ValueError when using PatchTSTForClassification

🤗Transformers
1

I’m attempting to adapt the PatchTST blog post (Patch Time Series Transformer in Hugging Face) to use PatchTSTForClassification. I’m getting the following error:

ValueError: The model did not return a loss from the inputs, only the following keys: prediction_logits. For reference, the inputs it received are past_values.

I know that it should be receiving past_values and target_values during training which doesn’t seem to be happening. In addition, I’m not sure how to get it to return loss? Any help is much appreciated.

2

Does this example help, PatchTST?

from transformers import PatchTSTConfig, PatchTSTForClassification

# classification task with two input channel2 and 3 classes
config = PatchTSTConfig(
    num_input_channels=2,
    num_targets=3,
    context_length=512,
    patch_length=12,
    stride=12,
    use_cls_token=True,
)
model = PatchTSTForClassification(config=config)

# during inference, one only provides past values
past_values = torch.randn(20, 512, 2)
outputs = model(past_values=past_values)
labels = outputs.prediction_logits

You need to get prediction_logits from output.