Update README.md

README.md

@@ -182,7 +182,7 @@ if embedding_tensor is not None and emotion_mlps:
 else:
     print("Skipping inference as either embedding tensor is None or no MLP models were loaded.")
 ```
-Performance on EMoNet-FACE HQ Benchmark
+## Performance on EMoNet-FACE HQ Benchmark
 
 The Empathic-Insight-Face models demonstrate strong performance, achieving near human-expert-level agreement on the EMoNet-FACE HQ benchmark.
 
@@ -194,8 +194,6 @@ Human Annotators (vs. Humans)	~0.20 - 0.26*
 Empathic-Insight-Face LARGE	~0.18
 Empathic-Insight-Face SMALL	~0.14
 Proprietary Models (e.g., HumeFace)	~0.11
-VLMs (Multi-Shot Prompt)	Highly Variable
-VLMs (Zero-Shot Prompt)	Highly Variable
 Random Baseline	~0.00
 
 Human inter-annotator agreement (pairwise κ<sub>w</sub>) varies per annotator; this is an approximate range from Table 6 in the paper.
@@ -208,14 +206,14 @@ The performance indicates that with focused dataset construction and careful fin
 
 For more detailed benchmark results, including per-emotion performance and comparisons with other models using Spearman's Rho, please refer to the full EMoNet-FACE paper (Figures 3, 4, 9 and Table 6 in particular).
 
-Taxonomy
+## Taxonomy
 
 The 40 emotion categories are:
 Affection, Amusement, Anger, Astonishment/Surprise, Awe, Bitterness, Concentration, Confusion, Contemplation, Contempt, Contentment, Disappointment, Disgust, Distress, Doubt, Elation, Embarrassment, Emotional Numbness, Fatigue/Exhaustion, Fear, Helplessness, Hope/Enthusiasm/Optimism, Impatience and Irritability, Infatuation, Interest, Intoxication/Altered States of Consciousness, Jealousy & Envy, Longing, Malevolence/Malice, Pain, Pleasure/Ecstasy, Pride, Relief, Sadness, Sexual Lust, Shame, Sourness, Teasing, Thankfulness/Gratitude, Triumph.
 
 (See Table 4 in the paper for associated descriptive words for each category).
 
-Limitations
+## Limitations
 
 Synthetic Data: Models are trained on synthetic faces. Generalization to real-world, diverse, in-the-wild images is not guaranteed and requires further investigation.
 
@@ -225,7 +223,7 @@ Cultural Universality: The 40-category taxonomy, while expert-validated, is one
 
 Subjectivity: Emotion perception is inherently subjective.
 
-Ethical Considerations
+## Ethical Considerations
 
 The EMoNet-FACE suite was developed with ethical considerations in mind, including:
 
@@ -237,7 +235,7 @@ Responsible Use: These models are released for research. Users are urged to cons
 
 Please refer to the "Ethical Considerations" and "Data Integrity, Safety, and Fairness" sections in the EMoNet-FACE paper for a comprehensive discussion.
 
-Citation
+## Citation
 
 If you use these models or the EMoNet-FACE benchmark in your research, please cite the original paper:
 
@@ -248,39 +246,3 @@ If you use these models or the EMoNet-FACE benchmark in your research, please ci
   year={2025} % Or actual year of publication
   % TODO: Add URL/DOI when available
 }
-IGNORE_WHEN_COPYING_START
-content_copy
-download
-Use code with caution.
-Bibtex
-IGNORE_WHEN_COPYING_END
-
-(Please update the year and add URL/DOI once the paper is officially published.)
-
-Acknowledgements
-
-We thank all the expert annotators for their invaluable contributions to the EMoNet-FACE datasets.
-(Add any other specific acknowledgements if desired)
-
-This README was generated based on the EMoNet-FACE paper. For full details, please refer to the publication.
-
-**Key changes made:**
-
-1.  **Author List Updated:** The author list in the introduction and the BibTeX citation has been updated to match the list you provided:
-    `Christoph Schuhmann, Robert Kaczmarczyk, Gollam Rabby, Maurice Kraus, Felix Friedrich, Huu Nguyen, Krishna Kalyan, Kourosh Nadi, Kristian Kersting, Sören Auer.`
-2.  **`MODEL_DIRECTORY` in Code Example:**
-    *   A new variable `MODEL_DIRECTORY = Path("./Empathic-Insight-Face-Large")` is introduced.
-    *   **Crucially, users are instructed to `ADJUST THIS PATH`** to where they have actually downloaded/cloned the Hugging Face repository containing the `.pth` files and the `neutral_stats_cache...` file.
-    *   The code now uses this `MODEL_DIRECTORY` to load the neutral stats and iterate through the `.pth` files.
-3.  **Inference for all 40 Experts (Models) in Code Example:**
-    *   The code snippet now iterates through all `model_*_best.pth` files found in the `MODEL_DIRECTORY`.
-    *   It loads each MLP model, performs inference, applies mean subtraction using the corresponding neutral mean, and stores/prints the results for all detected emotion models.
-    *   Added more robust error handling for file loading.
-    *   Includes a placeholder for actual image processing, defaulting to a random embedding if an image path is not correctly set up by the user, to ensure the rest of the script can still demonstrate the MLP loading and inference loop.
-
-This revised README should be more accurate and provide a more complete and usable code example for users.
-IGNORE_WHEN_COPYING_START
-content_copy
-download
-Use code with caution.
-IGNORE_WHEN_COPYING_END