Missing functions in directoty “models”

[firethatpotato]

Thanks for your great work but I have some troubles.

When I run "train_control_var_hpu.py" to validate using command in "Inference" in README, some errors indicating that there are some missing functions or classes in models.

Traceback (most recent call last): File "/home/N3_3090U2/anaconda3/envs/controlvar/lib/python3.9/site-packages/torch/multiprocessing/spawn.py", line 76, in _wrap fn(i, *args) File "/home/N3_3090U2/Desktop/controlVAR/ControlVAR-main/train_control_var_gpu.py", line 595, in process var = build_control_var(vae=vqvae, depth=args.depth, patch_nums=args.v_patch_nums, mask_type=args.mask_type, File "/home/N3_3090U2/Desktop/controlVAR/ControlVAR-main/models/__init__.py", line 37, in build_control_var return MaskVAR( **NameError: name 'MaskVAR' is not defined** File "/home/N3_3090U2/Desktop/controlVAR/ControlVAR-main/train_control_var_gpu.py", line 697, in run mp.spawn(process, File "/home/N3_3090U2/Desktop/controlVAR/ControlVAR-main/train_control_var_gpu.py", line 704, in <module> run(process, args.gpus, args) torch.multiprocessing.spawn.ProcessRaisedException:

'VisualProgressAutoreg' cannot be found, either.

[lxa9867]

Hi, please remove the VisualProgressAutoreg class as it was not released. It was not used in the pipeline.

[firethatpotato]

Hi, please remove the VisualProgressAutoreg class as it was not released. It was not used in the pipeline.

Thank you! How about the function "MaskVAR"？ This function is the return of "build_control_var", which is used to initiate var:

train_control_var_hpu.py:

var = build_control_var(vae=vqvae, depth=args.depth, patch_nums=args.v_patch_nums, mask_type=args.mask_type,
                         cond_drop_rate=1.1 if args.uncond else 0.1, bidirectional=args.bidirectional,
                         separate_decoding=args.separate_decoding, separator=args.separator, type_pos=args.type_pos,
                         indep=args.indep, multi_cond=args.multi_cond)

models init.py:

def build_control_var(
    vae: VQVAE, depth: int,
    patch_nums=(1, 2, 3, 4, 5, 6, 8, 10, 13, 16),   # 10 steps by default
    aln=1, aln_gamma_init=1e-3, shared_aln=False, layer_scale=-1,
    tau=4, cos_attn=False,
    flash_if_available=True, fused_if_available=True,
    mask_type='replace', cond_drop_rate=0.1, bidirectional=False, separate_decoding=False, separator=False,
    type_pos=False, indep=False, multi_cond=False,
):
    if mask_type == 'replace':
        mask_factor = 1
    elif mask_type == 'interleave_append':
        mask_factor = 2
    else:
        raise NotImplementedError
    return **MaskVAR**(
        vae_local=vae, patch_nums=patch_nums,
        depth=depth, embed_dim=depth*64, num_heads=depth, drop_path_rate=0.1 * depth/24,
        aln=aln, aln_gamma_init=aln_gamma_init, shared_aln=shared_aln, layer_scale=layer_scale,
        tau=tau, cos_attn=cos_attn, cond_drop_rate=cond_drop_rate,
        flash_if_available=flash_if_available, fused_if_available=fused_if_available, mask_factor=mask_factor,
        bidirectional=bidirectional, separate_decoding=separate_decoding, separator=separator, type_pos=type_pos,
        indep=indep, multi_cond=multi_cond,
    )

"MaskVAR" cannot be found.

[lxa9867]

Sorry about that. This should be a typo, please replace it as ControlVAR

[firethatpotato]

Sorry about that. This should be a typo, please replace it as ControlVAR

It works. Thanks for your reply!

[lxa9867]

Cool! Closing this issue.