Concrete-ml

Feature
* Add a XGBoost regression tutorial ([1911](https://github.com/zama-ai/concrete-ml-internal/issues/1911)) ([`174f6f7`](https://github.com/zama-ai/concrete-ml-internal/commit/174f6f7b8cf2594eefe89c4570e0cfbfeb85d306))
* Add encrypted sentiment analysis demo ([`fd684df`](https://github.com/zama-ai/concrete-ml-internal/commit/fd684dfccc06a74fc45b872389ef93ff39709fe4))
* Make net_inputs and net_ouputs optional ([`5ed9476`](https://github.com/zama-ai/concrete-ml-internal/commit/5ed9476ba19e04df90631a7503ce8a938627f594))
* Add RandomForestRegressor ([`0c7853c`](https://github.com/zama-ai/concrete-ml-internal/commit/0c7853c4f6213a179407583b876bec5e463d7178))
* Add XGBRegressor ([`9736557`](https://github.com/zama-ai/concrete-ml-internal/commit/97365570457757781ba51f3a9992812013dd947e))
* Add tree-regressor ([`5b12d53`](https://github.com/zama-ai/concrete-ml-internal/commit/5b12d53df5d0ec2ed153f37556ab031c1f2dfc33))
* Add Ridge, Lasso and ElasticNet regression models. ([`675c7b3`](https://github.com/zama-ai/concrete-ml-internal/commit/675c7b31b956d50f7555a6be1f78215491c77236))
* Import Quantized Brevitas ONNX graphs and upgrade QAT notebook ([`13d8d74`](https://github.com/zama-ai/concrete-ml-internal/commit/13d8d7471fac0a0c34653a5cac0fd57f34bcf770))

Fix
* Remove pygraphviz dependancy ([`f708ba7`](https://github.com/zama-ai/concrete-ml-internal/commit/f708ba7a6fe2a31d3302e88c571336b91696adc9))
* Flaky client server ([1927](https://github.com/zama-ai/concrete-ml-internal/issues/1927)) ([`d162cd6`](https://github.com/zama-ai/concrete-ml-internal/commit/d162cd6e0fdeb28ed4ae0672a7d021307cb66b92))
* Tweedie overflow underterministic bug ([`922d60e`](https://github.com/zama-ai/concrete-ml-internal/commit/922d60eb530fad648b49aa7c67ed553bf43286da))
* XGBRegressor verifies that n_targets is 1 ([`a15fe9d`](https://github.com/zama-ai/concrete-ml-internal/commit/a15fe9d24a965308b6fafdf2e8da965b1ea2a9c6))
* Make linear models with fit_intercept=False possible ([`90a50b4`](https://github.com/zama-ai/concrete-ml-internal/commit/90a50b404fde979f752620ea4fc70ce71b91c85d))
* Add quantize_inputs_with_net_outputs_precision to calibration process ([`1dd9ba3`](https://github.com/zama-ai/concrete-ml-internal/commit/1dd9ba32720a138af9ab6519ddeee14b9c5d1bec))

Documentation
* Update p_error with api call ([`d214216`](https://github.com/zama-ai/concrete-ml-internal/commit/d21421687f619aacbd0787fea606485f12e0fbea))
* Improve ClassifierComparison notebook ([`f45b79f`](https://github.com/zama-ai/concrete-ml-internal/commit/f45b79f3bd72315c0eb084f2a1a51ae57cff37f2))
* Integration of API docs with lazydocs ([`985d0d5`](https://github.com/zama-ai/concrete-ml-internal/commit/985d0d5a57d8c7599cc62132862a2a82eeab5f56))
* Major Revision of Inner Workings and integration of Quantization Aware Training ([`accfd3e`](https://github.com/zama-ai/concrete-ml-internal/commit/accfd3eea736e321e97c487153f3a21fcfb7c8b3))
* Improve contribution doc ([`0a0534a`](https://github.com/zama-ai/concrete-ml-internal/commit/0a0534ae29fa185c25d6935d72fac6b97f05c6b3))
* Adding new models to the docs ([`081d8f9`](https://github.com/zama-ai/concrete-ml-internal/commit/081d8f9a8d75bb2b9b155287966afe8aef923d46))
* Be more precise on installation ([`88fed0b`](https://github.com/zama-ai/concrete-ml-internal/commit/88fed0b517f02c12aeb79639c1fd1a847db70c2c))
* Improve our README ([`b2d81e4`](https://github.com/zama-ai/concrete-ml-internal/commit/b2d81e44e58fcf94f0992a908553e24b84b433f9))
* Decrease net_ouputs values from 8 to 5 bits in notebooks ([`49717bf`](https://github.com/zama-ai/concrete-ml-internal/commit/49717bf4084651a42ba752fc42b93c77f7773d13))
* Add comment about tqdm in linear.md ([`ba06e5e`](https://github.com/zama-ai/concrete-ml-internal/commit/ba06e5ecfc033fd5e270a08fc2760be225b29698))
* Update the LICENSE ([`30b2b27`](https://github.com/zama-ai/concrete-ml-internal/commit/30b2b27762cbcf88d35435b47b461729b43d0f5f))
* Add tqdm and remove inference slicing in titanic notebook ([`4419438`](https://github.com/zama-ai/concrete-ml-internal/commit/4419438834cb79db343cb21ce71de978f4febc2f))

