search

Nathan-Roll1 / PSST

Prosodic Speech Segmentation with Transformers
MIT License
23 stars 5 forks source link

Error while running the colab notebook #2

Closed staru09 closed 8 months ago

staru09 commented 8 months ago

ImportError: cannot import name 'is_torch_less_than_1_11' from 'transformers.pytorch_utils' (/usr/local/lib/python3.10/dist-packages/transformers/pytorch_utils.py)

The above exception was the direct cause of the following exception:

RuntimeError Traceback (most recent call last) RuntimeError: Failed to import optimum.exporters.onnx.main because of the following error (look up to see its traceback): cannot import name 'is_torch_less_than_1_11' from 'transformers.pytorch_utils' (/usr/local/lib/python3.10/dist-packages/transformers/pytorch_utils.py)

The above exception was the direct cause of the following exception:

RuntimeError Traceback (most recent call last) /usr/local/lib/python3.10/dist-packages/transformers/utils/import_utils.py in _get_module(self, module_name) 1390 return importlib.import_module("." + module_name, self.name) 1391 except Exception as e: -> 1392 raise RuntimeError( 1393 f"Failed to import {self.name}.{module_name} because of the following error (look up to see its" 1394 f" traceback):\n{e}"

Nathan-Roll1 commented 8 months ago

Thanks for bringing this to my attention -- I've updated (1) the path to the non-quantizd tutorial, and (2) forced the installation of an older version of the transformers library in the quantized notebook. Let me know if you are still encountering any issues.

staru09 commented 8 months ago

Thanks for fixing, It's working now. Just a quick question are these (!!!!!) some sort of delimiters, I am kinda new to intonation. Also any other resource to know about intonation units or detecting them in a sentence will be helpful.

Nathan-Roll1 commented 8 months ago

Yes, the '!!!!!' token represents prosodic (~intonation unit/~intonational phrase) boundaries , which are somewhat akin to what you might think of as phrases. I would recommend following some of the references in our paper (i.e. Du Bois et al. 1992) for a much more comprehensive definition. We've also included a selection of past approaches to the same task as well. Good luck!

staru09 commented 8 months ago

Thanks a lot

staru09 commented 8 months ago

Also, this is a sample text file from that dataset ALAN: ... (H) I got a story to tell you, 2.572 3.820 it's a shaggy dog story, 3.820 6.645 but as long as we’re talking about that recorder and six-hundred dollars, 6.645 8.378 (H)= I’ll tell you the story, 8.378 10.178 (H)= oh gosh, 10.178 10.678 uh (Hx), 10.678 12.335 Rae and I and Sue and Buddy, 12.335 14.727 ... took a trip, 14.727 17.012 ... to Mexico City, so these commas are showing some sort of intonational boundary?