NER: ModernBERT-base on MultiNERD (EN+DE), F1=0.9063
Browse files
README.md
ADDED
|
@@ -0,0 +1,78 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: transformers
|
| 3 |
+
license: apache-2.0
|
| 4 |
+
base_model: answerdotai/ModernBERT-base
|
| 5 |
+
tags:
|
| 6 |
+
- ner
|
| 7 |
+
- knowledge-platform
|
| 8 |
+
- modernbert
|
| 9 |
+
- multilingual
|
| 10 |
+
- patents
|
| 11 |
+
- scientific-papers
|
| 12 |
+
- generated_from_trainer
|
| 13 |
+
metrics:
|
| 14 |
+
- precision
|
| 15 |
+
- recall
|
| 16 |
+
- f1
|
| 17 |
+
- accuracy
|
| 18 |
+
model-index:
|
| 19 |
+
- name: knowledge-platform-ner
|
| 20 |
+
results: []
|
| 21 |
+
---
|
| 22 |
+
|
| 23 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 24 |
+
should probably proofread and complete it, then remove this comment. -->
|
| 25 |
+
|
| 26 |
+
# knowledge-platform-ner
|
| 27 |
+
|
| 28 |
+
This model is a fine-tuned version of [answerdotai/ModernBERT-base](https://huggingface.co/answerdotai/ModernBERT-base) on an unknown dataset.
|
| 29 |
+
It achieves the following results on the evaluation set:
|
| 30 |
+
- Loss: 0.0606
|
| 31 |
+
- Precision: 0.8951
|
| 32 |
+
- Recall: 0.9178
|
| 33 |
+
- F1: 0.9063
|
| 34 |
+
- Accuracy: 0.9811
|
| 35 |
+
|
| 36 |
+
## Model description
|
| 37 |
+
|
| 38 |
+
More information needed
|
| 39 |
+
|
| 40 |
+
## Intended uses & limitations
|
| 41 |
+
|
| 42 |
+
More information needed
|
| 43 |
+
|
| 44 |
+
## Training and evaluation data
|
| 45 |
+
|
| 46 |
+
More information needed
|
| 47 |
+
|
| 48 |
+
## Training procedure
|
| 49 |
+
|
| 50 |
+
### Training hyperparameters
|
| 51 |
+
|
| 52 |
+
The following hyperparameters were used during training:
|
| 53 |
+
- learning_rate: 2e-05
|
| 54 |
+
- train_batch_size: 16
|
| 55 |
+
- eval_batch_size: 32
|
| 56 |
+
- seed: 42
|
| 57 |
+
- gradient_accumulation_steps: 2
|
| 58 |
+
- total_train_batch_size: 32
|
| 59 |
+
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
|
| 60 |
+
- lr_scheduler_type: cosine
|
| 61 |
+
- lr_scheduler_warmup_steps: 0.1
|
| 62 |
+
- num_epochs: 3
|
| 63 |
+
|
| 64 |
+
### Training results
|
| 65 |
+
|
| 66 |
+
| Training Loss | Epoch | Step | Validation Loss | Precision | Recall | F1 | Accuracy |
|
| 67 |
+
|:-------------:|:-----:|:-----:|:---------------:|:---------:|:------:|:------:|:--------:|
|
| 68 |
+
| 0.1276 | 1.0 | 8020 | 0.0766 | 0.8595 | 0.8361 | 0.8476 | 0.9728 |
|
| 69 |
+
| 0.0927 | 2.0 | 16040 | 0.0623 | 0.8659 | 0.8923 | 0.8789 | 0.9777 |
|
| 70 |
+
| 0.0422 | 3.0 | 24060 | 0.0694 | 0.8707 | 0.8949 | 0.8827 | 0.9778 |
|
| 71 |
+
|
| 72 |
+
|
| 73 |
+
### Framework versions
|
| 74 |
+
|
| 75 |
+
- Transformers 5.6.0
|
| 76 |
+
- Pytorch 2.5.1+cu121
|
| 77 |
+
- Datasets 4.8.4
|
| 78 |
+
- Tokenizers 0.22.2
|