Sansa Gong

`articles_embeddings_1.pkl`: [articles_embeddings_1.zip](https://github.com/summmeer/session-based-news-recommendation/files/10132052/articles_embeddings_1.zip) `publish_time_1.txt` can be generated in `mind_preprocess.py`.

`articles_category` is a dictionary, pseudo code: ``` articles_category = {} for each news_id, news_category: articles_category[news_id] = news_category pickle.dump(articles_category, open('../articles_category.pkl', 'wb')) ``` Environment info: Python: 3.6.9, TensorFlow: 1.14.0

You may need to comment this line `gap.append(bucketized(t['active_t'])) # if use Adressa dataset` and uncomment these lines: https://github.com/summmeer/session-based-news-recommendation/blob/fd795a4a05588f641e5e8ad74ff80b477eead8ab/sampler.py#L89-L94

Yes, `interval=None` should be removed, this is for ablation.

function `bucketized` takes input seconds into 12 categories: [0, 11]. https://github.com/summmeer/session-based-news-recommendation/blob/fd795a4a05588f641e5e8ad74ff80b477eead8ab/sampler.py#L18 So the embedding dim should be 12, `vocab_size=12` https://github.com/summmeer/session-based-news-recommendation/blob/fd795a4a05588f641e5e8ad74ff80b477eead8ab/model_combine.py#L106 Or you can change `bucketized` function, it is the same,...

The log is normal, the question mark stands for batch size, which is unknown when building the graph for TensorFlow. I didn't encounter NaN before, maybe you can print the...

[articles_embeddings+titles.zip](https://github.com/summmeer/session-based-news-recommendation/files/10811688/articles_embeddings%2Btitles.zip)

Yes, I guess you can directly run it. The missing file is [here](https://drive.google.com/file/d/10HuyNTP8uAeiugyagLlIKqKO1SHEv5jM/view?usp=sharing).

The category file is easy to obtain, plz refer to https://github.com/summmeer/session-based-news-recommendation/issues/8#issuecomment-1334697445 The second question: yes, it's ok to do the replace

No, I think the memory usage is totally tolerable.