Feature
* Allow recompiling from onnx model ([`9b69e73`](https://github.com/zama-ai/concrete-ml/commit/9b69e73667efa233c4649b6cf4379b15a8d7d6e3))
* Adding support for p_error ([`fe03441`](https://github.com/zama-ai/concrete-ml/commit/fe03441331b0b2eac558fe6feb19c4c8450771d9))
* Adding GELU activation ([`c732d15`](https://github.com/zama-ai/concrete-ml/commit/c732d150fb30ad14ae787b2586a22d4c91269e01))
* Add random_state to models for client server reproducibility ([`6f887d0`](https://github.com/zama-ai/concrete-ml/commit/6f887d0ac590821c648e6aa6a504cff7fbc52790))
* Add QAT notebook ([`64c4512`](https://github.com/zama-ai/concrete-ml/commit/64c4512c47181f6df94cbc378605d44221ba82ce))
* Import Brevitas QAT networks ([`6c40a0c`](https://github.com/zama-ai/concrete-ml/commit/6c40a0c7c3765012766592a74db99430c5a373e5))
* Integration of the encrypt decrypt api ([`3c2a68a`](https://github.com/zama-ai/concrete-ml/commit/3c2a68a24da3e129c1650676f96642cceb0e0256))
* Support more input types in predict() ([`76e142c`](https://github.com/zama-ai/concrete-ml/commit/76e142ccd565793a6929cafd3ce82b07665541f0))
* Support more input types in fit() ([`1fa74f7`](https://github.com/zama-ai/concrete-ml/commit/1fa74f7bbf5a1ffd2ac3a3927e78ff7e0cf4d5e9))
* Add Round and Pow operators ([`d6880ce`](https://github.com/zama-ai/concrete-ml/commit/d6880ce9c8cd037504624d182c59d573483ef071))
* Ability to import Quantization Aware Training networks ([`c1bb947`](https://github.com/zama-ai/concrete-ml/commit/c1bb9479c7d4f55c33eb1a0c7a488c57515b2b8e))
* Compile user supplied ONNX to support keras/tf ([`fae3dc5`](https://github.com/zama-ai/concrete-ml/commit/fae3dc5b226f8b3ae98a54807cb661767e69beb3))
* Implement Generalized Linear Regression models ([`8e8e025`](https://github.com/zama-ai/concrete-ml/commit/8e8e025771f8a6abff95c752697d9d40aae9d2ab))
* Add SoftSign activation ([`3ce338e`](https://github.com/zama-ai/concrete-ml/commit/3ce338e9a29be1ce8a69763b24dda90a5828f812))
* Adding more activations ([`e43ce5c`](https://github.com/zama-ai/concrete-ml/commit/e43ce5c2481e39f7cb45529a912bb652d81c54a1))
* Implement Poisson Regression ([`09eefa5`](https://github.com/zama-ai/concrete-ml/commit/09eefa5faaad8d24ba31a439b270d4a0b54f21ea))
* Use the 8b of precision of Concrete Numpy ([`249c712`](https://github.com/zama-ai/concrete-ml/commit/249c712a0757e3971d2bb1a7d375f718abdf5c6b))
* Add ONNX flatten support ([`c5f215f`](https://github.com/zama-ai/concrete-ml/commit/c5f215f5ec7132662c99c59b8819d8564f77005f))
* Handle more tree-based classifiers ([`950cc6c`](https://github.com/zama-ai/concrete-ml/commit/950cc6c7b0cf4563f6745a7966f2d709f51fb78b))
* Add Batch Normalization ONNX operator ([`7969739`](https://github.com/zama-ai/concrete-ml/commit/796973954b69cbeb7b362d9f1217c7c2b3468330))
* Add Where, Greater, Mul, Sub ONNX operator support ([`f939149`](https://github.com/zama-ai/concrete-ml/commit/f9391495274a5e0f24b7a46c830c429fb6f80e55))
* Add ONNX Average Pooling and Pad operator ([`40f1ef9`](https://github.com/zama-ai/concrete-ml/commit/40f1ef9455931be72397912fd69fa759800f8468))
* Add more activation functions ([`26b2221`](https://github.com/zama-ai/concrete-ml/commit/26b222181a8f0ed0bd6aec5f95eaf6eb94cb5413))

