You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This is fairly minimal, I wondered if there was a model card available that went into a bit more detail about
Training data, augmentation and preprocessing methods
Training metrics, losses, number of epochs, etc
Intended uses and uses to avoid
I tried using this model to extract vector embeddings for similarity search here NERC-CEH/plankton_ml#5 and the results seemed promising, but the model interpretations i tried making in NERC-CEH/plankton_ml#6 were completely inconclusive, it's at this point I would have been very interested to read a model card as part of the scivision distribution
The text was updated successfully, but these errors were encountered:
@metazool thanks for your interest in the model. You can find further details of the research in the Data Study Group report. Note the shared model is only for binary classification, so it's restricted to the taxonomy level, copepod / non-copepod. Please refer to the report for further limitations.
https://sci.vision/#/model/resnet50-plankton
This is fairly minimal, I wondered if there was a model card available that went into a bit more detail about
I tried using this model to extract vector embeddings for similarity search here NERC-CEH/plankton_ml#5 and the results seemed promising, but the model interpretations i tried making in NERC-CEH/plankton_ml#6 were completely inconclusive, it's at this point I would have been very interested to read a model card as part of the
scivision
distributionThe text was updated successfully, but these errors were encountered: