dbourget / philai-embeddings-v1.1

dbourget/philai-embeddings-v1.1

This is a sentence-transformers model: It maps sentences & paragraphs to a 1024 dimensional dense vector space and can be used for tasks like clustering or semantic search.

The model is based on BERT-large-uncased, further trained using the TSDAE technique described here on a corpus of open access philosophy from PhilArchive.org .

The aim was to make an embedding model that can correctly capture semantic similarity between sentences that use philosophical jargon. The aim was partly achieved. In our informal benchmarks, it performs better than Ada3-large. However, it is still far from perfect and we hope to release an improved model soon.

Usage (Sentence-Transformers)

Using this model becomes easy when you have sentence-transformers installed:

pip install -U sentence-transformers

Then you can use the model like this:

from sentence_transformers import SentenceTransformer
sentences = ["This is an example sentence", "Each sentence is converted"]

model = SentenceTransformer('dbourget/philai-embeddings-v1.1')
embeddings = model.encode(sentences)
print(embeddings)

Usage (HuggingFace Transformers)

Without sentence-transformers , you can use the model like this: First, you pass your input through the transformer model, then you have to apply the right pooling-operation on-top of the contextualized word embeddings.

from transformers import AutoTokenizer, AutoModel
import torch


def cls_pooling(model_output, attention_mask):
    return model_output[0][:,0]


# Sentences we want sentence embeddings for
sentences = ['This is an example sentence', 'Each sentence is converted']

# Load model from HuggingFace Hub
tokenizer = AutoTokenizer.from_pretrained('dbourget/philai-embeddings-v1.1')
model = AutoModel.from_pretrained('dbourget/philai-embeddings-v1.1')

# Tokenize sentences
encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt')

# Compute token embeddings
with torch.no_grad():
    model_output = model(**encoded_input)

# Perform pooling. In this case, cls pooling.
sentence_embeddings = cls_pooling(model_output, encoded_input['attention_mask'])

print("Sentence embeddings:")
print(sentence_embeddings)

Evaluation Results

For an automated evaluation of this model, see the Sentence Embeddings Benchmark : https://seb.sbert.net

Training

The model was trained with the parameters:

DataLoader :

torch.utils.data.dataloader.DataLoader of length 122952 with parameters:

{'batch_size': 28, 'sampler': 'torch.utils.data.sampler.RandomSampler', 'batch_sampler': 'torch.utils.data.sampler.BatchSampler'}

Loss :

sentence_transformers.losses.DenoisingAutoEncoderLoss.DenoisingAutoEncoderLoss

Parameters of the fit()-Method:

{
    "epochs": 5,
    "evaluation_steps": 61476,
    "evaluator": "sentence_transformers.evaluation.TripletEvaluator.TripletEvaluator",
    "max_grad_norm": 1,
    "optimizer_class": "<class 'torch.optim.adamw.AdamW'>",
    "optimizer_params": {
        "lr": 2e-05
    },
    "scheduler": "constantlr",
    "steps_per_epoch": null,
    "warmup_steps": 10000,
    "weight_decay": 0
}

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: BertModel 
  (1): Pooling({'word_embedding_dimension': 1024, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
)

Citing & Authors