Fix
* Make tree inference faster by creating new numpy boolean operators ([`206caa5`](https://github.com/zama-ai/concrete-ml/commit/206caa5eec8f2e5db2530cf78069af060682e64e))
* Set a compatible version for protobuf ([`97ccfc0`](https://github.com/zama-ai/concrete-ml/commit/97ccfc053ad639842314bd4ba58ef22f6c6e985e))
* Improve IRIS FCNN FHE accuracy and visualization ([`02e497c`](https://github.com/zama-ai/concrete-ml/commit/02e497c50ad1a9634cbd9860baf1b0c9aa1fcf7b))
* Replace __init__ call by set_params ([`111419e`](https://github.com/zama-ai/concrete-ml/commit/111419e568761f59c1414fa188ecff01176f7350))
* Fix wrong fit_benchmark in linear models ([`f257def`](https://github.com/zama-ai/concrete-ml/commit/f257defebd67fa6b7e8342b9fcf46d52c9e3ee47))
* Fix GridSearchCV on trees ([`b614285`](https://github.com/zama-ai/concrete-ml/commit/b61428596aab98c1f7f1fa31aedf69ff46394bb0))
* Support decision tree with custom classes ([`baa3b4d`](https://github.com/zama-ai/concrete-ml/commit/baa3b4d46773d8918c09a28ed41e4f69ef8c2dc3))

Documentation
* Major refresh of 0.3 doc ([`e5e3205`](https://github.com/zama-ai/concrete-ml/commit/e5e320518a5841cc7880ef27cdb3bdc4c6097404))
* Add sentiment classification notebook ([`68ae7d0`](https://github.com/zama-ai/concrete-ml/commit/68ae7d0f946a743dfff6539876d4127c12375365))
* Restrict hyperparameters in titanic notebook for faster inference ([`9b63c8a`](https://github.com/zama-ai/concrete-ml/commit/9b63c8a923f76f45003354b77e3ed3061f408eb4))
* QAT explanation ([`9430ba6`](https://github.com/zama-ai/concrete-ml/commit/9430ba627a5576457c57b3fb35384d31008b47d2))
* Document ONNX compilation ([`972d05e`](https://github.com/zama-ai/concrete-ml/commit/972d05e8e8322d899d37830c5311243ccd17f563))
* Explain quantized vs float ops and fusing ([`fb9b409`](https://github.com/zama-ai/concrete-ml/commit/fb9b4090606a5a37afeb4a0b4afc6f930a1cd6ae))
* Add doc for pandas support ([`34652ce`](https://github.com/zama-ai/concrete-ml/commit/34652ce4e9ae7c414c26a10372b336b4f7e9f3d4))
* Add notebook and docs client server api ([`99ff1e7`](https://github.com/zama-ai/concrete-ml/commit/99ff1e73edc9dfc1384588a85289b669dc8505ff))
* Developing custom models ([`1c6f571`](https://github.com/zama-ai/concrete-ml/commit/1c6f571ea27b276acd6fee689e75a4b84b1f367b))
* Explain built-in quantized Neural Networks ([`972d464`](https://github.com/zama-ai/concrete-ml/commit/972d4645cc763035a743dd7fdd7dfddc7f91dcfc))
* Add a notebook for Kaggle Titanic competition ([`0a44853`](https://github.com/zama-ai/concrete-ml/commit/0a448539db58307c6887b18e3045cc7e023fc853))

Fix
* Set a compatible version for protobuf ([`4dd46cf`](https://github.com/zama-ai/concrete-ml/commit/4dd46cf8b73dbc8f09370fb919d8177312ec8f9a))
* Force ONNX package version to 1.11.0 in CLM 0.2 ([`fa90586`](https://github.com/zama-ai/concrete-ml/commit/fa90586c1b3757be966cd658d09c4a23147a9eb7))

Documentation
* Update the LICENSE ([`38c1062`](https://github.com/zama-ai/concrete-ml/commit/38c106220cb10a3ab232fd22191b74ced5b49f6e))

Breaking Changes (as compared to 0.1.x)

* `run` method is renamed to `encrypt_run_decrypt` after changes in Concrete-Numpy 0.5.0. Individual APIs to encrypt/run/decrypt separately will be available in a further release of Concrete-ML

