Feature
* Lora fine-tuning in FHE + gpt2 use case example ([823](https://github.com/zama-ai/concrete-ml/issues/823)) ([`4d2f2e6`](https://github.com/zama-ai/concrete-ml/commit/4d2f2e6b43c7cd0e628cd668b95f0132287568b6))
* Add GPU support ([849](https://github.com/zama-ai/concrete-ml/issues/849)) ([`945aead`](https://github.com/zama-ai/concrete-ml/commit/945aead0922b08cd9b177f6082f3c60349b4d118))
* Adding support for python3.11 ([701](https://github.com/zama-ai/concrete-ml/issues/701)) ([`819dca7`](https://github.com/zama-ai/concrete-ml/commit/819dca7f35a6a8750f6756207bc7b0a4a443718f))
* Support embedding layers ([778](https://github.com/zama-ai/concrete-ml/issues/778)) ([`296bc8c`](https://github.com/zama-ai/concrete-ml/commit/296bc8c15864d27576891a26251519bd0cd3e63d))
* Support encrypted mul div ([690](https://github.com/zama-ai/concrete-ml/issues/690)) ([`a1bd9b8`](https://github.com/zama-ai/concrete-ml/commit/a1bd9b83eee6a1e91f18d4094ea108f743af2cde))
* Upgrade pytorch to 2.3.1, brevitas to 0.10 ([788](https://github.com/zama-ai/concrete-ml/issues/788)) ([`c3d7c81`](https://github.com/zama-ai/concrete-ml/commit/c3d7c812ef0f4bcc01d7d0db49a80cebf18276b4))
* Add resnet use case ([738](https://github.com/zama-ai/concrete-ml/issues/738)) ([`6d8a205`](https://github.com/zama-ai/concrete-ml/commit/6d8a20596af4c13da536cfd3441d3aba057aaaed))
Fix
* Pip installation failing on new python version ([903](https://github.com/zama-ai/concrete-ml/issues/903)) ([`946cbab`](https://github.com/zama-ai/concrete-ml/commit/946cbabac2d9d14ba67992126bef8517c65ff0f0))
* Test warning sklearn ([902](https://github.com/zama-ai/concrete-ml/issues/902)) ([`f33b029`](https://github.com/zama-ai/concrete-ml/commit/f33b029e20d502cd1642b7426999ef2383ba8535))
* Update numpy for py311 and hummingbird ([901](https://github.com/zama-ai/concrete-ml/issues/901)) ([`153a9eb`](https://github.com/zama-ai/concrete-ml/commit/153a9eb4a616f4785e84945a8572a4cc92d2b9db))
* Relax python version restrictions ([853](https://github.com/zama-ai/concrete-ml/issues/853)) ([`040c308`](https://github.com/zama-ai/concrete-ml/commit/040c3084f014b033fbe8e825845c49492f8b10ee))
* Use correct torch version for intel mac ([798](https://github.com/zama-ai/concrete-ml/issues/798)) ([`d5f8f6c`](https://github.com/zama-ai/concrete-ml/commit/d5f8f6c4fa7ac65fcd4dcbaa183c4414b4892f91))
* Use correct torch version for intel mac ([798](https://github.com/zama-ai/concrete-ml/issues/798)) ([`a8eab89`](https://github.com/zama-ai/concrete-ml/commit/a8eab8947f07a6e3da28b36ae131d1c1a6ac0cc3))
* Protobuf 2gb limit when checking onnx ([811](https://github.com/zama-ai/concrete-ml/issues/811)) ([`c8908fa`](https://github.com/zama-ai/concrete-ml/commit/c8908fa31fbd8f7b9a93fe7d5d0f527607b3325f))
* Fix dtype check in quantizer dequant ([`77ced60`](https://github.com/zama-ai/concrete-ml/commit/77ced609f9f71c0efb54693ac32ca0640b534b56))
* Dynamic import of `transformers` in hybrid model ([`829b68b`](https://github.com/zama-ai/concrete-ml/commit/829b68b5b993049f5e177c1cb94d0c753ea3e80b))
Documentation
* Add encrypted LORA fine-tuning documentation ([887](https://github.com/zama-ai/concrete-ml/issues/887)) ([`3f65ec2`](https://github.com/zama-ai/concrete-ml/commit/3f65ec2dc2325794f2e08d1850a82362fbf84d2b))
* Language edits - "Deep Learning" and "Guides" ([786](https://github.com/zama-ai/concrete-ml/issues/786)) ([`1988936`](https://github.com/zama-ai/concrete-ml/commit/19889369ff3bff8ef77ae07d2a35b6d438a6397c))
* Update resources ([780](https://github.com/zama-ai/concrete-ml/issues/780)) ([`77cbdd6`](https://github.com/zama-ai/concrete-ml/commit/77cbdd61d0e44ab5875f8ee1c57d0831577042da))