#Can't train, keeps saying saved index and thats it. Apollio

1 messages · Page 1 of 1 (latest)

balmy elk
#

Now my copy of apollio isn't working. I moved it to C drive and its the latest version.
[7:26 PM]
now this shows up.
[7:26 PM]
Sync graph is now activated! With sync graph enabled, the model undergoes a single epoch of training. Once the graphs are synchronized, training proceeds for the previously specified number of epochs.
Starting training...
Loaded pretrainedG rvc\pretraineds\pretraineds_custom\G-f032k-TITAN-Medium.pth
Process Process-1:
Traceback (most recent call last):
File "C:\ApplioV3.2.0\ApplioV3.2.0\rvc\train\train.py", line 284, in run
latest_checkpoint_path(hps.model_dir, "D.pth"), net_d, optim_d
File "C:\ApplioV3.2.0\ApplioV3.2.0\rvc\train\utils.py", line 139, in latest_checkpoint_path
x = f_list[-1]
IndexError: list index out of range

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "C:\ApplioV3.2.0\ApplioV3.2.0\env\lib\multiprocessing\process.py", line 315, in _bootstrap
self.run()
File "C:\ApplioV3.2.0\ApplioV3.2.0\env\lib\multiprocessing\process.py", line 108, in run
self._target(self._args, **self._kwargs)
File "C:\ApplioV3.2.0\ApplioV3.2.0\rvc\train\train.py", line 300, in run
torch.load(hps.pretrainG, map_location="cpu")["model"]
File "C:\ApplioV3.2.0\ApplioV3.2.0\env\lib\site-packages\torch\serialization.py", line 1014, in load
return _load(opened_zipfile,
File "C:\ApplioV3.2.0\ApplioV3.2.0\env\lib\site-packages\torch\serialization.py", line 1422, in _load
result = unpickler.load()
File "C:\ApplioV3.2.0\ApplioV3.2.0\env\lib\site-packages\torch\serialization.py", line 1392, in persistent_load
typed_storage = load_tensor(dtype, nbytes, key, _maybe_decode_ascii(location))
File "C:\ApplioV3.2.0\ApplioV3.2.0\env\lib\site-packages\torch\serialization.py", line 1357, in load_tensor
storage = zip_file.get_storage_from_record(name, numel, torch.UntypedStorage)._typed_storage()._untyped_storage
RuntimeError: PytorchStreamReader failed reading file data/7: invalid header or archive is corrupted
Traceback (most recent call last):
File "C:\ApplioV3.2.0\ApplioV3.2.0\rvc\train\train.py", line 720, in <module>
main()
File "C:\ApplioV3.2.0\ApplioV3.2.0\rvc\train\train.py", line 159, in main
edit_config(model_config_file)
File "C:\ApplioV3.2.0\ApplioV3.2.0\rvc\train\train.py", line 148, in edit_config
config_data["train"]["log_interval"] = steps
NameError: free variable 'steps' referenced before assignment in enclosing scope
Saved index file 'C:\ApplioV3.2.0\ApplioV3.2.0\logs\Snootz32k\added_IVF122_Flat_nprobe_1_v2.index'

exotic lantern
balmy elk
pallid sable
#

@grand ginkgo Do you know anything about this perhaps?

grand ginkgo
#

That error never happened to me

exotic lantern
#

@grand ginkgo @pallid sable @balmy elk I found out a small cause, the error for me pops up when I turn on the Sync graph option on Applio (On pre-zip both V 3.2 and 3.2.1, so its something along those lines, probably an assignment variable is indented one too many times or something.

#

The free variable 'steps' referenced before assignment in enclosing scope Saved index file 'C:\ApplioV3.2.0\ApplioV3.2.0\logs\Snootz32k\added_IVF122_Flat_nprobe_1_v2.index'

#

I'm unsure if the others are stemming from same issue, because my terminal only shows me the referenced before assignment error.

# 
        for filename in os.listdir(logs_path):
            match = re.match(pattern, filename)
            if match:
                steps = int(match.group(1))

        def edit_config(config_file):
            with open(config_file, "r", encoding="utf8") as json_file:
                config_data = json.load(json_file)

            config_data["train"]["log_interval"] = steps

            with open(config_file, "w", encoding="utf8") as json_file:
                json.dump(
                    config_data,
                    json_file,
                    indent=2,
                    separators=(",", ": "),
                    ensure_ascii=False,
                )

        edit_config(model_config_file)
        edit_config(rvc_config_file)
#

I imagine the error is just stemming from its not matching something, so steps never gets assigned.

#

Although steps is later used and showing in terminal as having a value: Training has been successfully completed with 1 epoch, 37 steps and 42.172 loss gen.