diffuseq-v2: TypeError: load_state_dict() takes 1 positional argument but 3 were given about diffuseq HOT 3 OPEN

Thank you for pointing out. I mixed them up during the version update. I will update this soon.

from diffuseq.

Workaround for now in case anyone else has this issue: I changed this line back to
dist_util.load_state_dict(args.model_path, map_location="cpu")
and had to remove CUDA_VISIBLE_DEVICES=2 from my command for torch to recognise my GPU (probably because I only have one GPU) and now it's working. In fact, it's working far faster than the main branch, very impressive :)

from diffuseq.

The error:

Traceback (most recent call last):
  File "/mnt/c/Users/Z/Documents/GitHub/DiffuSeq/sample_seq2seq_dpmSolver.py", line 222, in <module>
    main()
  File "/mnt/c/Users/Z/Documents/GitHub/DiffuSeq/sample_seq2seq_dpmSolver.py", line 66, in main
    dist_util.load_state_dict(args.model_path, False, "amp", map_location="cpu")
TypeError: load_state_dict() takes 1 positional argument but 3 were given

from diffuseq.