search

time-series-foundation-models / lag-llama

Lag-Llama: Towards Foundation Models for Probabilistic Time Series Forecasting
Apache License 2.0
1.08k stars 121 forks source link

Trying to do fine-tuning with my own dataset #78

Closed vitu1610 closed 2 weeks ago

vitu1610 commented 2 weeks ago

Hello,

When I start fine-tuning, I receive the following error:

File [.\venv\lib\site-packages\gluonts\torch\model\estimator.py:205](./venv/lib/site-packages/gluonts/torch/model/estimator.py:205), in PyTorchLightningEstimator.train_model(self, training_data, validation_data, from_predictor, shuffle_buffer_length, cache_data, ckpt_path, **kwargs)
    202 trainer_kwargs = {**self.trainer_kwargs, "callbacks": callbacks}
    203 trainer = pl.Trainer(**trainer_kwargs)
--> 205 trainer.fit(
    206     model=training_network,
...
    136 if self.allow_missing:
    137     return {f: data[f] for f in self.input_fields if f in data}
--> 138 return {f: data[f] for f in self.input_fields}

KeyError: 'data_id'

Could you please clarify the origin of the ‘data_id’ field in my dataset, as it seems to be present despite my dataset not having any field with that name?

Very best

vitu1610 commented 2 weeks ago

While conducting a local evaluation of the model, it appears that the dependencies and scripts listed on the GitHub repository at https://github.com/time-series-foundation-models/lag-llama/ were not up-to-date in my PC. Upon performing the corresponding updates, the problem was resolved.