https://gitcode.net/greenplum/pytorch-widedeep/-/commit/0fdbfdcecba0e61cc083faa7ad2d00d9c88ebce4first step towards adding an example on how to reproduce a kaggle notebook...2023-07-24T17:01:50+02:00Javierjrzaurin@gmail.comfirst step towards adding an example on how to reproduce a kaggle notebook (details in the code) with this library
https://gitcode.net/greenplum/pytorch-widedeep/-/commit/21759eefe4fe56513ffec70f500511fe3cca165bAdded scripts on how to use the library for recsys in response to issue #133....2023-07-27T10:50:04+02:00Javierjrzaurin@gmail.comAdded scripts on how to use the library for recsys in response to issue #133. Also Added a simple/basic transformer model for the text component before integrating with HF. Also added the option of specify the dimension of the feed forward network
https://gitcode.net/greenplum/pytorch-widedeep/-/commit/d09446e53cb5ccb041b897d16e71290f44cc96c1Added unit tests. Need to write a notebook. Test is on GPU and ready to merge2023-07-27T12:55:45+02:00Javierjrzaurin@gmail.comhttps://gitcode.net/greenplum/pytorch-widedeep/-/commit/5b5e680871de6f151022bafca2447de5480e8890Added the 1st of two notebooks to illustrate the use of the library in the...2023-07-28T16:53:38+02:00Javierjrzaurin@gmail.comAdded the 1st of two notebooks to illustrate the use of the library in the context of recommendation systems
https://gitcode.net/greenplum/pytorch-widedeep/-/commit/04e9d38b500c37dc03dce13de0431d20ff39da5aAdded the notebooks to illustrate how to use the library to build...2023-07-28T22:25:31+02:00Javierjrzaurin@gmail.comAdded the notebooks to illustrate how to use the library to build recommendation algos. A couple of bugs to fix and ready to merge and publish
https://gitcode.net/greenplum/pytorch-widedeep/-/commit/8813ceeabaf4a9b37a5b8824a761c9fe12719715added movielens dataset and tests2023-07-30T21:24:14+02:00Pavol Mulinkamulinka.pavol@gmail.comhttps://gitcode.net/greenplum/pytorch-widedeep/-/commit/b6a1033639089c12352d64a2879d04921a03eabcMerge pull request #182 from jrzaurin/recsys_movielens_dataset2023-07-30T21:32:29+01:00Javierjrzaurin@gmail.com
added movielens dataset and testshttps://gitcode.net/greenplum/pytorch-widedeep/-/commit/d30203a0a02a75b5bf2a421ff0aac35c8b1c93b5Fixed a bug related to the padding idx and the fast ai transforms. Also...2023-07-31T13:11:32+01:00Javierjrzaurin@gmail.comFixed a bug related to the padding idx and the fast ai transforms. Also adjusted the scripts to show how one can use the 'load_movielens100k' function in the library
https://gitcode.net/greenplum/pytorch-widedeep/-/commit/801c597ae6d5d9224af43d9d495d59e81062f3b1Adjusted the notebooks to show how one can use the 'load_movielens100k' funct...2023-07-31T13:59:17+01:00Javierjrzaurin@gmail.comhttps://gitcode.net/greenplum/pytorch-widedeep/-/commit/e75c119073312e6bee047e52ad89a900125f524bbump version to 1.3.12023-07-31T17:36:48+01:00Javierjrzaurin@gmail.comhttps://gitcode.net/greenplum/pytorch-widedeep/-/commit/cd1ff79ae327ac025fe65e6dafe76f305bfd2cbbMerge pull request #183 from jrzaurin/wide_deep_recsys2023-07-31T18:01:57+01:00Javierjrzaurin@gmail.com
Wide deep recsyshttps://gitcode.net/greenplum/pytorch-widedeep/-/commit/52ae96b5829b60ce1f119db047574c96c950a571Merge remote-tracking branch 'origin/master' into flash_attention2023-08-02T10:42:06+01:00Javierjrzaurin@gmail.comhttps://gitcode.net/greenplum/pytorch-widedeep/-/commit/eb02f25f07908c0285fb3a0be7cfdb5186781576Added linear attention from the paper 'Transformers are RNNs: Fast...2023-08-02T13:12:53+01:00Javierjrzaurin@gmail.comAdded linear attention from the paper 'Transformers are RNNs: Fast Autoregressive Transformers with Linear Attention'. This now needs to be turned into an encoder and offer as an optional model
