Update README.md

README.md

@@ -6,7 +6,6 @@
 
 🔗  **[Github](https://github.com/ta012/SSLAM) | [Paper](https://openreview.net/pdf?id=odU59TxdiB) | [Open Review](https://openreview.net/forum?id=odU59TxdiB) | [Pre-trained Models(HuggingFace)](https://huggingface.co/ta012/SSLAM) | [Pre-trained Models(Google Drive)](https://drive.google.com/drive/folders/1G0icv-hdqDEqnfP4EFszMXhFnWWM09gT?usp=sharing) **
 
----
 
 # 📋 Table of Contents
 - [Why SSLAM?](#why-sslam)
@@ -16,29 +15,22 @@
   - [Inference Installation](#inference-installation)
   - [Model Weights](#model-weights)
   - [Using SSLAM](#using-sslam)
-- [Training Mode](#training-mode)
-  - [Training Installation](#training-installation)
-  - [Data Preparation](#️data-preparation)
-  - [Pre-Training](#pre-training)
-- [Checklist](#checklist)
 - [Acknowledgements](#acknowledgements)
 - [Citation](#citation)
 
----
 
 ## 🔍Why SSLAM? 
 🔊 **Real-world audio is polyphonic**—multiple overlapping sound sources are common in everyday environments.  
 ❌ **Existing SSL models focus on monophonic audio,** limiting their ability to generalize to real-world scenarios. Their benchmarks are primarily monophonic, and their pre-training does not account for polyphonic environments.   
 💡 **SSLAM bridges this gap** by introducing **self-supervised learning from audio mixtures**, enabling robust learning across **both monophonic and polyphonic soundscapes**.
 
----
 
 ## 🎼Key Features
 ✅ **Self-Supervised Learning from Audio Mixtures (SSLAM)** – improving robustness to real-world polyphonic audio  (multiple overlapping sounds).  
 ✅ **Source Retention Loss** – ensures the integrity of each sound source even in complex mixtures.  
 ✅ **SOTA Performance** – Achieves **+3.9% mAP improvement** on AudioSet-2M and **+9.1% on polyphonic datasets**.  
 
----
+
 ## 📊Results
 
 ### 1. Standard Audio-SSL Benchmark Datasets
@@ -47,7 +39,7 @@
 ### 2. Polyphonic Datasets
 ![Polyphonic Datasets](assets/poly_results.png)
 
----
+
 ## **🔍️Inference Mode**
 > **Note**: If you are already using [EAT](https://github.com/cwx-worst-one/EAT/tree/main) in your evaluation/inference pipeline, you can simply replace the weights with SSLAM weights, as the inference and evaluation code is identical to EAT.
 
@@ -59,7 +51,7 @@ conda create --prefix /path/to/sslam_eval_env -y python=3.9.13
 /path/to/sslam_eval_env/bin/python -m pip install pip==24.0 # downgrade pip
 /path/to/sslam_eval_env/bin/pip install -r SSLAM_Inference/requirements_sslam_eval.txt
 ```
----
+
 
 ## 📦Model Weights
 
@@ -94,53 +86,7 @@ cd SSLAM_Inference/scripts
 bash evaluate_AS2M_finetuned.sh # Reported mAP: 50.2
 ```
 
----
-## **📈Training Mode**
-We cover the self-supervised pre-training, fine-tuning and linear evaluation under this section.
-
-#### **📥Training Installation** 
-
-For training its better to install the fairseq in editable mode,
-
-```bash
-conda create --prefix /path/to/sslam_env -y python=3.9.13 ## env used for training
-/path/to/sslam_env/bin/python -m pip install pip==24.0 # downgrade pip
-cd SSLAM/
-git clone https://github.com/facebookresearch/fairseq.git
-
-##IMPORTANT: Copy the Pre-Training/SSLAM_Stage2 directory to SSLAM/fairseq 
-## so that the resultant path is SSLAM/fairseq/SSLAM_Stage2/.
-cd fairseq/
-
-## install all requirements apart from fairseq
-/path/to/sslam_env/bin/pip install -r SSLAM_Stage2/requirements_sslam.txt
-## install fairseq in editable mode
-/path/to/sslam_env/bin/pip install --editable ./
-```
-#### 🗄️Data Preparation
-We utilised AudioSet-2M (full set) for pre-training. For this phase, only the `train.tsv` file is required. Refer to [train.tsv for AudioSet-20K](data_manifests/manifest_as20k/train.tsv) to prepare the train.tsv file for your downloaded copy of AudioSet-2M.
-
-#### 🚀Pre-Training
-
-**Note:** This repository focuses solely on Stage 2 pre-training, which introduces our novel SSLAM pre-training strategy. 
-
-To begin Stage 2, you'll need a Stage 1 checkpoint. In our complete pre-training process, Stage 1 mirrors the approach in [EAT](https://github.com/cwx-worst-one/EAT/tree/main) and achieves similar performance. For convenience, we use the EAT checkpoint as the Stage 1 checkpoint.
-
-Download the [EAT](https://github.com/cwx-worst-one/EAT/tree/main) epoch 10 checkpoint using the link provided by the [EAT](https://github.com/cwx-worst-one/EAT/tree/main) repository: [EAT-base_epoch10_pt.pt](https://drive.google.com/file/d/10pklbY_fKraQUIBizSg1kv4lJXNWxpxl/view?usp=sharing).
-
-*Only the contents of the **models/** folder and a few parameters in the pre-training script differ between Stage 1 and Stage 2.*
 
-```bash
-cd SSLAM/fairseq/SSLAM_Stage2/scripts/
-bash pretrain_stage2.sh
-```
-
-
-## 📌Checklist 
-- [x] Inference Mode
-- [x] Pre-Training
-
----
 
 ## 🙏Acknowledgements