README updated

README.md

@@ -1,12 +1,3 @@
----
-license: mit
-base_model:
-- aletlvl/Nicheformer
-tags:
-- single-cell
-- transcriptomics
-- biology
----
 # Nicheformer
 
 Nicheformer is a transformer-based model designed for understanding and predicting cellular niches and their interactions. The model uses masked language modeling to learn representations of cellular contexts and their relationships.
@@ -41,8 +32,13 @@ from transformers import AutoModelForMaskedLM, AutoTokenizer
 import anndata as ad
 
 # Load model and tokenizer
-model = AutoModelForMaskedLM.from_pretrained("aletlvl/Nicheformer")
-tokenizer = AutoTokenizer.from_pretrained("aletlvl/Nicheformer")
+model = AutoModelForMaskedLM.from_pretrained("aletlvl/Nicheformer", trust_remote_code=True)
+tokenizer = AutoTokenizer.from_pretrained("aletlvl/Nicheformer", trust_remote_code=True)
+
+# Set technology mean for HF tokenizer
+technology_mean_path = 'technology_mean.npy'
+technology_mean = np.load(technology_mean_path)
+tokenizer._load_technology_mean(technology_mean)
 
 # Load your single-cell data
 adata = ad.read_h5ad("your_data.h5ad")
@@ -50,8 +46,13 @@ adata = ad.read_h5ad("your_data.h5ad")
 # Tokenize the data
 inputs = tokenizer(adata)
 
-# Get predictions
-outputs = model(**inputs)
+# Get embeddings
+embeddings = model.get_embeddings(
+              input_ids=inputs["input_ids"],
+              attention_mask=inputs["attention_mask"],
+              layer=-1,
+              with_context=False
+            )
 ```
 
 ## Training Data
@@ -74,7 +75,6 @@ The model was trained on single-cell gene expression data from various tissues a
 - Performance may vary depending on the quality and type of input data
 - The model works best with data from supported species and technologies
 
-
 ## License
 
 This model is released under the MIT License. See the LICENSE file for more details.
@@ -89,7 +89,6 @@ This is the official repository for **Nicheformer: a foundation model for single
 
 [![Preprint](https://img.shields.io/badge/preprint-available-brightgreen)](https://www.biorxiv.org/content/10.1101/2024.04.15.589472v1)  
 
-
 ## Citation
 
 If you use our tool or build upon our concepts in your own work, please cite it as
@@ -130,4 +129,4 @@ We provide the Nicheformer pretraining weights on Mendeley data, they can be dow
 For questions and help requests, you can reach out (preferably) on GitHub or email to the corresponding author. 
 
 
-[issue-tracker]: https://github.com/theislab/nicheformer/issues
+[issue-tracker]: https://github.com/theislab/nicheformer/issues