So you want to create a dataset that has each of the collapsed sequences put back into individual sequences, where the frequency of each is based on the count. There isn't a wrapped Galaxy tool that I know of to do this. And if you used the tool Collapse sequences (Galaxy Version 1.0.0) - FASTX-toolkit based, then the original sequence identifiers are no longer available.
A line-command script could be written to do this (and wrapped as a Galaxy tool). If you are interested in creating this, start here (use Planemo): https://galaxyproject.org/tools/
Next time, save back the original uncollapsed fasta dataset. It can be downloaded locally. Then after you confirm the download was successful, the dataset can be perm deleted from the history. This way, you can always upload it again if needed.
Thanks, Jen, Galaxy team