Feature
* Some breaking changes in Concrete Numpy API. ([`ecbb26e`](https://github.com/zama-ai/concrete-ml/commit/ecbb26ec89b2a9ba86d8cc5f7ade3c1b50837341))
* Using Concrete Numpy 0.5 ([`ee5987b`](https://github.com/zama-ai/concrete-ml/commit/ee5987bf03e0a90ef64b337cf530bbdcface8f60))
* Add multiclass xgboost ([`9247f35`](https://github.com/zama-ai/concrete-ml/commit/9247f35cbb50b98b23d4ce73e2cc69424cf23e4a))
* Adding a `set_version_and_push` command to the makefile ([`ab853c2`](https://github.com/zama-ai/concrete-ml/commit/ab853c2726dcecb93c085060ed81053008f99c64))
* Add multiclass capability to decision trees ([`df0a64d`](https://github.com/zama-ai/concrete-ml/commit/df0a64d2eb0c83244fd5a2b7a9f130368a4d509b))

Fix
* Fixing Pypi's Homepage button link ([`bf52514`](https://github.com/zama-ai/concrete-ml/commit/bf52514b876291462281d7f52039a646eddffbea))

Breaking
* The API `.forward_fhe.run()` has been renamed into `.forward_fhe.encrypt_run_decrypt()` ([`ecbb26e`](https://github.com/zama-ai/concrete-ml/commit/ecbb26ec89b2a9ba86d8cc5f7ade3c1b50837341))

Documentation
* Reorganize index page ([`f1bafff`](https://github.com/zama-ai/concrete-ml/commit/f1bafffcc9807dbaf24fe777466e652d23f03895))

Feature
* Add multiclass capability to decision trees ([`6f9651e`](https://github.com/zama-ai/concrete-ml/commit/6f9651e4052a7fd2c56532cb061fa122093d02a4))
* Update to Concrete Numpy 0.4.0 and update the theme. ([`27839be`](https://github.com/zama-ai/concrete-ml/commit/27839be8f8713d16c2efe5d764cf4372313be255))

Fix
* Fixing Pypi's Homepage button link ([`d1c0e5d`](https://github.com/zama-ai/concrete-ml/commit/d1c0e5d2b0f3edbde39c0cff62b8812d67f541cf))
* Broken links in README.md ([`0c47668`](https://github.com/zama-ai/concrete-ml/commit/0c476683a57274de5863678d9be100012c84e7dc))

Documentation
* Reorganize index page ([`349c947`](https://github.com/zama-ai/concrete-ml/commit/349c947ebda9fb00442d0860c9fa81990bcf2761))