https://gitcode.net/greenplum/pytorch-widedeep/-/commit/7a80f3e7068032ee1b57ef07900bcbdd6f188ed7implemented linear and 'standard' attention in a functional way so they are...2023-08-02T15:20:57+01:00Javierjrzaurin@gmail.comimplemented linear and 'standard' attention in a functional way so they are available via parameters passed to the main multi head attention class
https://gitcode.net/greenplum/pytorch-widedeep/-/commit/2a74d34f6855db80c32a37ddaa346b1da4721aa1test passed. Need to increase a bit test coverage for the tabtransformer and...2023-08-02T18:18:36+01:00Javierjrzaurin@gmail.comtest passed. Need to increase a bit test coverage for the tabtransformer and attention_layers, and review the docs
https://gitcode.net/greenplum/pytorch-widedeep/-/commit/b6362d1d31e6ac71557539421c3875b4286d57ddAdded some docs. Only thing left is test new attention mechanism on GPU2023-08-03T18:20:53+01:00Javierjrzaurin@gmail.comhttps://gitcode.net/greenplum/pytorch-widedeep/-/commit/d3657c32d9a352940eab1d4436794b57a1432751Added a example of flash and linear attention. Fix some small bugs in one...2023-08-04T13:12:37+01:00Javierjrzaurin@gmail.comAdded a example of flash and linear attention. Fix some small bugs in one example. Adjusted all new functionality to GPU usage
https://gitcode.net/greenplum/pytorch-widedeep/-/commit/67439c4220fa6fdf979b32f7c0c9ea4d9f748437Bumped to version 1.3.22023-08-04T15:17:32+01:00Javierjrzaurin@gmail.com
"This is the second of the two notebooks where we aim to illustrate how one could use this library to build recommendation algorithms using the example in this [Kaggle notebook](https://www.kaggle.com/code/matanivanov/wide-deep-learning-for-recsys-with-pytorch) as guidance. In the previous notebook we used `pytorch-widedeep` to build a model that replicated almost exactly that in the notebook. In this, shorter notebook we will show how one could use the library to explore other models, following the same problem formulation, this is: given a state of a user at a certain point in time having watched a series of movies, our goal is to predict which movie the user will watch next. \n",
"\n",
"Assuming that one has read (and run) the previous notebook, the required data will be stored in a local dir called `prepared_data`, so let's read it:"
"...remember that in the previous notebook we explained that we are not going to use a validation set here (in a real-world example, or simply a more realistic example, one should always use it).\n"
"Also remember that, in the previous notebook we discussed that the `'maxlen'` and `'max_movie_index'` parameters should be computed using only the train set. In particular, to properly do the tokenization, one would have to use ONLY train tokens and add a token for new 'unknown'/'unseen' movies in the test set. This can also be done with this library or manually, so I will leave it to the reader to implement that tokenzation appraoch."
"From now one things are pretty simple, moreover bearing in mind that in this example we are not going to use a wide component since, in pple, one would believe that the information in that component is also 'carried' by the movie sequences (However in the previous notebook, if one performs ablation studies, these suggest that most of the prediction power comes from the linear, wide model).\n",
"\n",
"In the example here we are going to explore one (of many) possibilities. We are simply going to encode the triplet `(user, item, rating)` and use it as a `deeptabular` component and the sequences of previously watched movies as the `deeptext` component. For the `deeptext` component we are going to use a basic encoder-only transformer model.\n",