborsRegressor (2019-08-30)

  • issue 44: Add an example to look into the performance of every node for a particular dataset (2019-08-30)

  • issue 43: LGBMClassifier has wrong shape (2019-08-29)

  • issue 42: Adds a graph which visually summarize the validating benchmark (ONNX). (2019-08-27)

  • issue 41: Enables to test multiple number of features at the same time (ONNX) (2019-08-27)

  • issue 40: Add a parameter to change the number of featuress when validating a model (ONNX). (2019-08-26)

  • issue 39: Add a parameter to dump all models even if they don’t produce errors when being validated (ONNX) (2019-08-26)

  • issue 24: support double for TreeEnsembleClassifier (python runtime ONNX) (2019-08-23)

  • issue 38: See issue on onnxmltools. onnx/onnxmltools#321 (2019-08-19)

  • issue 35: Supports parameter time_kwargs in the command line (ONNX) (2019-08-09)

  • issue 34: Add intervals when measuring time ratios between scikit-learn and onnx (ONNX) (2019-08-09)

  • issue 31: Implements shape inference for the python runtime (ONNX) (2019-08-06)

  • issue 15: Tells operator if the execution can be done inplace for unary operators (ONNX). (2019-08-06)

  • issue 27: Bug fix (2019-08-02)

  • issue 23: support double for TreeEnsembleRegressor (python runtime ONNX) (2019-08-02)

  • issue 26: Tests all converters in separate processeses to make it easier to catch crashes (2019-08-01)

  • issue 25: Ensures operator clip returns an array of the same type (ONNX Python Runtime) (2019-07-30)

  • issue 22: Implements a function to shake an ONNX model and test float32 conversion (2019-07-28)

  • issue 21: Add customized converters (2019-07-28)

  • issue 20: Enables support for TreeEnsemble operators in python runtime (ONNX). (2019-07-28)

  • issue 19: Enables support for SVM operators in python runtime (ONNX). (2019-07-28)

  • issue 16: fix documentation, visual graph are not being rendered in notebooks (2019-07-23)

  • issue 18: implements python runtime for SVM (2019-07-20)

  • issue 17: add a mechanism to use ONNX with double computation (2019-07-15)

  • issue 13: add automated benchmark of every scikit-learn operator in the documentation (2019-07-05)

  • issue 12: implements a way to measure time for each node of the ONNX graph (2019-07-05)

  • issue 11: implements a better ZipMap node based on dedicated container (2019-07-05)

  • issue 8: implements runtime for decision tree (2019-07-05)

  • issue 7: implement python runtime for scaler, pca, knn, kmeans (2019-07-05)

  • issue 10: implements full runtime with onnxruntime not node by node (2019-06-16)

  • issue 9: implements a onnxruntime runtime (2019-06-16)

  • issue 6: first draft of a python runtime for onnx (2019-06-15)

  • issue 5: change style highlight-ipython3 (2018-01-05)