Upd DT and README:

README.md

@@ -43,9 +43,11 @@ tags:
 
 
 ### 🧠 **MedSwin Medical Specialist Models**
-- **MedSwin TA** (default) - Task-Aware merged model
-- **MedSwin SFT** - Supervised Fine-Tuned model
-- **MedSwin KD** - Knowledge Distillation model  
+- **MedSwin DT** (85% acc) - DaRE-TIES merged model from SFT and KD.
+- **MedSwin NSL** (85% acc) - NuSLERP merged model from SFT and KD.
+- **MedSwin TA** (84% acc) - Task-Arimethic merged model from SFT and KD.
+- **MedSwin SFT** - (82% acc) Supervised Fine-Tuned model.
+- **MedSwin KD** - (83% acc) Knowledge Distillation from MedGemma-27b-it. 
 - Models download on-demand for efficient resource usage
 - Fine-tuned on MedAlpaca-7B for medical domain expertise
 
@@ -106,7 +108,7 @@ Sample [conversation](sample.md) between MAC-system and patient are recorded.
 2. **Configure Settings**: 
    - Enable/disable Document RAG
    - Enable/disable Web Search (MCP)
-   - Select medical model (MedSwin SFT/KD/TA)
+   - Select medical model (MedSwin SFT/KD/TA/DT/NSL)
 3. **Ask Questions**: Type your medical question in any language
 4. **Get Answers**: Receive comprehensive answers based on:
    - Your uploaded documents (if RAG enabled)
@@ -116,7 +118,7 @@ Sample [conversation](sample.md) between MAC-system and patient are recorded.
 
 ## 🔧 Technical Details
 
-- **Medical Models**: MedSwin/MedSwin-7B-SFT, MedSwin-7B-KD, MedSwin-Merged-TA-SFT-0.7
+- **Medical Models**: MedSwin/MedSwin-7B-SFT, MedSwin-7B-KD, MedSwin-Merged-TA-SFT-0.7, MedSwin-Merged-NuSLERP-KD-0.7, MedSwin-Merged-DaRE-TIES-KD-0.7
 - **Architecture**: MAC (Multi-Agent Collaboration) - Gemini Supervisor + MedSwin Specialist
 - **Translation**: Gemini MCP (gemini-2.5-flash-lite)
 - **Document Parsing**: Gemini MCP (PDF, Word, TXT, MD, JSON, XML, CSV)

config.py

@@ -8,6 +8,7 @@ MEDSWIN_MODELS = {
     "MedSwin KD": "MedSwin/MedSwin-7B-KD",
     "MedSwin TA": "MedSwin/MedSwin-Merged-TA-SFT-0.7",
     "MedSwin DT": "MedSwin/MedSwin-Merged-DaRE-TIES-KD-0.7",
+    "MedSwin NSL": "MedSwin/MedSwin-Merged-NuSLERP-KD-0.7",
 }
 DEFAULT_MEDICAL_MODEL = "MedSwin TA"
 EMBEDDING_MODEL = "abhinand/MedEmbed-large-v0.1"