This is a LinkTransformer model. At its core this model this is a sentence transformer model sentence-transformers model- it just wraps around the class. It is designed for quick and easy record linkage (entity-matching) through the LinkTransformer package. The tasks include clustering, deduplication, linking, aggregation and more. Notwithstanding that, it can be used for any sentence similarity task within the sentence-transformers framework as well. It maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search. Take a look at the documentation of sentence-transformers if you want to use this model for more than what we support in our applications.

This model has been fine-tuned on the model : sentence-transformers/paraphrase-multilingual-mpnet-base-v2. It is pretrained for the language : - en

This model was trained on a dataset prepared by linking product classifications from UN stats. This model is designed to link different products to their industrial classification (ISIC) - trained on variation brought on by product level correspondance. It was trained for 100 epochs using other defaults that can be found in the repo's LinkTransformer config file - LT_training_config.json

Using this model becomes easy when you have LinkTransformer installed:

Then you can use the model like this:

Any Sentence Transformers can be used as a backbone by simply adding a pooling layer. Any other transformer on HuggingFace can also be used by specifying the option add_pooling_layer==True The model was trained using SupCon loss. Usage can be found in the package docs. The training config can be found in the repo with the name LT_training_config.json To replicate the training, you can download the file and specify the path in the config_path argument of the training function. You can also override the config by specifying the training_args argument. Here is an example.

You can also use this package for deduplication (clusters a df on the supplied key column). Merging a fine class (like product) to a coarse class (like HS code) is also possible. Read our paper and the documentation for more!

You can evaluate the model using the LinkTransformer package's inference functions. We have provided a few datasets in the package for you to try out. We plan to host more datasets on Huggingface and our website (Coming soon) that you can take a look at.

The model was trained with the parameters:

DataLoader:

torch.utils.data.dataloader.DataLoader of length 281 with parameters:

Loss:

Parameters of the fit()-Method:

{
    "epochs": 100,
    "evaluation_steps": 141,
    "evaluator": "sentence_transformers.evaluation.SequentialEvaluator.SequentialEvaluator",
    "max_grad_norm": 1,
    "optimizer_class": "<class 'torch.optim.adamw.AdamW'>",
    "optimizer_params": {
        "lr": 2e-05
    },
    "scheduler": "WarmupLinear",
    "steps_per_epoch": null,
    "warmup_steps": 28100,
    "weight_decay": 0.01
}

LinkTransformer( (0): Transformer({'max_seq_length': 128, 'do_lower_case': False}) with Transformer model: XLMRobertaModel (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False}) )

## Citing & Authors

@misc{arora2023linktransformer, title={LinkTransformer: A Unified Package for Record Linkage with Transformer Language Models}, author={Abhishek Arora and Melissa Dell}, year={2023}, eprint={2309.00789}, archivePrefix={arXiv}, primaryClass={cs.CL} }