Update README.md

README.md

@@ -1,95 +1,116 @@
 ---
-library_name: sentence-transformers
-pipeline_tag: sentence-similarity
+library_name: hierarchy-transformers
+pipeline_tag: feature-extraction
 tags:
-- sentence-transformers
+- hierarchy-transformers
 - feature-extraction
-- sentence-similarity
+- hierarchy-encoding
+- subsumption-relationships
 - transformers
-
+license: apache-2.0
+language:
+- en
 ---
 
 # Hierarchy-Transformers/HiT-MPNet-WordNetNoun-Hard
 
-This is a [sentence-transformers](https://www.SBERT.net) model: It maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search.
-
-<!--- Describe your model here -->
-
-## Usage (Sentence-Transformers)
+A **Hi**erarchy **T**ransformer Encoder (HiT) model that explicitly encodes entities according to their hierarchical relationships.
 
-Using this model becomes easy when you have [sentence-transformers](https://www.SBERT.net) installed:
+### Model Description
 
-```
-pip install -U sentence-transformers
-```
+<!-- Provide a longer summary of what this model is. -->
 
-Then you can use the model like this:
+HiT-MPNet-WordNetNoun-Hard is a HiT model trained on WordNet's noun hierarchy with **hard** negative sampling.
 
-```python
-from sentence_transformers import SentenceTransformer
-sentences = ["This is an example sentence", "Each sentence is converted"]
+- **Developed by:** [Yuan He](https://www.yuanhe.wiki/), Zhangdie Yuan, Jiaoyan Chen, and Ian Horrocks
+- **Model type:** Hierarchy Transformer Encoder (HiT)
+- **License:** Apache license 2.0
+- **Hierarchy**: WordNet (Noun)
+- **Dataset**: Download `wordnet.zip` from the [Zenodo link](https://zenodo.org/doi/10.5281/zenodo.10511042)
+- **Pre-trained model:** [sentence-transformers/all-mpnet-base-v2](https://huggingface.co/sentence-transformers/all-mpnet-base-v2)
+- **Training Objectives**: Jointly optimised on *hyperbolic clustering* and *hyperbolic centripetal* losses
 
-model = SentenceTransformer('Hierarchy-Transformers/HiT-MPNet-WordNetNoun-Hard')
-embeddings = model.encode(sentences)
-print(embeddings)
-```
+### Model Sources
 
+<!-- Provide the basic links for the model. -->
 
+- **Repository:** https://github.com/KRR-Oxford/HierarchyTransformers
+- **Paper:** [Language Models as Hierarchy Encoders](tbd)
 
-## Usage (HuggingFace Transformers)
-Without [sentence-transformers](https://www.SBERT.net), 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.
+## Usage
 
-```python
-from transformers import AutoTokenizer, AutoModel
-import torch
+<!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
 
+HiT models are used to encode entities (presented as texts) and predict their hierarhical relationships in hyperbolic space. 
 
-#Mean Pooling - Take attention mask into account for correct averaging
-def mean_pooling(model_output, attention_mask):
-    token_embeddings = model_output[0] #First element of model_output contains all token embeddings
-    input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float()
-    return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9)
+### Get Started
 
+Install `hierarchy_transformers` (check our [repository](https://github.com/KRR-Oxford/HierarchyTransformers)) through `pip` or `GitHub`.
 
-# Sentences we want sentence embeddings for
-sentences = ['This is an example sentence', 'Each sentence is converted']
+Use the code below to get started with the model.
 
-# Load model from HuggingFace Hub
-tokenizer = AutoTokenizer.from_pretrained('Hierarchy-Transformers/HiT-MPNet-WordNetNoun-Hard')
-model = AutoModel.from_pretrained('Hierarchy-Transformers/HiT-MPNet-WordNetNoun-Hard')
+```python
+from hierarchy_transformers import HierarchyTransformer
+from hierarchy_transformers.utils import get_torch_device
 
-# Tokenize sentences
-encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt')
+# set up the device (use cpu if no gpu found)
+gpu_id = 0
+device = get_torch_device(gpu_id)
 
-# Compute token embeddings
-with torch.no_grad():
-    model_output = model(**encoded_input)
+# load the model
+model = HierarchyTransformer.load_pretrained('Hierarchy-Transformers/HiT-MiniLM-L12-WordNet', device)
 
-# Perform pooling. In this case, mean pooling.
-sentence_embeddings = mean_pooling(model_output, encoded_input['attention_mask'])
+# entity names to be encoded.
+entity_names = ["computer", "personal computer", "fruit", "berry"]
 
-print("Sentence embeddings:")
-print(sentence_embeddings)
+# get the entity embeddings
+entity_embeddings = model.encode(entity_names)
 ```
 
+### Default Probing for Subsumption Prediction
 
+Use the entity embeddings to predict the subsumption relationships between them.
 
-## Evaluation Results
-
-<!--- Describe how your model was evaluated -->
+```python
+# suppose we want to compare "personal computer" and "computer", "berry" and "fruit"
+child_entity_embeddings = model.encode(["personal computer", "berry"], convert_to_tensor=True)
+parent_entity_embeddings = model.encode(["computer", "fruit"], convert_to_tensor=True)
+
+# compute the hyperbolic distances and norms of entity embeddings
+dists = model.manifold.dist(child_entity_embeddings, parent_entity_embeddings)
+child_norms = model.manifold.dist0(child_entity_embeddings)
+parent_norms = model.manifold.dist0(parent_entity_embeddings)
+
+# use the empirical function for subsumption prediction proposed in the paper
+# `centri_score_weight` and the overall threshold are determined on the validation set
+subsumption_scores = dists + centri_score_weight * (parent_norms - child_norms)
+```
 
-For an automated evaluation of this model, see the *Sentence Embeddings Benchmark*: [https://seb.sbert.net](https://seb.sbert.net?model_name=Hierarchy-Transformers/HiT-MPNet-WordNetNoun-Hard)
+Training and evaluation scripts are available at [GitHub](https://github.com/KRR-Oxford/HierarchyTransformers).
+Technical details are presented in the [paper](tbd).
 
 
 
 ## Full Model Architecture
 ```
 HierarchyTransformer(
-  (0): Transformer({'max_seq_length': 384, 'do_lower_case': False}) with Transformer model: MPNetModel 
-  (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})
+  (0): Transformer({'max_seq_length': 128, 'do_lower_case': False}) with Transformer model: BertModel 
+  (1): Pooling({'word_embedding_dimension': 384, '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
+## Citation
+
+<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
+
+**BibTeX:**
+
+Preprint on Arxiv:
+
+[More Information Needed]
+
+
+
+## Model Card Contact
 
-<!--- Describe where people can find more information -->
+For any queries or feedback, please contact Yuan He (yuan.he@cs.ox.ac.uk).