Feature
* Add tests for more torch functions that are supported, mention them in the docs ([`0478854`](https://github.com/zama-ai/concrete-ml/commit/047885489a0e8fd0580c56729a9d143c7dfd7705))
* Add FHE in xgboost notebook ([`1367d4e`](https://github.com/zama-ai/concrete-ml/commit/1367d4ed1903ecbb20766fbd80c23f92946b4435))
* Make all classifier demos run in FHE for the datasets and in VL for the domain grid ([`d95af58`](https://github.com/zama-ai/concrete-ml/commit/d95af588fe308d30e53efd9352c23414ff1e1f51))
* Remove workaround reshape and remaining 3dmatmul ([`28ea1eb`](https://github.com/zama-ai/concrete-ml/commit/28ea1eb65f7b8d494a3cd5ef9ca5d69397f0a2f1))
* Change predict to predict_proba for average_precision ([`a057881`](https://github.com/zama-ai/concrete-ml/commit/a057881231f12ae1b67d29154b2dfe538aa716f5))
* Allow FHE on xgboost ([`7b5c118`](https://github.com/zama-ai/concrete-ml/commit/7b5c118192103700cce05fd5f7844c5ff9291ee1))
* Add CNN notebook ([`4acca2f`](https://github.com/zama-ai/concrete-ml/commit/4acca2f183d83bbafd5162db43227a9109aeea01))
* Optimize QuantizeAdd to use TLUs when one of the inputs is a constant([`1ffcdfb`](https://github.com/zama-ai/concrete-ml/commit/1ffcdfbb98367ded06b1be757c02453a05a7e048))
* Different n_bits for weights/activations/outputs ([`321d151`](https://github.com/zama-ai/concrete-ml/commit/321d151dac5edd78f5cae793ea8be5814ac55901))
* Add virtual lib management to SklearnLinearModelMixin ([`596d16e`](https://github.com/zama-ai/concrete-ml/commit/596d16e052592c6460aca746a34052c1015224f4))
* Add quantized CNN. ([`1a78593`](https://github.com/zama-ai/concrete-ml/commit/1a785937e5e6b85bec6eee00be3e9c3b305afdba))
* Start refactoring tree based models ([`8e62cf8`](https://github.com/zama-ai/concrete-ml/commit/8e62cf8b8dc2e3f6b28d05bc600479a2b29b1dba))
* Set symmetric quantization by default in PTQ ([`8fcd307`](https://github.com/zama-ai/concrete-ml/commit/8fcd307529103f8b5a7e576df5bb1f5116a64b34))
* Add random forest + benchmark ([`5630f17`](https://github.com/zama-ai/concrete-ml/commit/5630f17f7406277b1b917376a8ee1fadf1e0f133))
* Allow base_score with xgboost ([`17d5cc4`](https://github.com/zama-ai/concrete-ml/commit/17d5cc471b5e79924c198462280d1b51ba29b7f9))
* Add predict_proba to logistic regression ([`9aaeec5`](https://github.com/zama-ai/concrete-ml/commit/9aaeec56945e1d9f2941be31f01721ed47fcc464))
* Add xgboost ([`699603d`](https://github.com/zama-ai/concrete-ml/commit/699603d67d9d4b806ea833849f7e0611b8588781))
* Add NN regression benchmarks ([`9de2ba4`](https://github.com/zama-ai/concrete-ml/commit/9de2ba439aeb157b519cdc583310744f85822df8))
* Add symetric quantization (needed for tree output) ([`4a173ee`](https://github.com/zama-ai/concrete-ml/commit/4a173ee02c2340f2564791bb49fef64890f1e53f))
* Implement LinearSVC ([`d048077`](https://github.com/zama-ai/concrete-ml/commit/d0480773391cc1e0e7cab19e05517c24b9b29b39))
* Implement LinearSVRegression ([`36df77e`](https://github.com/zama-ai/concrete-ml/commit/36df77ee269bc9cacbb1608087a6629392822e00))
* Remove identity nodes from ONNX models ([`9719c08`](https://github.com/zama-ai/concrete-ml/commit/9719c087251cc41b0bee80eb3f939ccee1898236))
* Add binary + multiclass logistic regression ([`85c25df`](https://github.com/zama-ai/concrete-ml/commit/85c25dfec05cf15254e701a5acd0e9bcb08278f1))
* Improve r2 test for low variance targets. ([`44ec0b3`](https://github.com/zama-ai/concrete-ml/commit/44ec0b38c7d68e74c3308aa625dbe3267cace9a3))
* Add sklearn linear regression model ([`060a4c6`](https://github.com/zama-ai/concrete-ml/commit/060a4c6235d288d614d6da5e6dc0e04f08bdcb46))
