Broever101

What's the way to save and load a model currently? I'm saving it like so ```julia BSON.@save bsonname bert_model wordpiece tokenizer ``` And loading it using `load_pretrain_bert(bsonname)` but it throws...

> Simply `BSON.@save` and `BSON.@load`. I guess the error is probably because you forget to `using Transformers` before loading. And yes it's better to do `cpu(model)` before saving. Weird. I...

> Did you ever resolve this issue? I'm having the same. Using the training example Peter linked

> The main issue for the pretrain example is that the code is really really old and I haven't find time to update that. But I didn't make it a...

So these are all the weights that are out of shape ``` weights[5] = reshape(weights[5], (1150, 1)) weights[6] = reshape(weights[6], (1150, 1)) weights[10] = reshape(weights[10], (1150, 1)) weights[11] = reshape(weights[11],...

> > While running the example code: > > ```python > > from pyserum.connection import get_live_markets, get_token_mints > > print("tokens: ") > > print(get_token_mints()) > > print("markets: ") > >...