Closed openvino-dev-samples closed 1 week ago
View / edit / reply to this conversation on ReviewNB
eaidova commented on 2024-09-04T05:08:54Z ----------------------------------------------------------------
Line #4. class blip2model(torch.nn.Module):
general code style in python to use CamelCase for classes nameing and lowercase underscore for class fields , variables and arguments, so
blip2model -> Blip2Model
Qformer -> q_former
View / edit / reply to this conversation on ReviewNB
eaidova commented on 2024-09-04T05:08:55Z ----------------------------------------------------------------
Line #12. if model_type.value == "Blip2":
I see a lot of if-else statemet with the same actions, maybe it is better to move into reusable functions?
View / edit / reply to this conversation on ReviewNB
eaidova commented on 2024-09-04T05:08:55Z ----------------------------------------------------------------
I think it is not only Blip2 issue, CLIP generally mostly trained on English data, I think translation will be useful for other models as well. Can we make this description is more generic? Can we also high-light that we use optimum-intel for that and provide link on model card?
View / edit / reply to this conversation on ReviewNB
eaidova commented on 2024-09-04T05:08:56Z ----------------------------------------------------------------
Line #3. trans_model_path = "ov_models/trans_model"
maybe it is better to be more specific and add languages in dir naming (potentially, if it will be requested in future, we can extend it on other languages)?
View / edit / reply to this conversation on ReviewNB
eaidova commented on 2024-09-04T05:08:57Z ----------------------------------------------------------------
maybe it will be also good to mention about translation in this description?
Check out this pull request on
See visual diffs & provide feedback on Jupyter Notebooks.
Powered by ReviewNB