* Add virtual lib basic class ([`ad32509`](https://github.com/zama-ai/concrete-ml/commit/ad3250900ab6e499f982c1fa2b363701baf2e2d4))
* Improve NN benchmarks ([`ae8313e`](https://github.com/zama-ai/concrete-ml/commit/ae8313e7cf8b76acf3b131e95ec08a313beb958e))
* Add NN benchmarks and sklearn wrapper for FHE NNs ([`e73a514`](https://github.com/zama-ai/concrete-ml/commit/e73a5144dbe04f4560c28906b48d74cf7ecd7f73))
* More efficient numpy_gemm, since traced ([`609f1df`](https://github.com/zama-ai/concrete-ml/commit/609f1df90577e215fae168898f70a6bacab51a6c))
* Integrate hummingbird ([`01c3a4a`](https://github.com/zama-ai/concrete-ml/commit/01c3a4a3d02e28ee0f2e72ab58e67ae7c7f322c4))
* Add ONNX quantized implementation for MatMul and Add ([`716fc43`](https://github.com/zama-ai/concrete-ml/commit/716fc43850ed945ab6f005ec24817172934ac161))
* Allow multiple inputs for a QuantizedModule ([`1fa530d`](https://github.com/zama-ai/concrete-ml/commit/1fa530db7e58abacc0a53b31a1d0f66b74949e5f))
* Allow QuantizedModule to handle complicated NN topologies ([`da91e40`](https://github.com/zama-ai/concrete-ml/commit/da91e403e4c72c7724a056b53436c81b833f74ad))
* Let's allow (alpha, beta) == (1, 0) in Gemm ([`4b9927a`](https://github.com/zama-ai/concrete-ml/commit/4b9927ae11463ab419a6af9005d9effec7b0f0fe))
* Manage constant folding in PTQ ([`a0c56d7`](https://github.com/zama-ai/concrete-ml/commit/a0c56d748b580cb4206754daaacdfdfc10200231))
* Replace numpy.isclose with r2 score ([`65f0a6e`](https://github.com/zama-ai/concrete-ml/commit/65f0a6e2c1d00d58a3f0babd4a3ff1128f9bf3df))
* Replace the torch quantization functions with ones usable with ONNX ([`ecdeb50`](https://github.com/zama-ai/concrete-ml/commit/ecdeb50ec9c519e5683cec1bfeeca4f30cee1777))
* Add test when input is float to quantized module ([`d58910d`](https://github.com/zama-ai/concrete-ml/commit/d58910d53c98e4c2e558c131ae9de3d2da47a2a2))
* Let user chose its error type ([`e5d7440`](https://github.com/zama-ai/concrete-ml/commit/e5d7440951d1eb01bc824b0c1117ec1329824bd6))
* Post training quantization for ONNX repr ([`8b051df`](https://github.com/zama-ai/concrete-ml/commit/8b051df6b2f067d0c03547e4ee8aac1f4acbfbab))
* Adding more activations and numpy functions ([`73d885c`](https://github.com/zama-ai/concrete-ml/commit/73d885cbffcf31adba86669548e707405df41bfe))
* Let's have relu and relu6 ([`f64c3bf`](https://github.com/zama-ai/concrete-ml/commit/f64c3bfa2bb55accc2b104095e30a0f3c116b6a0))
* Add quantized tanh ([`ca9c6e5`](https://github.com/zama-ai/concrete-ml/commit/ca9c6e54643c61c81ba43a2fcba09fb916ee9f8b))
* Add classification benchmarks, fix bugs in DecisionTreeClassifier ([`d66d7bf`](https://github.com/zama-ai/concrete-ml/commit/d66d7bfce15ef91cb94b92c73f65d53aa7bf1cec))
* Provide quantized versions of ONNX ops ([`b63eca2`](https://github.com/zama-ai/concrete-ml/commit/b63eca266204dd6972971d22e8e94fefbbe00c16))
* Add darglint as a pluggin of flake8 ([`bb568e2`](https://github.com/zama-ai/concrete-ml/commit/bb568e208bf9d2b509edc90e1353e6e94ba215fd))
* Use ONNX as intermediate format to convert torch models to numpy ([`072bd63`](https://github.com/zama-ai/concrete-ml/commit/072bd631271c61b528c8279c0f2aef7db84904b5))
* Add decision trees + update notebook ([`db163f5`](https://github.com/zama-ai/concrete-ml/commit/db163f593a517adea86a778b21e7505636c00f26))
* Restore quantized model benchmarks ([`d1cfc4e`](https://github.com/zama-ai/concrete-ml/commit/d1cfc4ef1dbb1b804e597c1c968e62689e2fa936))
* Port quantization and torch from concrete-numpy. ([`a525e8b`](https://github.com/zama-ai/concrete-ml/commit/a525e8bfcf974aaff1b31071a9027b41f4cc3581))

