Justin Zhao comments

Results 42 comments of


                                            Justin Zhao

Impossibility to use a tokenizer with auto_transformer

@sergsb that seems reasonable to me. I think that's what @arnavgarg1 was going for in https://github.com/ludwig-ai/ludwig/pull/3632, specifically [here](https://github.com/ludwig-ai/ludwig/pull/3632/files#diff-9433924b5853fa7de42f7848db61b31247c8502f84f421d9039c3868795d1729R63-R68).

ModuleNotFoundError: No module named 'ray.air.checkpoint'

Hi @andysingal, which version of ray are you using? Ludwig requires ray v2.3.1.

ModuleNotFoundError: No module named 'ray.air.checkpoint'

Hi @andysingal, In order to use datasets hosted on kaggle, you'll need to make sure that your Kaggle credentials are set up properly and you've logged into kaggle to agree...

ModuleNotFoundError: No module named 'ray.air.checkpoint'

@andysingal The error in your previous comment is: ``` OSError: Could not find kaggle.json. Make sure it's located in /root/.kaggle. Or use the environment method. ``` Even though you are...

Question: Frequently predict same samples during training

Hi @MarselScheer, Ludwig uses the [ProgressTracker](https://github.com/ludwig-ai/ludwig/blob/master/ludwig/utils/trainer_utils.py#L84) to store metrics and other artifacts during evaluation during training. The progress tracker is provided to the `on_eval_end()` and `on_epoch_end()` callbacks. Some ideas for...

preprocess_for_training Caught exception during model preprocessing: maximum recursion depth exceeded while getting the str of an object

@Bdl-1989 Odd error. Are you able to use the tokenizer for `/mnt/d/AI/Models/llama-7b-hf` on your outside of Ludwig? ```python from transformers import AutoTokenizer import pandas as pd tokenizer = AutoTokenizer.from_pretrained("/mnt/d/AI/Models/llama-7b-hf") #...

wandb get stuck in running and finally lead to OOM

Hi @MarselScheer, Thank you for reporting your experience using wandb on Ludwig. I'm not very familiar with the wandb integration, but I have noticed that the `wandb.finish()` is missing from...

wandb get stuck in running and finally lead to OOM

Could you try patching https://github.com/ludwig-ai/ludwig/pull/3710 to see if you observe the same OOM in your first setup?

"Encounted `nan` values in tensor. Will be removed.", UserWarning

Hi @msmmpts, NaN values are more likely if you use a really high learning rate. I would recommend retrying with a learning rate that's an order magnitude smaller, like `0.0001`.

RuntimeError if just one class in multi-class output feature

Hi @Peetee06, it's great to hear about your interest to contribute to Ludwig! We have a contributing guide here: https://github.com/ludwig-ai/ludwig/blob/master/CONTRIBUTING.md Also, I recommend joining [our slack](https://join.slack.com/t/ludwig-ai/shared_invite/zt-mrxo87w6-DlX5~73T2B4v_g6jj0pJcQ)!