1 __author__ = 'tsungyi'
      2 
----> 3 import pycocotools._mask as _mask
      4 
      5 # Interface for manipulating masks stored in RLE format.

pycocotools/_mask.pyx in init pycocotools._mask()

ValueError: numpy.ndarray size changed, may indicate binary incompatibility. Expected 88 from C header, got 80 from PyObject

got this error for controllable talknet colab

i wonder if there's a way to improve autotune

sometimes pitch is misinterpreted from reference

I find it sometimes gets it in the wrong octave more often than not

This is the cleanest I've ever managed to get Talknet to sound without autotune or anything. It's just barely spot on!

can anyone help me with this error code?: Traceback (most recent call last): File "/content/controllable_talknet.py", line 973, in generate_audio tnmodel = TalkNetSpectModel.restore_from(talknet_path) File "/usr/local/lib/python3.8/dist-packages/nemo/core/classes/modelPT.py", line 479, in restore_from raise FileNotFoundError(f"Can't find {restore_path}") FileNotFoundError: Can't find /content/models/16ospTK5cBvEJqiHYhF2u1E9u6ritT2zJ/TalkNetSpect.nemo

this popped up after I wanted to generate an audio file

Did you tried restarting and starting from scratch?

yes

Are you using the latest notebook in #841437191073955920 ?

umm i don't know

for me thats the first time I am using such a tool, so pls excuse my inexperience

Use the one that's in #841437191073955920 and tell me if the error persists

error at step 6: IndexError Traceback (most recent call last) in 96 97 f, p = forward_extractor(target_tokens, log_probs, blank_id) ---> 98 durs = backward_extractor(f, p) 99 100 dur_key = Path(dl.dataset.collection[sample_idx].audio_file).stem in backward_extractor(f, p) 45 s, t = n - 1, m 46 while s > 0: ---> 47 durs[s - 1] += 1 48 s -= p[s, t] 49 t -= 1 IndexError: index 4982107084641271886 is out of bounds for axis 0 with size 79

i have the local talknet syntehsis tool but it wont download the model from id

any fix?

Cannot retrieve the public link of the file. You may need to change the permission to 'Anyone with the link', or have had many accesses. You may still be able to access the file from the browser:

Yes

I'll drop the fix here in a moment

thanks mate

i tried to install the latest gdown but it keps reverting back to 4.3.2 in the reqirements.txt file

here, try this

move this bat file to the Talknet folder after downloading it

and then run it

ok thx brother

you're a legend bro thankyou

np np

Sorry for the dumb Q, I'm new here. What's the difference between list1 and list2?

can anyone help me? I try step 6 but it always gets an error, can anyone explain this? --------------------------------------------------------------------------- ImportError Traceback (most recent call last) in 8 9 import json ---> 10 from nemo.collections.asr.models import EncDecCTCModel 11 asr_model = EncDecCTCModel.from_pretrained(model_name="asr_talknet_aligner").cpu().eval() 12 12 frames /usr/local/lib/python3.8/dist-packages/pytorch_lightning/metrics/utils.py in 20 from torchmetrics.utilities.data import dim_zero_mean as _dim_zero_mean 21 from torchmetrics.utilities.data import dim_zero_sum as _dim_zero_sum ---> 22 from torchmetrics.utilities.data import get_num_classes as _get_num_classes 23 from torchmetrics.utilities.data import select_topk as _select_topk 24 from torchmetrics.utilities.data import to_categorical as _to_categorical ImportError: cannot import name 'get_num_classes' from 'torchmetrics.utilities.data' (/usr/local/lib/python3.8/dist-packages/torchmetrics/utilities/data.py)