Kan-gpt

------------------
- Release: version 1.0.0 🚀 [Aditya NG]
- Version increment. [Aditya NG]
- Merge pull request 11 from wektorz/fix_missing_dataset_dowload.
[Aditya]

add missing dataset dowload to setup that caused train to fail in jupyter notebook
- Update KAN_GPT.ipynb. [Wiktor Zdrojewski]

add missing dowload that caused train to fail
- Docs(mkdocs): results added. [Aditya NG]
- Docs(README,media): results and metrics. [Aditya NG]
- Feat(train.py): metrics added. [Aditya NG]
- Release: version 0.4.0 🚀 [Aditya NG]

------------------
- Release: version 0.4.0 🚀 [Aditya NG]
- Version increment. [Aditya NG]
- Fix(kan_gpt/efficient_kan/__init__.py): missing init file. [Aditya NG]
- Feat(README.md): results added. [Aditya NG]
- Docs(mkdocs): documentation for mkdocs. [Aditya NG]
- Feat(sweep): slower learning rate. [Aditya NG]
- Fix(sweep): max_iters. [Aditya NG]
- Fix(sweep): cuda device. [Aditya NG]
- Ci(main): disabled wandb in ci. [Aditya NG]
- Test(tests/test_train.py): disable wandb during testing. [Aditya NG]
- Fix(sweep): added cuda clean. [Aditya NG]
- Feat(sweep): reduced batch size. [Aditya NG]
- Feat(download_dataset): functions to download tinyshakespeare and
webtext. [Aditya NG]

------------------
- Release: version 0.3.0 🚀 [Aditya NG]
- Docs(README.md): dataset mention. [Aditya NG]
- Feat(sweep): sweep script for getting a vast hyperparam sweep. [Aditya
NG]
- Feat(tinyshakespeare): support for another dataset. [Aditya NG]
- Test(tests/test_prompt.py): eval code. [Aditya NG]
- Test(efficient_kan,original_kan): coverage. [Aditya NG]

------------------
- Release: version 0.2.0 🚀 [Aditya NG]
- Version increment. [Aditya NG]
- Docs(KAN_GPT.ipynb): benchmark runs between mlp and kan gpts. [Aditya
NG]
- Feat(train): date id of asset. [Aditya NG]
- Feat(train,prompt): saving model after training, prompting a model
given the saved model path. [Aditya NG]
- Feat(kan_gpt/dataset.py): lengthwise dataset iteration. [Aditya NG]
- Fix(dataset): index error. [Aditya NG]

------------------
- Release: version 0.1.3 🚀 [Aditya NG]
- Fix(dataset): padding and range fix. [Aditya NG]

------------------
- Release: version 0.1.2 🚀 [Aditya NG]
- Feat(kan_gpt/efficient_kan/): integration with efficient kan reduced
training memory footprint. [Aditya NG]
- Test(gpt-pico): reducing memory footprint. [Aditya NG]