Could not sync with existing tensorboard event files

wandb --version && python --version && uname

• Weights and Biases version: 0.8.36
• Python version: Python 3.7.6
• Operating System: MSYS_NT-10.0-18362

Description

Describe what you were trying to get done. Tell us what happened, what went wrong, and what you expected to happen.

I am using Anaconda to manage my Python environment and I installed required dependencies for wandb and also pytorch >1.5.0 and tensorboard = 2.2.1. I followed the instructions to log in and sync using the command line interface, but first it requires me to install tensorflow (according to the documentation it should not be a hard requirement) which I did, but then when I tried to sync, it returns an internal error.

What I Did

wandb sync log_dir

wandb: Syncing log_dir to:
wandb: revived-sun-1 https://app.wandb.ai/username/project_name/runs/2cswv2su
wandb: Found tfevents file, converting...
2020-05-21 00:05:17.302868: W tensorflow/stream_executor/platform/default/dso_loader.cc:55] Could not load dynamic library 'cudart64_101.dll'; dlerror: cudart64_101.dll not found
2020-05-21 00:05:17.310526: I tensorflow/stream_executor/cuda/cudart_stub.cc:29] Ignore above cudart dlerror if you do not have a GPU set up on your machine.
WARNING:tensorflow:From h:\anaconda3\envs\conda_env_name\lib\site-packages\tensorflow\python\summary\summary_iterator.py:68: tf_record_iterator (from tensorflow.python.lib.io.tf_record) is deprecated and will be removed in a future version.
Instructions for updating:
Use eager execution and:
`tf.data.TFRecordDataset(path)`
Traceback (most recent call last):
  File "h:\anaconda3\envs\conda_env_name\lib\runpy.py", line 193, in _run_module_as_main
    "__main__", mod_spec)
  File "h:\anaconda3\envs\conda_env_name\lib\runpy.py", line 85, in _run_code
    exec(code, run_globals)
  File "H:\Anaconda3\envs\conda_env_name\Scripts\wandb.exe\__main__.py", line 7, in <module>
  File "h:\anaconda3\envs\conda_env_name\lib\site-packages\click\core.py", line 829, in __call__
    return self.main(*args, **kwargs)
  File "h:\anaconda3\envs\conda_env_name\lib\site-packages\click\core.py", line 782, in main
    rv = self.invoke(ctx)
  File "h:\anaconda3\envs\conda_env_name\lib\site-packages\click\core.py", line 1259, in invoke
    return _process_result(sub_ctx.command.invoke(sub_ctx))
  File "h:\anaconda3\envs\conda_env_name\lib\site-packages\click\core.py", line 1066, in invoke
    return ctx.invoke(self.callback, **ctx.params)
  File "h:\anaconda3\envs\conda_env_name\lib\site-packages\click\core.py", line 610, in invoke
    return callback(*args, **kwargs)
  File "h:\anaconda3\envs\conda_env_name\lib\site-packages\click\decorators.py", line 21, in new_func
    return f(get_current_context(), *args, **kwargs)
  File "h:\anaconda3\envs\conda_env_name\lib\site-packages\wandb\cli.py", line 154, in wrapper
    return func(*args, **kwargs)
  File "h:\anaconda3\envs\conda_env_name\lib\site-packages\wandb\cli.py", line 447, in sync
    path, run_id=id, project=project, entity=entity, ignore_globs=globs)
  File "h:\anaconda3\envs\conda_env_name\lib\site-packages\wandb\wandb_run.py", line 339, in from_directory
    summary.update(wbtf.stream_tfevents(path, file_api, run, namespace=namespace))
  File "h:\anaconda3\envs\conda_env_name\lib\site-packages\wandb\tensorflow\__init__.py", line 55, in stream_tfevents
    parsed = tf_summary_to_dict(summary, namespace=namespace)
  File "h:\anaconda3\envs\conda_env_name\lib\site-packages\wandb\tensorboard\__init__.py", line 339, in tf_summary_to_dict
    if not wandb.run.config.get(key):
AttributeError: 'NoneType' object has no attribute 'config'

[cvphelps]

Hi there, thanks for letting us know about this issue. We’d like to fix syncing problems in the new version of our client, and I’ll check with my colleague @raubitsj to see if this case is supported.

We’re working on a new beta of the Weights & Biases client that improves support for environments like Windows. Would you be interested in giving it a try?

1. Join the Slack forum and go to the wandb-cling channel.
2. Install CLI-NG following directions from the GitHub repo.
3. Try tracking a model with the new client, no code changes necessary.

Message us in the wandb-cling channel— we’d love to hear any questions or feedback.

[aj7tesh]

"“Could not sync with existing tensorboard event files wandb/client#1051"” https://github.com/wandb/client-ng/issues/23#issuecomment-651258932 Above feature is not yet there,right ?