File size: 5,065 Bytes
3f07d8a 068adce 3f07d8a 068adce 3f07d8a 068adce 3f07d8a |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 |
---
language: fr
license: mit
tags:
- flair
- token-classification
- sequence-tagger-model
base_model: hmbyt5-preliminary/byt5-small-historic-multilingual-span20-flax
inference: false
widget:
- text: 'Parmi les remèdes recommandés par la Société , il faut mentionner celui que
M . Schatzmann , de Lausanne , a proposé :'
---
# Fine-tuned Flair Model on LeTemps French NER Dataset (HIPE-2022)
This Flair model was fine-tuned on the
[LeTemps French](https://github.com/hipe-eval/HIPE-2022-data/blob/main/documentation/README-letemps.md)
NER Dataset using hmByT5 as backbone LM.
The LeTemps dataset consists of NE-annotated historical French newspaper articles from mid-19C to mid 20C.
The following NEs were annotated: `loc`, `org` and `pers`.
# ⚠️ Inference Widget ⚠️
Fine-Tuning ByT5 models in Flair is currently done by implementing an own [`ByT5Embedding`][0] class.
This class needs to be present when running the model with Flair.
Thus, the inference widget is not working with hmByT5 at the moment on the Model Hub and is currently disabled.
This should be fixed in future, when ByT5 fine-tuning is supported in Flair directly.
[0]: https://github.com/stefan-it/hmBench/blob/main/byt5_embeddings.py
# Results
We performed a hyper-parameter search over the following parameters with 5 different seeds per configuration:
* Batch Sizes: `[8, 4]`
* Learning Rates: `[0.00015, 0.00016]`
And report micro F1-score on development set:
| Configuration | Run 1 | Run 2 | Run 3 | Run 4 | Run 5 | Avg. |
|-------------------|--------------|--------------|--------------|--------------|--------------|--------------|
| bs8-e10-lr0.00016 | [0.6553][1] | [0.6628][2] | [0.6699][3] | [0.6524][4] | [0.6542][5] | 65.89 ± 0.65 |
| bs4-e10-lr0.00015 | [0.6603][6] | [0.6651][7] | [0.654][8] | [0.6575][9] | [0.6575][10] | 65.89 ± 0.37 |
| bs4-e10-lr0.00016 | [0.6423][11] | [0.6595][12] | [0.6625][13] | [0.6657][14] | [0.6538][15] | 65.68 ± 0.82 |
| bs8-e10-lr0.00015 | [0.6502][16] | [0.6541][17] | [0.6607][18] | [0.6496][19] | [0.6629][20] | 65.55 ± 0.54 |
[1]: https://hf.co/stefan-it/hmbench-letemps-fr-hmbyt5-bs8-wsFalse-e10-lr0.00016-poolingfirst-layers-1-crfFalse-1
[2]: https://hf.co/stefan-it/hmbench-letemps-fr-hmbyt5-bs8-wsFalse-e10-lr0.00016-poolingfirst-layers-1-crfFalse-2
[3]: https://hf.co/stefan-it/hmbench-letemps-fr-hmbyt5-bs8-wsFalse-e10-lr0.00016-poolingfirst-layers-1-crfFalse-3
[4]: https://hf.co/stefan-it/hmbench-letemps-fr-hmbyt5-bs8-wsFalse-e10-lr0.00016-poolingfirst-layers-1-crfFalse-4
[5]: https://hf.co/stefan-it/hmbench-letemps-fr-hmbyt5-bs8-wsFalse-e10-lr0.00016-poolingfirst-layers-1-crfFalse-5
[6]: https://hf.co/stefan-it/hmbench-letemps-fr-hmbyt5-bs4-wsFalse-e10-lr0.00015-poolingfirst-layers-1-crfFalse-1
[7]: https://hf.co/stefan-it/hmbench-letemps-fr-hmbyt5-bs4-wsFalse-e10-lr0.00015-poolingfirst-layers-1-crfFalse-2
[8]: https://hf.co/stefan-it/hmbench-letemps-fr-hmbyt5-bs4-wsFalse-e10-lr0.00015-poolingfirst-layers-1-crfFalse-3
[9]: https://hf.co/stefan-it/hmbench-letemps-fr-hmbyt5-bs4-wsFalse-e10-lr0.00015-poolingfirst-layers-1-crfFalse-4
[10]: https://hf.co/stefan-it/hmbench-letemps-fr-hmbyt5-bs4-wsFalse-e10-lr0.00015-poolingfirst-layers-1-crfFalse-5
[11]: https://hf.co/stefan-it/hmbench-letemps-fr-hmbyt5-bs4-wsFalse-e10-lr0.00016-poolingfirst-layers-1-crfFalse-1
[12]: https://hf.co/stefan-it/hmbench-letemps-fr-hmbyt5-bs4-wsFalse-e10-lr0.00016-poolingfirst-layers-1-crfFalse-2
[13]: https://hf.co/stefan-it/hmbench-letemps-fr-hmbyt5-bs4-wsFalse-e10-lr0.00016-poolingfirst-layers-1-crfFalse-3
[14]: https://hf.co/stefan-it/hmbench-letemps-fr-hmbyt5-bs4-wsFalse-e10-lr0.00016-poolingfirst-layers-1-crfFalse-4
[15]: https://hf.co/stefan-it/hmbench-letemps-fr-hmbyt5-bs4-wsFalse-e10-lr0.00016-poolingfirst-layers-1-crfFalse-5
[16]: https://hf.co/stefan-it/hmbench-letemps-fr-hmbyt5-bs8-wsFalse-e10-lr0.00015-poolingfirst-layers-1-crfFalse-1
[17]: https://hf.co/stefan-it/hmbench-letemps-fr-hmbyt5-bs8-wsFalse-e10-lr0.00015-poolingfirst-layers-1-crfFalse-2
[18]: https://hf.co/stefan-it/hmbench-letemps-fr-hmbyt5-bs8-wsFalse-e10-lr0.00015-poolingfirst-layers-1-crfFalse-3
[19]: https://hf.co/stefan-it/hmbench-letemps-fr-hmbyt5-bs8-wsFalse-e10-lr0.00015-poolingfirst-layers-1-crfFalse-4
[20]: https://hf.co/stefan-it/hmbench-letemps-fr-hmbyt5-bs8-wsFalse-e10-lr0.00015-poolingfirst-layers-1-crfFalse-5
The [training log](training.log) and TensorBoard logs (only for hmByT5 and hmTEAMS based models) are also uploaded to the model hub.
More information about fine-tuning can be found [here](https://github.com/stefan-it/hmBench).
# Acknowledgements
We thank [Luisa März](https://github.com/LuisaMaerz), [Katharina Schmid](https://github.com/schmika) and
[Erion Çano](https://github.com/erionc) for their fruitful discussions about Historic Language Models.
Research supported with Cloud TPUs from Google's [TPU Research Cloud](https://sites.research.google/trc/about/) (TRC).
Many Thanks for providing access to the TPUs ❤️
|