{"id":13934997,"url":"https://github.com/suriyadeepan/practical_seq2seq","last_synced_at":"2026-01-24T17:12:17.963Z","repository":{"id":86896565,"uuid":"77355961","full_name":"suriyadeepan/practical_seq2seq","owner":"suriyadeepan","description":"A simple, minimal wrapper for tensorflow's seq2seq module, for experimenting with datasets rapidly","archived":false,"fork":false,"pushed_at":"2018-01-12T00:54:03.000Z","size":7300,"stargazers_count":570,"open_issues_count":56,"forks_count":271,"subscribers_count":29,"default_branch":"master","last_synced_at":"2024-08-08T23:19:43.287Z","etag":null,"topics":["chatbot","neural-conversation-models","seq2seq","tensorflow"],"latest_commit_sha":null,"homepage":"http://suriyadeepan.github.io/2016-12-31-practical-seq2seq/","language":"Jupyter Notebook","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"gpl-3.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/suriyadeepan.png","metadata":{"files":{"readme":"README.markdown","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null}},"created_at":"2016-12-26T03:42:08.000Z","updated_at":"2024-07-29T15:44:50.000Z","dependencies_parsed_at":"2023-05-29T18:15:07.025Z","dependency_job_id":null,"html_url":"https://github.com/suriyadeepan/practical_seq2seq","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/suriyadeepan%2Fpractical_seq2seq","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/suriyadeepan%2Fpractical_seq2seq/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/suriyadeepan%2Fpractical_seq2seq/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/suriyadeepan%2Fpractical_seq2seq/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/suriyadeepan","download_url":"https://codeload.github.com/suriyadeepan/practical_seq2seq/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":226666460,"owners_count":17665034,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["chatbot","neural-conversation-models","seq2seq","tensorflow"],"created_at":"2024-08-07T23:01:20.854Z","updated_at":"2026-01-24T17:12:17.956Z","avatar_url":"https://github.com/suriyadeepan.png","language":"Jupyter Notebook","funding_links":[],"categories":["Jupyter Notebook"],"sub_categories":[],"readme":"# Seq2Seq Wrapper for Tensorflow\n\n![](https://img.shields.io/badge/python-3-brightgreen.svg) ![](https://img.shields.io/badge/tensorflow-0.12.0-yellowgreen.svg)\n\nTo make life easier for beginners looking to experiment with seq2seq model. Read the article I wrote on seq2seq - [**Practical seq2seq**](http://suriyadeepan.github.io/2016-12-31-practical-seq2seq/), for more details.\n\n\n## Datasets\n\n- [x] [CMU Pronouncing Dictionary](https://github.com/suriyadeepan/datasets/tree/master/seq2seq/CMUdict) : Phoneme sequence to word (sequence of alphabets)\n![](/img/phoneme.png)\n\n_______________________________________________________________\n\n- [x] [Twitter Chat Log](https://github.com/suriyadeepan/datasets/tree/master/seq2seq/twitter) : Sentence to Sentence\n\n![](/img/twitter01.png)\n![](/img/twitter02.png)\n\n**Update**\n\nAfter tweaking the dataset to bring down the overall % of unknown tokens to 3%, increasing vocabulary to 8000, the [results](https://gist.github.com/suriyadeepan/5d4aaf96c7bd89b908921804c683fee8) are promising.\n\n| Query\t\t\t\t\t| Reply\t\t\t\t\t|\n| ------------- | ------------- |\n| donald trump won last nights presidential debate according to snap online polls | thought he was a joke |\n| trump campaign please never delete this | he was a joke |\n| s new scoop is unk the castro connection how trumps company violated the us embargo against cuba | trump is a disgrace to the debate and the media is not a liar |\n| who won the first presidential debate | trump will be a better time for a man |\n| just wanna live in unk everything is 10x better there | i was just on the same side and i was like it was a good time |\n| the lil girl i forgot her name scares the shit out of me n probably always will lmaooo | she was so cute and she was so cute and she was a bitch |\n| question of the day who do you think won last nights debate weigh in using | hillary is a trump supporter and a woman who has a better job than a woman |\n| trump is a total embarrassment hes like a unk child who happens to be a professional liar and con man | trump is a lot of money to get a great country |\n| chill the fuck out | thats what i said |\n| so fun to see what the special effects team created for | wow that is the best thing ive ever seen |\n| she was fine too unk voice | she is a good guy |\n| the important ones are still here | i dont think so |\n| been drinking pumpkin spice protein shake every morning for a week and now i literally cannot even | me too i just got it |\n| lmao i just lost my job and i cant even leave early to be in the unk video today | literally the same thing |\n| hey happy birthday have a nice day | thank you |\n\n_______________________________________________________________\n\n- [x] [Cornell Movie Dialog Corpus](https://github.com/suriyadeepan/datasets/tree/master/seq2seq/cornell_movie_corpus)\n\n\\~4% unknown tokens in dataset; vocabulary of 8000.\n\n| Query\t\t\t\t\t| Reply\t\t\t\t\t|\n| ------------- | ------------- |\n| when are you going to let the police know | you dont know what youre talking about |\n| you can do it | yeah i think so |\n| wait are you saying that i dont appreciate | i know you know what i mean |\n| youre terrible | yeah but im not |\n| nah i dont think so | why not |\n| hes pissing in our faces again and were just taking it | i cant |\n| im sorry i dont know what else to say except im sorry | why would you like to be |\n| nobodys gonna get you now get inside | i cant go back |\n| shall i leave | of course |\n| well i really think hes got a chance | i know |\n| youd better be quiet sandy | shut up |\n| jesus christ you scared the shit out of me | whats going on |\n| well im sorry im really sorry ellie | its okay |\n| my lady this play will end badly i will tell | lets get out of here |\n| im going to miss you | no youre not |\n| what do you mean | i dont know i dont know what i mean |\n| my god these people are insane | we dont know what they are |\n| this isnt a date | no what is it |\n| you ought to go home and take care of that | i cant do that |\n| is something wrong | no no no |\n\n\n## Credits\n\n- Borrowed most of the code for [seq2seq_wrapper.py](/seq2seq_wrapper.py) from [mikesj-public](https://github.com/mikesj-public/rnn_spelling_bee/blob/master/spelling_bee_RNN.ipynb)\n- Borrowed the twitter dataset from this dude [Marsan-Ma](https://github.com/Marsan-Ma/)\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fsuriyadeepan%2Fpractical_seq2seq","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fsuriyadeepan%2Fpractical_seq2seq","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fsuriyadeepan%2Fpractical_seq2seq/lists"}