Fix
* Remove fixmes, add HardSigmoid ([`847db99`](https://github.com/zama-ai/concrete-ml/commit/847db992f7ae01342f987ab3a38068d8c20bdeb9))
* Docs ([`8096acc`](https://github.com/zama-ai/concrete-ml/commit/8096accbb6b248a83383699e769d2b38dbed4707))
* Safer default parameter for ensemble methods ([`8da0988`](https://github.com/zama-ai/concrete-ml/commit/8da0988edeba21accb753e9dc41a33531df9febc))
* Increase n_bits for clear vs quantized comparison for decision tree ([`b9f1206`](https://github.com/zama-ai/concrete-ml/commit/b9f1206b9dd10348757451e408f528f5e4aa9ff1))
* Fix notebook on macOS + some warnings ([`ab2a821`](https://github.com/zama-ai/concrete-ml/commit/ab2a821a55af729457da4d93eda56385c462bf9d))
* Xgboost handle the edge case where n_estimators = 1 ([`3673584`](https://github.com/zama-ai/concrete-ml/commit/36735846d1c61df9f14e4d3add273a4f87d856d4))
* Issues in Classifier Comparison notebook ([`3053085`](https://github.com/zama-ai/concrete-ml/commit/30530850f12182ade18fa092c71332167165bd74))
* One more bug about convergence ([`c6cee4e`](https://github.com/zama-ai/concrete-ml/commit/c6cee4e56025d0399d7a46f27822a55c64e2ec55))
* Fix convergence issues in tests ([`7b92bd8`](https://github.com/zama-ai/concrete-ml/commit/7b92bd87ad62d5749643dd2392b203a975251d4a))
* Remove metric evaluation for n_bits < 16 ([`7c4bd0e`](https://github.com/zama-ai/concrete-ml/commit/7c4bd0e7950e3fb204a916745779e12604d00dc5))
* Wrong xgboost init ([`2ed49b6`](https://github.com/zama-ai/concrete-ml/commit/2ed49b651acf7e7ab98e915e55ff928f8985ce87))
* Workaround while 518 is being investigated ([`7f521f9`](https://github.com/zama-ai/concrete-ml/commit/7f521f90ea8a2f37d1d88adb713879657443585b))
* Looks like a mistake ([`69e9b15`](https://github.com/zama-ai/concrete-ml/commit/69e9b157b3dd2252e9a2e0bbb540cfb68dbaee25))
* Speedup qnn tests ([`9d07f5c`](https://github.com/zama-ai/concrete-ml/commit/9d07f5c5f2ab5fe159fa4a94d6f82b231b04c69d))
* Workaround for segfaults on macOS ([`798662f`](https://github.com/zama-ai/concrete-ml/commit/798662f5b1d252bb2b7d6e095f7f2641ddf77f43))
* Remove check_r2_score with argmax predictions ([`7d52750`](https://github.com/zama-ai/concrete-ml/commit/7d5275035454e8c8cbfcc22fa86d803573e94208))
* Review ([`82abb12`](https://github.com/zama-ai/concrete-ml/commit/82abb127aa6d92df8f89aa59a637ad5d0a2fbd4f))
* Fully connected notebook ([`1f7b92e`](https://github.com/zama-ai/concrete-ml/commit/1f7b92e2623ebf45d3104b43acda9d2b68bb5c15))
* When we test determinism, it is fine if there is an issue in the underlying Concrete Numpy ([`6595495`](https://github.com/zama-ai/concrete-ml/commit/65954951a4895aed67bc73f93b9b0bc2dacc97b7))
* Change the md5, even if the licence hasn't changed ([`182084f`](https://github.com/zama-ai/concrete-ml/commit/182084f4df41b27aa3aae62fce4ad5349d75beb1))
* Decision tree bug ([`84a65e4`](https://github.com/zama-ai/concrete-ml/commit/84a65e42959f01edb7d046730d84f5b9dfa7da33))
* Remove gpl lib + update sphinx-zama-theme ^2.2.0 ([`65aa1b2`](https://github.com/zama-ai/concrete-ml/commit/65aa1b2c5dc6230e14209bd86ec2d39133729f7f))
* Remove Hardsigmoid and Tanhshrink for a moment, since there are issues with precision ([`51c0bc5`](https://github.com/zama-ai/concrete-ml/commit/51c0bc5929f49691247092aaa56196e689552731))
* Remove fc comparison fhe vs quantization ([`1c527be`](https://github.com/zama-ai/concrete-ml/commit/1c527bee2383ed4a803aa53e9a890895610b6caa))
* Use right imports in docs ([`9fe43bf`](https://github.com/zama-ai/concrete-ml/commit/9fe43bf9b8bd801c9a8ceac9b8455205ee4d2354))
* Change qvalues to values in quantized module and fix iris notebook mistake ([`11c5616`](https://github.com/zama-ai/concrete-ml/commit/11c5616c5a92cbef39a74b60f6a99200b9336c83))
* Wrong fixture for a list + flaky test for decision tree + add fixture for model check is good execution ([`cc3c0b6`](https://github.com/zama-ai/concrete-ml/commit/cc3c0b67aeb11e610f6b6f89941c98fc0b44736c))
* Add missing docstrings ([`0c164f5`](https://github.com/zama-ai/concrete-ml/commit/0c164f588358b88d50327ccda5c710ab85dd5ad4))
* Fix docstrings which are incomplete thanks to darglint ([`45d4fca`](https://github.com/zama-ai/concrete-ml/commit/45d4fca9e7abcdc372ba738d16927b4ed15aad1f))

