error: No such file or directory about sosd

[andreaskipf]

Hi Jingyang, thanks for reaching out. The dataset is called `fb_200M_uint64` (not `20M`). Did you try running `reproduce.sh` in the root directory?

…

On Wed, Apr 14, 2021 at 5:10 AM Jinyang Li ***@***.***> wrote: Hi, I'm trying to run your code and get this error (the dataset is fb_20M_uint64): [image: image] <https://user-images.githubusercontent.com/24719703/114648597-b4e76300-9d11-11eb-9039-c36dc6a9a748.png> Could you please suggest how to solve this problem? Thank you! — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#19>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAUJBEQFJ5DGEGLOKOQVBR3TIUBS3ANCNFSM424P7CYQ> .

[niceIrene]

Hi, I had the same problem. I changed the downsample.py file to get a smaller dataset, say fb_20M_uint. When I was running the build_rmis.sh script, by using the command: "build_rmi_set fb_100M_uint64" in build_rmis.sh, it reported the same error. I looks like line15: RMI/target/release/rmi data/$DATA_NAME --param-grid ${JSON_PATH} -d rmi_data/ --threads 8 --zero-build-time is not working.

[niceIrene]

It looks like in order to run the other datasets, we need to add something to scripts/rmi_specs as you did for the other datasets?

[RyanMarcus]

In order to build RMIs for a new dataset (or a downsampled one), you'll have to generate a configuration for training the RMI on that dataset. The config files for each dataset are stored in the scripts/rmi_specs directory. You could try copying the RMI config for the larger dataset (i.e., cp fb_200M_uint64.json fb_20M_uint64.json), but this will likely produce suboptimal results.

The best way to try the smaller dataset would be to run the RMI optimizer to create a configuration for the smaller dataset specifically.

[niceIrene]

Thanks!