Yanay1
commented
4 months ago Both of the models we made available for download have the same output dimension which is fixed (1280).
Closed fengzhanying closed 4 months ago
Yanay1
commented
4 months ago Both of the models we made available for download have the same output dimension which is fixed (1280).
Can the parameter "--output_dim" be used? I set it to be 128 and I got this error:
Traceback (most recent call last): File "eval_single_anndata.py", line 155, in
main(args, accelerator)
File "eval_single_anndata.py", line 85, in main
processor.run_evaluation()
File "evaluate.py", line 145, in run_evaluation
run_eval(self.adata, self.name, self.pe_idx_path, self.chroms_path,
File "evaluate.py", line 203, in run_eval
model.load_state_dict(torch.load(args.model_loc, map_location="cpu"),
File "/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1671, in load_state_dict
raise RuntimeError('Error(s) in loading state_dict for {}:\n\t{}'.format(
RuntimeError: Error(s) in loading state_dict for TransformerModel:
size mismatch for decoder.1.0.weight: copying a param with shape torch.Size([1280, 1024]) from checkpoint, the shape in current model is torch.Size([128, 1024]).
size mismatch for decoder.1.0.bias: copying a param with shape torch.Size([1280]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for decoder.1.1.weight: copying a param with shape torch.Size([1280]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for decoder.1.1.bias: copying a param with shape torch.Size([1280]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for decoder.2.0.weight: copying a param with shape torch.Size([1280, 1280]) from checkpoint, the shape in current model is torch.Size([128, 128]).
size mismatch for decoder.2.0.bias: copying a param with shape torch.Size([1280]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for decoder.2.1.weight: copying a param with shape torch.Size([1280]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for decoder.2.1.bias: copying a param with shape torch.Size([1280]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for decoder.3.weight: copying a param with shape torch.Size([1280, 1280]) from checkpoint, the shape in current model is torch.Size([128, 128]).
size mismatch for decoder.3.bias: copying a param with shape torch.Size([1280]) from checkpoint, the shape in current model is torch.Size([128]).
size mismatch for binary_decoder.0.0.weight: copying a param with shape torch.Size([2048, 2560]) from checkpoint, the shape in current model is torch.Size([2048, 1408]).
What is the cause of this error?