Abhijnan/AxomiyaBERTa

Size of the model is 48MB

Example of usage:

tokenizer = AutoTokenizer.from_pretrained("Abhijnan/assamese_albert")
model = AutoModel.from_pretrained("Abhijnan/assamese_albert")

assamese_text = 'সঁচাই পৃথিৱীখন যে ইমান সুন্দৰ'
assamese_token = tokenizer(assamese_text ,padding='longest', return_tensors="pt")
print(assamese_token)
outputs = model(assamese_token["input_ids"],assamese_token["attention_mask" ],  output_attentions=True) 
print(outputs[0].shape)  #for getting the last hidden states of the sequence

'''
Output:
--------

{'input_ids': tensor([[   3, 8799, 6403,   38,  630, 1092,    4]]), 'token_type_ids': tensor([[0, 0, 0, 0, 0, 0, 0]]), 'attention_mask': tensor([[1, 1, 1, 1, 1, 1, 1]])}

torch.Size([1, 7, 768])

Created by:
Abhijnan Nath
Email: abhijnan.nath@colostate.edu
Situated Grounding and Natural Language (SIGNAL) Lab (https://www.signallab.ai/)
Computer Science Department, CSU Fort Collins.
Linkedin: (https://www.linkedin.com/in/abhijnan-nath-737727169/
Github : https://github.com/AbhijnanNath)