pyannote-pipeline

pyannote-pipeline copied to clipboard

Error while optimizing hyperparameters of clustering pipeline

2

Hello! I wanted to try out the hyperparameter otimization as described in the pyannote-pipeline ReadMe. When calling `optimizer.tune()` I got the following error: ``` File "/usr/local/lib/python3.10/dist-packages/pyannote/pipeline/optimizer.py", line 269, in objective...

HannaLangenberg

I followed the notebook "Adapting pyannote.audio 2.1 pretrained speaker diarization pipeline to your own data" to adapt the Pyannote.audio 2.1 Speaker Diarization Pipeline to my own data. I am using...

nalli-hu

hbredin

Update to latest optuna

1

https://github.com/optuna/optuna/releases/tag/v3.0.0

hbredin

Simplify CLI

1

The whole `{root_dir}/train/{protocol}.{subset}/apply/{date}/{protocol}.{subset}.rttm` directory scheme is (a bit) cumbersome. Would be nice to simplify the whole thing: ## Training ```bash pyannote-pipeline train {/path/to/config.yml} {protocol} ``` would creates two files in...

hbredin

https://scikit-learn.org/dev/auto_examples/miscellaneous/plot_pipeline_display.html#sphx-glr-auto-examples-miscellaneous-plot-pipeline-display-py

hbredin

Dump the output of the pipeline after each better trial, here: https://github.com/pyannote/pyannote-pipeline/blob/afed9a67618cd5ebb6f19a441568a01d3c89fb0c/pyannote/pipeline/experiment.py#L296-L301

hbredin

Optimization visualization

2

Would be nice to add support for parallel axes plot: https://twitter.com/bokeh/status/1233782933477109760

hbredin

Multi-objective optimization is available since Optuna 2.4 Could be useful for optimization diarization pipelines for both purity and coverage, or detection pipelines for both missed detections and false alarms.

hbredin

- [x] deploy to PyPI - [x] test - [ ] documentation

hbredin