The following error occurs when using mamba_chunk_scan_combined in mamba-ssm-2.2.0:
File "/usr/local/lib/python3.10/dist-packages/torch/_tensor.py", line 522, in backward
torch.autograd.backward(
File "/usr/local/lib/python3.10/dist-packages/torch/autograd/init.py", line 266, in backward
Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass
RuntimeError: function MambaChunkScanCombinedFnBackward returned an incorrect number of gradients (expected 16, got 14)
However, mamba-ssm-2.1.0 does not have this problem.
The following error occurs when using
mamba_chunk_scan_combined
inmamba-ssm-2.2.0
: File "/usr/local/lib/python3.10/dist-packages/torch/_tensor.py", line 522, in backward torch.autograd.backward( File "/usr/local/lib/python3.10/dist-packages/torch/autograd/init.py", line 266, in backward Variable._execution_engine.run_backward( # Calls into the C++ engine to run the backward pass RuntimeError: function MambaChunkScanCombinedFnBackward returned an incorrect number of gradients (expected 16, got 14) However,mamba-ssm-2.1.0
does not have this problem.