Documentation
* Refresh notebooks ([`ff771aa`](https://github.com/zama-ai/concrete-ml/commit/ff771aa9fa8a3ced5a111c4a499f7eed2eaa4fc8))
* Update the theme ([`0d1e672`](https://github.com/zama-ai/concrete-ml/commit/0d1e672e7054b650d14224bc342eb47d5e356b6a))
* Update simple example readme ([`21d9a77`](https://github.com/zama-ai/concrete-ml/commit/21d9a775855fc1dd8e5452bee0dd1475b39f7a53))
* Readme ([`029237a`](https://github.com/zama-ai/concrete-ml/commit/029237a7ec3c6b7458587710b99d1bc0ad72a41d))
* Update compute with quantization ([`b836811`](https://github.com/zama-ai/concrete-ml/commit/b836811cbd1046304e4ec8bde56a14e08c79551d))
* Rewrite the developer section for Quantization, show how to to work with quantized operators ([`436e71e`](https://github.com/zama-ai/concrete-ml/commit/436e71ec1dcbb568d144c75bb5a920b96245c9d6))
* Add Pruning docs ([`33b044f`](https://github.com/zama-ai/concrete-ml/commit/33b044fa7dd8d42d401c7c92e1b56876f4d8a7e7))
* Add info on skorch ([`6b3ca04`](https://github.com/zama-ai/concrete-ml/commit/6b3ca04a234c721bf7b689442b036e2983b627ab))
* Adding documentation ([`ed9ee3f`](https://github.com/zama-ai/concrete-ml/commit/ed9ee3fbe5a3d57e5c79cf035463cb7da32d9f65))
* Adding documentation ([`c4e73ec`](https://github.com/zama-ai/concrete-ml/commit/c4e73ec1a5c51a6036b890ed4be2e2985376cb0b))
* Improve quantization explanation in the User Guide ([`4508282`](https://github.com/zama-ai/concrete-ml/commit/45082828207be0222adf0bf0ac12f944923f1e58))
* Add a summary of our results ([`1046cc2`](https://github.com/zama-ai/concrete-ml/commit/1046cc26cfcea410afc24bab5b2d14259a9378ef))
* Write Virtual Lib documentation for release ([`4f68f3f`](https://github.com/zama-ai/concrete-ml/commit/4f68f3f12b65bceaedc25f724e53d2944daec126))
* Add hummingbird usage ([`05103b3`](https://github.com/zama-ai/concrete-ml/commit/05103b3550083057e0e1b868b4a67592751f4863))
* Update docs for release ([`95a1669`](https://github.com/zama-ai/concrete-ml/commit/95a166910a5a767ccacd8513aeccdda2147cbad3))
* Update our project setup doc ([`beef6c9`](https://github.com/zama-ai/concrete-ml/commit/beef6c967250242889cc5325850b205941418775))
* Update README ([`51ed1be`](https://github.com/zama-ai/concrete-ml/commit/51ed1bec70f9e32d93f141d7882741a88dab87ea))
* Add automatic ToC to README ([`4d51c96`](https://github.com/zama-ai/concrete-ml/commit/4d51c964c20fbf3cdfd09f83a4880110f00eccd3))
* Add source in docs ([`37227c6`](https://github.com/zama-ai/concrete-ml/commit/37227c6e24f252b36497aef8b2321785434131b9))
* Small update to the docker set up instructions ([`833d6e4`](https://github.com/zama-ai/concrete-ml/commit/833d6e4ee611e38084ed36e9b6aacd2a1a4f4d61))
* Update contributing to mention `make conformance` ([`bff86ca`](https://github.com/zama-ai/concrete-ml/commit/bff86ca0a3e0d36469ed6def3c11071ad05112ce))
* No need to update releasing.md ([`179e235`](https://github.com/zama-ai/concrete-ml/commit/179e235b046461ac3443837e92e934f929a523b4))
* Add a pruning section. ([`c977a32`](https://github.com/zama-ai/concrete-ml/commit/c977a324f0c0db9d236e3cb7a3d9debff45c7c58))
* No RF or SVM dedicated notebook ([`6307508`](https://github.com/zama-ai/concrete-ml/commit/6307508ca9730c17cd92a78de58dc41b7b3edb0b))
* Warn the user that GLM and PoissonRegression are currently not natively in the package ([`e3e0234`](https://github.com/zama-ai/concrete-ml/commit/e3e0234f73076fec88776d5b493228a1e47c89b4))
* Add Random Forest to our classifier comparison ([`858f193`](https://github.com/zama-ai/concrete-ml/commit/858f193888ca7714d3bbadbcb19ba1ce10d97e2b))
* Add XGBClassifier to our classifier comparison ([`eff1b15`](https://github.com/zama-ai/concrete-ml/commit/eff1b15149bde4bf48d0f907ea90afd838e68d11))
* Update our documentation ([`2b16560`](https://github.com/zama-ai/concrete-ml/commit/2b16560055ce2d14844a2631fbb19321eab3520b))
* Add a comparison of our classifiers ([`ce0d24b`](https://github.com/zama-ai/concrete-ml/commit/ce0d24b3d0a06ecf48b8424b978ae770336c7589))
* Make the plan for the documentation ([`0306cee`](https://github.com/zama-ai/concrete-ml/commit/0306cee62db302b1a48d95e28e180722c2173417))
* Add a sentence about quantized module 237 ([`a440de3`](https://github.com/zama-ai/concrete-ml/commit/a440de3860823f3fe8be78514e16d591349f0948))
* Use 2.1.0 theme ([`4fb1445`](https://github.com/zama-ai/concrete-ml/commit/4fb1445d24eab96e4d33bcf33e88b6039fc33216))
* Add starter docs for how ONNX is used internally ([`16978b6`](https://github.com/zama-ai/concrete-ml/commit/16978b607b3ba83445a536deca7e37750a36c36b))
* Add relevant docs from concrete-numpy ([`235322a`](https://github.com/zama-ai/concrete-ml/commit/235322ae13a0b24a32300832bbf5b6445d89d31c))
* Check mdformat ([`c29504a`](https://github.com/zama-ai/concrete-ml/commit/c29504a028b6f3d2c8ca71df5a3f8c0d43cd3afe))