Ellen Zhong

Noted! What is the format that the original particles are in? I assume just in a .star or .cs file? Right now, you can copy over the micrograph coordinates (`--ref-star`...

Mixed precision training has been available since cryodrgn version 0.3.0 via nvidia's [apex](https://nvidia.github.io/apex/) library. I believe the latest versions of pytorch (1.6+) now natively support mixed precision mode. Eventually, we...

Fixed in commit 005afc2563dcefffe17985d562f4d4fe9196129a ``` $ cryodrgn parse_ctf_csparc J150/P10_J150_passthrough_particles.cs -o ctf.pkl -D 128 --Apix 2.49 2021-04-18 13:38:00 1313694 particles 2021-04-18 13:38:00 Image size (pix) : 128 2021-04-18 13:38:00 A/pix :...

Can you provide the command you used? Is your downsampled image stack formatted as a single `.mrcs` file you could provide to `write_starfile.py` instead?

I see -- you want the `.star` file to reference the original (full resolution) images. Currently only a single `.mrcs` or a `.txt` file containing a list of `.mrcs` are...

Note I recently converted the write_starfile.py script into a cryodrgn command, e.g. `cryodrgn write_starfile ...`. This will be in the upcoming cryoDRGN 1.0 version.

Can you provide the command you ran and some more details on the dataset? How many particles and what are the dimensions of the dataset? Also good news, I have...

Wow that is a lot of epochs! Yes, I have run into nans during training before, but only with zdim=1 and usually triggered by impurities/non-standard images in the dataset. By...

I'm not sure cryoSPARC has the functionality to be able to reimport a particle stack that refers to a (subset of) previously imported particles. I agree this would be a...

Thanks for the question! I would recommend a C1 refinement and to start the refinement from an _asymmetric initial model_. The goal is to get the true pose for each...