Rocket

Is your feature request related to a problem? Please describe.

It may be hard to understand why a post is in one of our timelines, especially in the case of someone boosting/linking someone else you don't follow, or when the same post appears multiple times because it got boosted.

Describe the solution you'd like

Adding a little "X boosted/liked" indicator to each concerned card

Dear team,

I am in stuck when convert very large numpy array to your TSDatasets.
These are what I have tried to fix my issue:

• when building time-series, I used tensorflow.keras.preprocessing.timeseries_dataset_from_array. After this step, the memory is still fine
• I concatenate all batch data into numpy array, this step produces problem so I use numpy memmap to avoi

@giovaniceotto. Thank you so much.I have run the Monte Carlo analysis successfully.And the new reference documents are so comprehensive and detailed. Besides，I made some changes to the sample code ，so that I can get simulation data of every timestep and do more in-depth analysis. All data will be saved in an excel file in the location you specify.

The specific methods are as follows:
**Fir