Easydel

Latest version: v0.0.67

Safety actively analyzes 638773 Python packages for vulnerabilities to keep your Python projects secure.

Scan your dependencies

Page 2 of 3

0.0.50

What's Changed
* Optimize mean loss and accuracy calculation by yhavinga in https://github.com/erfanzar/EasyDeL/pull/100
* Mixtral Models are fully supported and they are `PJIT-compatible`
* A Wider range of models now support FlashAttention on TPU
* Qwen 1, Qwen 2, PHI 2, Robert is new Added Models which support FlashAttention on TPU and `EasyBIT`
* LoRA support for the trainer is now Added (`EasyDeLXRapTureConfig`)
* Adding EasyDel Serve Engine APIs
* Adding Prompter (Beta and might be removed in future updates)
* The Training Process is now 21 % Faster in `0.0.50` than `0.0.42`.
* Transform Functions are now Automated for all the models (Except `MosaicMPT` for this one you still have to use static methods)
* The Trainer APIs have changed and now it's faster, more dynamic, and more hackable.
* Default Version of the JAX now changed to 0.4.22 for `FJFormer` custom Pallas kernels usage.

New Contributors
* yhavinga made their first contribution at https://github.com/erfanzar/EasyDeL/pull/100

**Full Changelog**: https://github.com/erfanzar/EasyDeL/compare/0.0.42...0.0.50

0.0.42

New Features:

- `EasyDelState` is added
- Auto Convertors from torch > huggingface > jax > flax > EasyDel are added
- Trainer has a lot of improvements

**Full Changelog**: https://github.com/erfanzar/EasyDeL/compare/0.0.41...0.0.42

0.0.41

- API Changes
- making CausalLanguageModel Trainer separated from others
- Custom Errors added
- Timer bugs fixed
- `AutoEasyDelForCasualLM` is now more automated and falcon bugs have been fixed
- 4D Mesh being used for better partitioning
- And many many more

**Full Changelog**: https://github.com/erfanzar/EasyDeL/compare/0.0.40...0.0.41

0.0.40

What's Changed
* Updating JAXBeta branch from main branch by erfanzar in https://github.com/erfanzar/EasyDeL/pull/42
* Update Beta Branch by erfanzar in https://github.com/erfanzar/EasyDeL/pull/48
* Update V0.0.40 Beta (Adding Flash Attention, Adding 8,6,4 Bit models ,improving Documentations) by erfanzar in https://github.com/erfanzar/EasyDeL/pull/52
* Fix eval batch loop (beta branch) by w11wo in https://github.com/erfanzar/EasyDeL/pull/51
* Support Sphinx Docstring Format by w11wo in https://github.com/erfanzar/EasyDeL/pull/53
* Update Beta Branch by erfanzar in https://github.com/erfanzar/EasyDeL/pull/55
* Update Beta Branch by erfanzar in https://github.com/erfanzar/EasyDeL/pull/56
* Updating Mistral and Llama Models by erfanzar in https://github.com/erfanzar/EasyDeL/pull/57
* Updating Beta Branch by erfanzar in https://github.com/erfanzar/EasyDeL/pull/58
* Changing Mesh by erfanzar in https://github.com/erfanzar/EasyDeL/pull/60
* Updating Beta by erfanzar in https://github.com/erfanzar/EasyDeL/pull/62
* 4D Mesh now supported for all the Models and BITs improved by erfanzar in https://github.com/erfanzar/EasyDeL/pull/64

**Full Changelog**: https://github.com/erfanzar/EasyDeL/compare/0.0.38...0.0.40

0.0.38

Changes and Latest Commits:
- [remove setting default Device to fix PJRT TPU Device Mismatch Bug](https://github.com/erfanzar/EasyDeL/commit/3bcc24952d21680d36a70761b28352ed403a108e)

- [Update issue templates](https://github.com/erfanzar/EasyDeL/commit/4850599d0e56cdd37c19bacb5edd88437bfed638)

- [Create CONTRIBUTING.md](https://github.com/erfanzar/EasyDeL/commit/3585ac7361920652fe3104f3f3172a0afa95edb8)

- [TODO: Adding Auto Evaluators for Models](https://github.com/erfanzar/EasyDeL/commit/1d28721b6d7aaeccd0fc0425957add9a73092874)

- [TODO:Adding TF Dataset for optional Use](https://github.com/erfanzar/EasyDeL/commit/b1e6937afcf40dc7a7928baa5a0e97412f10a291)

- [TODO: Adding RLTrainer](https://github.com/erfanzar/EasyDeL/commit/8a82685f51bb30d159a492e660f0918304f7dee4)

- [Adding BasicRL Model With ValueHead](https://github.com/erfanzar/EasyDeL/commit/689d54c3355412f59cae3b80b81d8847a4e19dc1)

**Full Changelog**: https://github.com/erfanzar/EasyDeL/compare/0.0.37...0.0.38

0.0.37

What's Changed
* Adding Llama Mojo by erfanzar in https://github.com/erfanzar/EasyDeL/pull/43
* [Version 0.0.37 imports are optimized and python 3.12 is supported](https://github.com/erfanzar/EasyDeL/commit/20ca2e6cb551e44cac90bd0bd4d21f5aa52fa4ed)
* [Chaning Imports due to Memory crashes on python 3.10 on mid-end CPUs](https://github.com/erfanzar/EasyDeL/commit/2a42d9c83399b0f712d9ccc8304b45ce098fbc34)
* [Adding do_sample as defualt op for PytorchServer](https://github.com/erfanzar/EasyDeL/commit/c981e239c3d6b8231b31b3b5ffb39fd5807ee532)

**Full Changelog**: https://github.com/erfanzar/EasyDeL/compare/0.0.31...0.0.37

Page 2 of 3

© 2024 Safety CLI Cybersecurity Inc. All Rights Reserved.