out,state = tflib.ops.FreeRunIm2LatexAttention('AttLSTM',emb_seqs,ctx,EMB_DIM,ENC_DIM,DEC_DIM,D,H,W)

[cltdevelop]

Hello!
Thank you for sharing your code!
I have something wrong when I run this code.
Is there something wrong in the attention.py. The order of parameters in out,state = tflib.ops.FreeRunIm2LatexAttention('AttLSTM',emb_seqs,ctx,EMB_DIM,ENC_DIM,DEC_DIM,D,H,W) is not right? I mean the parameter named ctx should in the second place.

[m--i]

Is that related with the error ValueError: Dimension must be 3 but is 4 for 'transpose' (op: 'Transpose') with input shapes: [?,?,80], [4]

which occurs at
out,state = tflib.ops.FreeRunIm2LatexAttention('AttLSTM',emb_seqs,ctx,EMB_DIM,ENC_DIM,DEC_DIM,D,H,W)

in attention.py.

Is there a solution for it yet?

[m--i]

If you try to map all parameters given in attention.py and FreeRunIm2LatexAttention(…) in ops.py there a value for output_dim missing in attention.py.

Even in ops.py an explanation for this parameter is missing.

Which is the correct parameter to pass?