Closed WhaleGe closed 1 year ago
@WhaleGe, seems you have some problem fetching the json, can you confirm you can access https://celltypist.cog.sanger.ac.uk/models/models.json
To use your own model, you can provide a path to the model. here, ./model_from_Elmentaite_2021.pkl
@WhaleGe, seems you have some problem fetching the json, can you confirm you can access https://celltypist.cog.sanger.ac.uk/models/models.json
To use your own model, you can provide a path to the model. here,
./model_from_Elmentaite_2021.pkl
Yes, I can't fetch the json. So, I provide a path to the model. But I found that I have to modify the /home/hw/.celltypist/data/models/models.json file. Like this, the program to run.
{
"description": "CellTypist model list",
"last_update": "2022-12-20 19:43:06.565435",
"models": [
{
"filename": "model_from_Elmentaite_2021.pkl",
"url": "",
"version": "v1",
"date": "2022-12-20 19:43:06.565435",
"details": "",
"No_celltypes": 134,
"source": "",
"default": true
}
]
}
@WhaleGe, the error comes from your inability to access the json file. It's better to solve the url issue at your end such that you can use a number of built-in models.
Another way to solve your problem (if you really have problem fetching the json) is to put your model in /home/hw/.celltypist/data/models/, and use celltypist.annotate(adata_James, model = 'model_from_Elmentaite_2021.pkl')
@WhaleGe, the error comes from your inability to access the json file. It's better to solve the url issue at your end such that you can use a number of built-in models.
Another way to solve your problem (if you really have problem fetching the json) is to put your model in /home/hw/.celltypist/data/models/, and use
celltypist.annotate(adata_James, model = 'model_from_Elmentaite_2021.pkl')
Ok I got it, thanks for your help
Thank you for providing such a useful tool, but when I run your the tutorial of Best practice in large-scale cross-dataset label transfer using CellTypist, why can't I read the locally saved model_from_Elmentaite_2021.pkl.