programmersilvanus/ner-xlmr

1---2license: mit3base_model: xlm-roberta-base4tags:5- silvanus6metrics:7- precision8- recall9- f110- accuracy11model-index:12- name: xlm-roberta-base-ner-silvanus13  results:14  - task:15      name: Token Classification16      type: token-classification17    dataset:18      name: id_nergrit_corpus19      type: id_nergrit_corpus20      config: ner21      split: validation22      args: ner23    metrics:24    - name: Precision25      type: precision26      value: 0.91891891891891927    - name: Recall28      type: recall29      value: 0.927272727272727230    - name: F131      type: f132      value: 0.923076923076923133    - name: Accuracy34      type: accuracy35      value: 0.985851877822921636language:37- id38- en39- es40- it41- sk42pipeline_tag: token-classification43widget:44- text: >-45    Kebakaran hutan dan lahan terus terjadi dan semakin meluas di Kota46    Palangkaraya, Kalimantan Tengah (Kalteng) pada hari Rabu, 15 Nopember 202347    20.00 WIB. Bahkan kobaran api mulai membakar pondok warga dan mendekati48    permukiman. BZK #RCTINews #SeputariNews #News #Karhutla #KebakaranHutan49    #HutanKalimantan #SILVANUS_Italian_Pilot_Testing50  example_title: Indonesia51- text: >-52    Wildfire rages for a second day in Evia destroying a Natura 2000 protected53    pine forest. - 5:51 PM Aug 14, 201954  example_title: English55- text: >-56    3 nov 2023 21:57 - Incendio forestal obliga a la evacuación de hasta 85057    personas cerca del pueblo de Montichelvo en Valencia.58  example_title: Spanish59- text: >-60    Incendi boschivi nell'est del Paese: 2 morti e oltre 50 case distrutte nello61    stato del Queensland.62  example_title: Italian63- text: >-64    Lesné požiare na Sicílii si vyžiadali dva ľudské životy a evakuáciu hotela65    http://dlvr.it/SwW3sC - 23. septembra 2023 20:5766  example_title: Slovak67---68 69<!-- This model card has been generated automatically according to the information the Trainer had access to. You70should probably proofread and complete it, then remove this comment. -->71 72# xlm-roberta-base-ner-silvanus73 74This model is a fine-tuned version of [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) on the Indonesian NER dataset.75It achieves the following results on the evaluation set:76- Loss: 0.056777- Precision: 0.918978- Recall: 0.927379- F1: 0.923180- Accuracy: 0.985981 82## Model description83 84The XLM-RoBERTa model was proposed in [Unsupervised Cross-lingual Representation Learning at Scale](https://arxiv.org/abs/1911.02116) by Alexis Conneau, Kartikay Khandelwal, Naman Goyal, Vishrav Chaudhary, Guillaume Wenzek, Francisco Guzmán, Edouard Grave, Myle Ott, Luke Zettlemoyer and Veselin Stoyanov. It is based on Facebook's RoBERTa model released in 2019. It is a large multi-lingual language model, trained on 2.5TB of filtered CommonCrawl data.85 86- **Developed by:** See [associated paper](https://arxiv.org/abs/1911.02116)87- **Model type:** Multi-lingual model88- **Language(s) (NLP) or Countries (images):** XLM-RoBERTa is a multilingual model trained on 100 different languages; see [GitHub Repo](https://github.com/facebookresearch/fairseq/tree/main/examples/xlmr) for full list; model is fine-tuned on a dataset in English89- **License:** More information needed90- **Related Models:** [RoBERTa](https://huggingface.co/roberta-base), [XLM](https://huggingface.co/docs/transformers/model_doc/xlm)91    - **Parent Model:** [XLM-RoBERTa](https://huggingface.co/xlm-roberta-base)92- **Resources for more information:** [GitHub Repo](https://github.com/facebookresearch/fairseq/tree/main/examples/xlmr)93 94## Intended uses & limitations95 96This model can be used to extract multilingual information such as location, date and time on social media (Twitter, etc.). This model is limited by an Indonesian language training data set to be tested in 4 languages (English, Spanish, Italian and Slovak) using zero-shot transfer learning techniques to extract multilingual information.97 98## Training and evaluation data99 100This model was fine-tuned on Indonesian NER datasets.101Abbreviation|Description102-|-103O|Outside of a named entity104B-LOC |Beginning of a location right after another location105I-LOC |Location106B-DAT |Beginning of a date right after another date107I-DAT |Date108B-TIM |Beginning of a time right after another time109I-TIM |Time110 111## Training procedure112 113### Training hyperparameters114 115The following hyperparameters were used during training:116- learning_rate: 2e-05117- train_batch_size: 8118- eval_batch_size: 8119- seed: 42120- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08121- lr_scheduler_type: linear122- num_epochs: 3123 124### Training results125 126| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1     | Accuracy |127|:-------------:|:-----:|:----:|:---------------:|:---------:|:------:|:------:|:--------:|128| 0.1394        | 1.0   | 827  | 0.0559          | 0.8808    | 0.9257 | 0.9027 | 0.9842   |129| 0.0468        | 2.0   | 1654 | 0.0575          | 0.9107    | 0.9190 | 0.9148 | 0.9849   |130| 0.0279        | 3.0   | 2481 | 0.0567          | 0.9189    | 0.9273 | 0.9231 | 0.9859   |131 132 133### Framework versions134 135- Transformers 4.35.0136- Pytorch 2.1.0+cu118137- Datasets 2.14.6138- Tokenizers 0.14.1