could you tell me what the ouput means when I did like this

model = ReformerLM(
            vocab_size=model_config['vocab_size'],
            d_in=model_config['n_seq_len'],
            d_out=model_config['n_embd'],
            n_layers=model_config['n_layers'],
            n_heads=model_config['n_heads'],
            attn_k=model_config['attn_k'],
            attn_v=model_config['attn_v'],
        ) 
inputs:  [batch_size,seq_len,dim]=(4,150,100)
output =  model(inputs)

thanks

Thanks for your sharing model!
However, it seems that you have directly split inputs into (#chunks), while the original paper used Random Projection, and then take the argmax as its hash value to determine the chunk?

Additionally, have you add the LSH part into the project? Since I am confused how to implement it in code, I would appreciate it very much if you can give some guidance!

Thanks in advance!

hi, I don't find the implementation of the LSH attention. Is it in progress?
thx