sizhelee
commented
1 week ago - You can find the corresponding data(s) about the original POP909 dataset and the pretrained models in the Polyffusion repository.
- For the ModuleNotFoundError, we have updated the 'inference_sdf.py' file. Please refer to the latest version.
- For the error about the dataloader, we re-split the dataset in our paper (which is not the same as POP909). The new split file is updated.
about data : data/musicalion_solo_piano_4_bin_pnt"
about pre_model: pretrained/a2s/a2s-stage3a.pt"
about pre_model: pretrained/pnotree_20/train_20-last-model.pt" I can't find the above data,
when i run train python diffbgm/main.py --model ldm_chd8bar --output_dir ./result/symmv (diff-bgm) D:\2023project\project\0win_os\13Audio\Diff-BGM>python diffbgm/main.py --model ldm_chd8bar --output_dir ./result/symmv load train valid set with: {'use_track': [0, 1, 2]} D:\2023project\project\0win_os\13Audio\Diff-BGM\diffbgm\data\dataset.py:99: UserWarning: To copy construct from a tensor, it is recommended to use sourceTensor.clone().detach() or sourceTensor.clone().detach().requiresgrad(True), rather than torch.tensor(sourceTensor). self.visual = torch.tensor(self.visual[idx_ls, :]) Dataloader ready: batch_size=32, num_workers=4, pin_memory=True, {'use_track': [0, 1, 2]} Total parameters: 41476098 Creating new log folder as ./result/symmv/07-24_141153 { "attention_levels": [ 2, 3 ], "batch_size": 32, "channel_multipliers": [ 1, 2, 4, 4 ], "channels": 64, "chd_hidden_dim": 512, "chd_input_dim": 36, "chd_n_step": 32, "chd_z_dim": 512, "chd_z_input_dim": 512, "cond_type": "visual", "d_cond": 512, "fp16": true, "img_h": 128, "img_w": 128, "in_channels": 2, "latent_scaling_factor": 0.18215, "learning_rate": 5e-05, "linear_end": 0.012, "linear_start": 0.00085, "max_epoch": 100, "max_grad_norm": 10, "n_heads": 4, "n_res_blocks": 2, "n_steps": 1000, "num_workers": 4, "out_channels": 2, "pin_memory": true, "tf_layers": 1, "use_enc": true } Epoch 0: 0%| | 0/655 [00:11<?, ?it/s] Traceback (most recent call last): File "D:\2023project\project\0win_os\13Audio\Diff-BGM\diffbgm\main.py", line 71, in
config.train()
File "D:\2023project\project\0win_os\13Audio\Diff-BGM\diffbgm\train__init.py", line 49, in train
learner.train(max_epoch=self.params.max_epoch)
File "D:\2023project\project\0win_os\13Audio\Diff-BGM\diffbgm\learner.py", line 133, in train
for _step, batch in enumerate(
File "H:\Anaconda3\envs\diff-bgm\lib\site-packages\tqdm\std.py", line 1195, in iter
for obj in iterable:
File "H:\Anaconda3\envs\diff-bgm\lib\site-packages\torch\utils\data\dataloader.py", line 441, in iter__
return self._get_iterator()
File "H:\Anaconda3\envs\diff-bgm\lib\site-packages\torch\utils\data\dataloader.py", line 388, in _get_iterator
return _MultiProcessingDataLoaderIter(self)
File "H:\Anaconda3\envs\diff-bgm\lib\site-packages\torch\utils\data\dataloader.py", line 1042, in init
w.start()
File "H:\Anaconda3\envs\diff-bgm\lib\multiprocessing\process.py", line 121, in start
self._popen = self._Popen(self)
File "H:\Anaconda3\envs\diff-bgm\lib\multiprocessing\context.py", line 224, in _Popen
return _default_context.get_context().Process._Popen(process_obj)
File "H:\Anaconda3\envs\diff-bgm\lib\multiprocessing\context.py", line 327, in _Popen
return Popen(process_obj)
File "H:\Anaconda3\envs\diff-bgm\lib\multiprocessing\popen_spawn_win32.py", line 93, in init
reduction.dump(process_obj, to_child)
File "H:\Anaconda3\envs\diff-bgm\lib\multiprocessing\reduction.py", line 60, in dump
ForkingPickler(file, protocol).dump(obj)
AttributeError: Can't pickle local object 'get_train_val_dataloaders..'
Traceback (most recent call last):
File "", line 1, in
File "H:\Anaconda3\envs\diff-bgm\lib\multiprocessing\spawn.py", line 116, in spawn_main
exitcode = _main(fd, parent_sentinel)
File "H:\Anaconda3\envs\diff-bgm\lib\multiprocessing\spawn.py", line 126, in _main
self = reduction.pickle.load(from_parent)
EOFError: Ran out of input
when i infer use three model:
Please point me to any errors or code and data missing, thank you