Update README.md

README.md

@@ -37,7 +37,7 @@ base_model:
 - google-bert/bert-base-uncased
 base_model_relation: finetune
 model-index:
-- name: Emotion Analyzer Bert
+- name: Bert Emotion Classifier
   results:
   - task:
       type: multi-label-classification
@@ -156,10 +156,10 @@ model-index:
     source:
       name: Kaggle Evaluation Notebook
       url: >-
-        https://www.kaggle.com/code/ravindranlogasanjeev/evaluation-logasanjeev-emotions-analyzer-bert/notebook
+        https://www.kaggle.com/code/ravindranlogasanjeev/evaluation-logasanjeev-bert-emotion-classifier/notebook
 ---
 
-# Emotions Analyzer Bert
+# Bert Emotion Classifier
 
 Fine-tuned [BERT-base-uncased](https://huggingface.co/bert-base-uncased) on [GoEmotions](https://huggingface.co/datasets/google-research-datasets/go_emotions) for multi-label classification (28 emotions). This updated version includes improved Macro F1, ONNX support for efficient inference, and visualizations for better interpretability.
 
@@ -176,7 +176,7 @@ Fine-tuned [BERT-base-uncased](https://huggingface.co/bert-base-uncased) on [GoE
 
 ## Try It Out
 
-For accurate predictions with optimized thresholds, use the [Gradio demo](https://logasanjeev-emotions-analyzer-bert-demo.hf.space). The demo now includes preprocessed text and the top 5 predicted emotions, in addition to thresholded predictions. Example predictions:
+For accurate predictions with optimized thresholds, use the [Gradio demo](https://logasanjeev-bert-emotion-classifier-demo.hf.space). The demo now includes preprocessed text and the top 5 predicted emotions, in addition to thresholded predictions. Example predictions:
 
 - **Input**: "I’m thrilled to win this award! 😄"
   - **Output**: `excitement: 0.5836, joy: 0.5290`
@@ -194,7 +194,7 @@ For accurate predictions with optimized thresholds, use the [Gradio demo](https:
 - **Hamming Loss**: 0.0377
 - **Avg Positive Predictions**: 1.4789
 
-For a detailed evaluation, including class-wise accuracy, precision, recall, F1, MCC, support, and thresholds, along with visualizations, check out the [Kaggle notebook](https://www.kaggle.com/code/ravindranlogasanjeev/evaluation-logasanjeev-emotions-analyzer-bert/notebook).
+For a detailed evaluation, including class-wise accuracy, precision, recall, F1, MCC, support, and thresholds, along with visualizations, check out the [Kaggle notebook](https://www.kaggle.com/code/ravindranlogasanjeev/evaluation-logasanjeev-bert-emotion-classifier/notebook).
 
 ### Class-Wise Performance
 
@@ -256,29 +256,26 @@ The easiest way to use the model with PyTorch is to programmatically fetch and u
 Run the following Python script to download `inference.py` and make predictions:
 
 ```python
-!pip install transformers torch huggingface_hub emoji -q
+# pip install transformers torch huggingface_hub emoji -q
 
-import shutil
-import os
 from huggingface_hub import hf_hub_download
-from importlib import import_module
+import importlib.util
 
-repo_id = "logasanjeev/emotions-analyzer-bert"
-local_file = hf_hub_download(repo_id=repo_id, filename="inference.py")
+# download inference script
+path = hf_hub_download(repo_id="logasanjeev/bert-emotion-classifier", filename="inference.py")
 
-current_dir = os.getcwd()
-destination = os.path.join(current_dir, "inference.py")
-shutil.copy(local_file, destination)
-
-inference_module = import_module("inference")
-predict_emotions = inference_module.predict_emotions
+# load module
+spec = importlib.util.spec_from_file_location("inference", path)
+inference = importlib.util.module_from_spec(spec)
+spec.loader.exec_module(inference)
 
+# run prediction
 text = "I’m thrilled to win this award! 😄"
-result, processed = predict_emotions(text)
-print(f"Input: {text}")
-print(f"Processed: {processed}")
-print("Predicted Emotions:")
-print(result)
+result, processed = inference.predict_emotions(text)
+
+print("Input:", text)
+print("Processed:", processed)
+print("Predicted Emotions:", result)
 ```
 
 #### Expected Output:
@@ -323,29 +320,26 @@ For faster and more efficient inference using ONNX, you can use `onnx_inference.
 Run the following Python script to download `onnx_inference.py` and make predictions:
 
 ```python
-!pip install transformers onnxruntime huggingface_hub emoji numpy -q
+# pip install transformers torch huggingface_hub emoji -q
 
-import shutil
-import os
 from huggingface_hub import hf_hub_download
-from importlib import import_module
+import importlib.util
 
-repo_id = "logasanjeev/emotions-analyzer-bert"
-local_file = hf_hub_download(repo_id=repo_id, filename="onnx_inference.py")
+# download inference script
+path = hf_hub_download(repo_id="logasanjeev/bert-emotion-classifier", filename="inference.py")
 
-current_dir = os.getcwd()
-destination = os.path.join(current_dir, "onnx_inference.py")
-shutil.copy(local_file, destination)
-
-onnx_inference_module = import_module("onnx_inference")
-predict_emotions = onnx_inference_module.predict_emotions
+# load module
+spec = importlib.util.spec_from_file_location("inference", path)
+inference = importlib.util.module_from_spec(spec)
+spec.loader.exec_module(inference)
 
+# run prediction
 text = "I’m thrilled to win this award! 😄"
-result, processed = predict_emotions(text)
-print(f"Input: {text}")
-print(f"Processed: {processed}")
-print("Predicted Emotions:")
-print(result)
+result, processed = inference.predict_emotions(text)
+
+print("Input:", text)
+print("Processed:", processed)
+print("Predicted Emotions:", result)
 ```
 
 #### Expected Output:
@@ -407,7 +401,7 @@ def preprocess_text(text):
     text = text.lower()
     return text
 
-repo_id = "logasanjeev/emotions-analyzer-bert"
+repo_id = "logasanjeev/bert-emotion-classifier"
 model = BertForSequenceClassification.from_pretrained(repo_id)
 tokenizer = BertTokenizer.from_pretrained(repo_id)