Skip to content
This repository has been archived by the owner on Oct 5, 2023. It is now read-only.

Cannot load and run embedding inference using local model #891

Open
handrew opened this issue Jun 30, 2023 · 0 comments
Open

Cannot load and run embedding inference using local model #891

handrew opened this issue Jun 30, 2023 · 0 comments

Comments

@handrew
Copy link

handrew commented Jun 30, 2023

Error

OSError: Cannot save file into a non-existent directory: '.zeno_cache/EMBEDDINGembeddingmodels'

my zeno.toml looks like:

functions = "functions"
view = "image-classification"
metadata = "metadata.csv"
data_column = "id"
data_path = "."
label_column = "label"
models = ["IMAGENET1K_V2", "models/my_own_model.pth"] # Add paths to model checkpoints, passed into @model

Potential Fix

The way zeno.classes.base.ZenoColumn is turned into a string cannot support models in subfolders. I suggest either replacing slashes with some other character, like underscores, or using the hash of the ZenoColumn in the run_inference function in data_processing.py.

I tried fixing it (using the former approach above) in ZenoColumn myself but I think it probably breaks some things elsewhere.

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant