Error p.attn_bias_ptr is not correctly aligned when testing

[poedator]

tried to test the code with run_A100.sh script but got this error:

$:~/sequoia/tests$ bash run_A100.sh
...
Traceback (most recent call last):
  File "/extra_disk_1/optimus/sequoia/tests/testbed.py", line 293, in <module>
    simulation_fast(target_model=target_model, draft_model=draft_model, dataloader=dataloader, T=args.T, top_p=args.P,
  File "/extra_disk_1/optimus/sequoia/tests/testbed.py", line 68, in simulation_fast
    spectree = SpecTree(prefix=input_ids.squeeze(0), device='cuda:0', temperature=T,
  File "/extra_disk_1/optimus/sequoia/tests/../Tree/SpecTree.py", line 68, in __init__
    draft_model_outputs = self.draft_model_engine.inference(input_ids=self.tokens[:self.num_nodes].unsqueeze(0), 
  File "/home/optimus/conda/envs/py9/lib/python3.9/site-packages/torch/utils/_contextlib.py", line 115, in decorate_context
    return func(*args, **kwargs)
  File "/extra_disk_1/optimus/sequoia/tests/../Engine/Engine.py", line 242, in inference
    return self.engine.model_run(input_ids=input_ids, storage_ids=storage_ids,
  File "/home/optimus/conda/envs/py9/lib/python3.9/site-packages/torch/utils/_contextlib.py", line 115, in decorate_context
    return func(*args, **kwargs)
  File "/extra_disk_1/optimus/sequoia/tests/../Engine/Engine.py", line 38, in model_run
    logits = self.model(input_ids=input_ids, 
  File "/home/optimus/conda/envs/py9/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl
    return self._call_impl(*args, **kwargs)
  File "/home/optimus/conda/envs/py9/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl
    return forward_call(*args, **kwargs)
  File "/extra_disk_1/optimus/sequoia/tests/../Engine/Llama_model.py", line 201, in forward
    outputs = self.model(
  File "/home/optimus/conda/envs/py9/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl
    return self._call_impl(*args, **kwargs)
  File "/home/optimus/conda/envs/py9/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl
    return forward_call(*args, **kwargs)
  File "/extra_disk_1/optimus/sequoia/tests/../Engine/Llama_model.py", line 59, in forward
    layer_outputs = decoder_layer(
  File "/home/optimus/conda/envs/py9/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl
    return self._call_impl(*args, **kwargs)
  File "/home/optimus/conda/envs/py9/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl
    return forward_call(*args, **kwargs)
  File "/extra_disk_1/optimus/sequoia/tests/../Engine/Llama_modules.py", line 334, in forward
    hidden_states = self.self_attn(
  File "/home/optimus/conda/envs/py9/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl
    return self._call_impl(*args, **kwargs)
  File "/home/optimus/conda/envs/py9/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl
    return forward_call(*args, **kwargs)
  File "/extra_disk_1/optimus/sequoia/tests/../Engine/Llama_modules.py", line 127, in forward
    attn_output = torch.nn.functional.scaled_dot_product_attention(
RuntimeError: p.attn_bias_ptr is not correctly aligned

my lib versions:

$:~/sequoia/tests$ pip list | grep -e transformers -e torch -e accelerate
accelerate                        0.26.1
torch                             2.1.0
torchaudio                        0.13.1
torchvision                       0.14.1
transformers                      4.37.2

[dreaming-panda]

Can you switch to torch 2.1.2 and transformers 4.36.2?