update

README.md

@@ -23,7 +23,7 @@ It takes a DNA sequence corresponding to a transcript and produces **one logit p
 Model name on Hugging Face:
 
 ```python
-genatator-caduceus-ps-multispecies-transcript-type-classification
+genatator-caduceus-ps-multispecies-transcript-type
 ````
 
 Architecture properties:
@@ -67,7 +67,7 @@ The model predicts a single binary output:
 
 The model returns **one logit per input sequence**.
 
-To obtain probabilities, apply a sigmoid to the logits.
+To obtain probabilities for the **lncRNA** class, apply a sigmoid to the logits.
 
 ---
 
@@ -76,7 +76,7 @@ To obtain probabilities, apply a sigmoid to the logits.
 ```python
 from transformers import AutoTokenizer, AutoModelForSequenceClassification
 
-repo_id = "shmelev/genatator-caduceus-ps-multispecies-transcript-type-classification"
+repo_id = "shmelev/genatator-caduceus-ps-multispecies-transcript-type"
 
 tokenizer = AutoTokenizer.from_pretrained(
     repo_id,
@@ -99,7 +99,7 @@ model.eval()
 import torch
 from transformers import AutoTokenizer, AutoModelForSequenceClassification
 
-repo_id = "shmelev/genatator-caduceus-ps-multispecies-transcript-type-classification"
+repo_id = "shmelev/genatator-caduceus-ps-multispecies-transcript-type"
 
 tokenizer = AutoTokenizer.from_pretrained(repo_id, trust_remote_code=True)
 model = AutoModelForSequenceClassification.from_pretrained(repo_id, trust_remote_code=True)
@@ -121,7 +121,7 @@ probs = torch.sigmoid(logits)
 print("Input shape:", input_ids.shape)
 print("Logits shape:", logits.shape)
 print("Probabilities shape:", probs.shape)
-print("Probabilities:", probs.squeeze(-1))
+print("Probabilities of lncRNA:", probs.squeeze(-1))
 ```
 
 Example output:
@@ -130,5 +130,5 @@ Example output:
 Input shape: torch.Size([2, sequence_length])
 Logits shape: torch.Size([2, 1])
 Probabilities shape: torch.Size([2, 1])
-Probabilities: tensor([0.91, 0.08])
-```
+Probabilities of lncRNA: tensor([0.08, 0.91])
+```