|
--- |
|
license: mit |
|
base_model: BAAI/bge-base-en-v1.5 |
|
tags: |
|
- generated_from_trainer |
|
model-index: |
|
- name: CONDITIONAL-multilabel-bge |
|
results: [] |
|
datasets: |
|
- GIZ/policy_classification |
|
library_name: transformers |
|
pipeline_tag: text-classification |
|
|
|
co2_eq_emissions: |
|
emissions: 28.4522411264774 |
|
source: codecarbon |
|
training_type: fine-tuning |
|
on_cloud: true |
|
cpu_model: Intel(R) Xeon(R) CPU @ 2.00GHz |
|
ram_total_size: 12.6747894287109 |
|
hours_used: 0.702 |
|
hardware_used: 1 x Tesla T4 |
|
--- |
|
|
|
<!-- This model card has been generated automatically according to the information the Trainer had access to. You |
|
should probably proofread and complete it, then remove this comment. --> |
|
|
|
# CONDITIONAL-multilabel-bge |
|
|
|
This model is a fine-tuned version of [BAAI/bge-base-en-v1.5](https://huggingface.co./BAAI/bge-base-en-v1.5) on the [Policy-Classification](https://huggingface.co./datasets/GIZ/policy_classification) dataset. |
|
It achieves the following results on the evaluation set: |
|
- Loss: 0.5295 |
|
- Precision-micro: 0.5138 |
|
- Precision-samples: 0.1866 |
|
- Precision-weighted: 0.5169 |
|
- Recall-micro: 0.7378 |
|
- Recall-samples: 0.1874 |
|
- Recall-weighted: 0.7378 |
|
- F1-micro: 0.6058 |
|
- F1-samples: 0.1852 |
|
- F1-weighted: 0.6065 |
|
|
|
## Model description |
|
|
|
The purpose of this model is to predict multiple labels simultaneously from a given input data. Specifically, the model will predict 2 labels - |
|
ConditionalLabel, UnconditionalLabel - that are relevant to a particular task or application |
|
- **Conditional**: In context of climate policy documents if certain Target/Action/Plan/Policy commitment is being made conditionally. |
|
- **Unconditional**: In context of climate policy documents if certain Target/Action/Plan/Policy commitment is being made unconditionally. |
|
|
|
## Intended uses & limitations |
|
|
|
The dataset sometimes does not include the sub-heading/heading which indicates that the paragraph belongs to Conditional/Unconditional category. |
|
But has been copied from the relevant document from those sub-headings. This makes the assessment of Conditonality very difficult. Annotator when given only the paragraph without |
|
the full long context had a difficulty in assessing the conditionality of commitments being made in paragraph. |
|
|
|
## Training and evaluation data |
|
|
|
- Training Dataset: 5901 |
|
| Class | Positive Count of Class| |
|
|:-------------|:--------| |
|
| ConditionalLabel | 1986 | |
|
| UnconditionalLabel | 1312 | |
|
|
|
|
|
- Validation Dataset: 1190 |
|
| Class | Positive Count of Class| |
|
|:-------------|:--------| |
|
| ConditionalLabel | 192 | |
|
| UnconditionalLabel | 136 | |
|
|
|
## Training procedure |
|
|
|
### Training hyperparameters |
|
|
|
The following hyperparameters were used during training: |
|
- learning_rate: 4.02e-05 |
|
- train_batch_size: 16 |
|
- eval_batch_size: 16 |
|
- seed: 42 |
|
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 |
|
- lr_scheduler_type: cosine |
|
- lr_scheduler_warmup_steps: 200 |
|
- num_epochs: 6 |
|
|
|
### Training results |
|
|
|
| Training Loss | Epoch | Step | Validation Loss | Precision-micro | Precision-samples | Precision-weighted | Recall-micro | Recall-samples | Recall-weighted | F1-micro | F1-samples | F1-weighted | |
|
|:-------------:|:-----:|:----:|:---------------:|:---------------:|:-----------------:|:------------------:|:------------:|:--------------:|:---------------:|:--------:|:----------:|:-----------:| |
|
| 0.5361 | 1.0 | 369 | 0.4405 | 0.3405 | 0.1655 | 0.4102 | 0.6311 | 0.1622 | 0.6311 | 0.4423 | 0.1622 | 0.4503 | |
|
| 0.3692 | 2.0 | 738 | 0.3437 | 0.4631 | 0.1794 | 0.4929 | 0.6890 | 0.1761 | 0.6890 | 0.5539 | 0.1762 | 0.5604 | |
|
| 0.182 | 3.0 | 1107 | 0.3915 | 0.4702 | 0.1857 | 0.4871 | 0.7470 | 0.1891 | 0.7470 | 0.5771 | 0.1854 | 0.5800 | |
|
| 0.0757 | 4.0 | 1476 | 0.4713 | 0.4960 | 0.1882 | 0.4986 | 0.7530 | 0.1908 | 0.7530 | 0.5981 | 0.1877 | 0.5987 | |
|
| 0.0298 | 5.0 | 1845 | 0.4971 | 0.5161 | 0.1840 | 0.5184 | 0.7317 | 0.1857 | 0.7317 | 0.6053 | 0.1829 | 0.6058 | |
|
| 0.0152 | 6.0 | 2214 | 0.5295 | 0.5138 | 0.1866 | 0.5169 | 0.7378 | 0.1874 | 0.7378 | 0.6058 | 0.1852 | 0.6065 | |
|
|
|
|label | precision |recall |f1-score| support| |
|
|:-------------:|:---------:|:-----:|:------:|:------:| |
|
|ConditionalLabel |0.490 |0.760 |0.595 | 192.0 | |
|
|UnconditionalLabel |0.555 |0.706 | 0.621 | 136.0 | |
|
| |
|
|
|
### Environmental Impact |
|
Carbon emissions were measured using [CodeCarbon](https://github.com/mlco2/codecarbon). |
|
- **Carbon Emitted**: 0.02845 kg of CO2 |
|
- **Hours Used**: 0.702 hours |
|
|
|
### Training Hardware |
|
- **On Cloud**: yes |
|
- **GPU Model**: 1 x Tesla T4 |
|
- **CPU Model**: Intel(R) Xeon(R) CPU @ 2.00GHz |
|
- **RAM Size**: 12.67 GB |
|
|
|
|
|
|
|
### Framework versions |
|
|
|
- Transformers 4.38.1 |
|
- Pytorch 2.1.0+cu121 |
|
- Datasets 2.18.0 |
|
- Tokenizers 0.15.2 |
