---
tags:
- time-series
- temporal-point-processes
- hawkes-processes
- scientific-ml
license: mit
library_name: transformers
---

# FIM-PP Model Card

`FIM-PP` is the Foundation Inference Model for marked temporal point processes.
It infers conditional intensity functions from a context set of event sequences and
supports zero-shot use as well as downstream fine-tuning.

## Loading

Install the `fim` package first, then load the model with Transformers:

```python
from transformers import AutoModel

model = AutoModel.from_pretrained("FIM4Science/FIM-PP", trust_remote_code=True)
model.eval()
```

## Notes

- The released checkpoint is configured for up to 22 event marks.
- The model expects Hawkes-style context and inference tensors as described in the
  OpenFIM point-process tutorial.
- If needed, the lower-level fallback remains available through
  `fim.models.hawkes.FIMHawkes.load_model(...)`.

## Reference

If you use this model, please cite:

```bibtex
@inproceedings{fim_pp,
  title={In-Context Learning of Temporal Point Processes with Foundation Inference Models},
  author={David Berghaus and Patrick Seifner and Kostadin Cvejoski and Cesar Ojeda and Ramses J. Sanchez},
  booktitle={The Fourteenth International Conference on Learning Representations},
  year={2026},
  url={https://openreview.net/forum?id=h9HwUAODFP}
}
```