search

thaihungle / SAM

Self-attentive Associative Memory & SAM-based Two-Memory Model
MIT License
46 stars 8 forks source link

File Not Found... #1

Closed ghost closed 3 years ago

ghost commented 3 years ago

Please could I know from where I need to download the files required to execute this? Please help. Inkedfilenotfound_LI

thaihungle commented 3 years ago

Hi, Please find the dataset here https://pytorch.org/text/_modules/torchtext/datasets/babi.html

ghost commented 3 years ago

I had downloaded the files in this link https://github.com/siddk/entity-network/tree/master/tasks and then put them in a folder named 'data' and put that folder into 'babi' in here : https://github.com/thaihungle/SAM/tree/master/babi .... and i executed this command : python run_all_babi.py and I got the following error: strip2

It says NoneType object no attribute 'strip',... so nothing is stored in token? I don't know what I'm doin wrong. Please help. Thanks .

thaihungle commented 3 years ago

This error is new to me. I see you are reading pik files. Did you download the raw dataset and run the code with raw data first? It should generate the valid pik for other runs. Also, it's best to use torch 1.0.0.

ghost commented 3 years ago

Please may I know from where I can download the raw dataset? In this link you had shared before : https://pytorch.org/text/_modules/torchtext/datasets/babi.html , I'm not sure if I'm supposed to paste it in some code file that's already there and run it ?

thaihungle commented 3 years ago

The official raw dataset is here http://www.thespermwhale.com/jaseweston/babi/tasks_1-20_v1-2.tar.gz You can find the data folder en-valid-10k Put it under your ./babi/data/

ghost commented 3 years ago

I did as per your instructions. I downloaded the raw dataset and put the necessary files under ./babi/data.... Now Im getting the following error: babi_error

So I went and looked into the data folder, and the pik files weren't generated. What do I do now? Please help!!!

thaihungle commented 3 years ago

I use python 3.6 and this only yields a warning (see attached image) Maybe you should try python 3.6 or modify the code to ignore empty string. image