Skip to content

I'm new to using pytorch #35

@MqcCheeze

Description

@MqcCheeze

I've been using Google colab with the command

!python3 -m transkun.train trained_model.pth --nProcess 1 --datasetPath "dataset" --datasetMetaFile_train "dataset/pt/train.pickle" --datasetMetaFile_val "dataset/pt/val.pickle" --modelConf "checkpoint/conf.json"

  1. I'm not sure why .pickle files are created instead of .pt files.
  2. I get an error "KeyError: 'duration' " when I run the above command

How can I create .pt files, or continue training as this is a complete roadblock for me

2026-01-19 20:40:49.392771: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:467] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered
WARNING: All log messages before absl::InitializeLog() is called are written to STDERR
E0000 00:00:1768855249.413320 6179 cuda_dnn.cc:8579] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered
E0000 00:00:1768855249.419458 6179 cuda_blas.cc:1407] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered
W0000 00:00:1768855249.434779 6179 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1768855249.434803 6179 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1768855249.434807 6179 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
W0000 00:00:1768855249.434810 6179 computation_placer.cc:177] computation placer already registered. Please check linkage and avoid linking the same target more than once.
2026-01-19 20:40:49.439353: I tensorflow/core/platform/cpu_feature_guard.cc:210] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.
To enable the following instructions: AVX2 AVX512F FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.
#0 loaded
loading dataset....
loading the annotation file...
Traceback (most recent call last):
File "", line 198, in _run_module_as_main
File "", line 88, in run_code
File "/content/Transkun/transkun/train.py", line 401, in
train(0, 1, saved_filename, runSeed, args)
File "/content/Transkun/transkun/train.py", line 84, in train
dataset = Data.DatasetMaestro(datasetPath, datasetPicklePath)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/content/Transkun/transkun/Data.py", line 468, in init
self.durations = [float(
["duration"]) for _ in self.data]
~^^^^^^^^^^^^
KeyError: 'duration'

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions