Update README.md

README.md

@@ -51,24 +51,7 @@ PyTorch Geometric: https://github.com/pyg-team/pytorch_geometric
 https://huggingface.co/spaces/awacke1/MN.Map.Hospitals.Top.Five
 ![image](https://user-images.githubusercontent.com/30595158/226150906-65fcdb27-b234-4500-8cd8-c6b88d1afa05.png)
 
-
-
-# Graph, Clinical Terminology, FHIR Apps and Services:
-
-CP Matplotlib, NetworkX, Streamlit, PyVis, st-click0detector, graphviz:
-https://huggingface.co/spaces/awacke1/CPVisGraph
-
-OMS and LOCUS:
-https://huggingface.co/spaces/awacke1/NLPGraphOMSandLOCUS
-
-https://huggingface.co/spaces/awacke1/Git-GPG-Git-Actions-01-GraphViz
-https://huggingface.co/spaces/awacke1/Dice-Roll-Treemap-Plotly
-https://huggingface.co/spaces/awacke1/GraphVis3
-https://huggingface.co/spaces/awacke1/GraphViz-Demo
-https://huggingface.co/spaces/awacke1/StreamlitGraphViz
-https://huggingface.co/spaces/awacke1/CardGameActivity-GraphViz
-
-Technical Architecture - Open Source Graph ML Libraries:
+# Technical Architecture - Open Source Graph ML Libraries:
 
 NetworkX: https://networkx.org/
 PyTorch GNN: https://github.com/microsoft/ptgnn
@@ -79,7 +62,6 @@ Deep Graph Library (DGL): https://github.com/dmlc
 PyTorch Geometric: https://github.com/pyg-team/pytorch_geometric
 
 
-
 # Clinical Terminology:
 # FHIR:
 https://huggingface.co/spaces/awacke1/Clinical-Terminology-FHIR-Assessment
@@ -188,105 +170,6 @@ Face Mesh Real Time - 30 Frames per second!
 ![image](https://user-images.githubusercontent.com/30595158/225768360-c64197ff-919f-47a9-8cc0-c6d5e73e5853.png)
 
 
-
-# ASR Voice and Virtual Assistants With Avatars
-1. https://huggingface.co/spaces/awacke1/ASR-openai-whisper-large
-2. https://huggingface.co/spaces/awacke1/ASR-voidful-wav2vec2-xlsr-multilingual-56
-3. https://huggingface.co/spaces/awacke1/ASR-nvidia-stt_en_conformer_ctc_large
-4. https://huggingface.co/spaces/awacke1/ASR-facebook-hubert-large-ls960-ft
-5. https://huggingface.co/spaces/awacke1/ASR-openai-whisper-tiny.en
-6. https://huggingface.co/spaces/awacke1/ASR-openai-whisper-tiny
-7. https://huggingface.co/spaces/awacke1/ASR-openai-whisper-medium
-8. https://huggingface.co/spaces/awacke1/ASR-nvidia-stt_en_conformer_transducer_xlarge
-9. https://huggingface.co/spaces/awacke1/ASR-openai-whisper-base
-10. https://huggingface.co/spaces/awacke1/ASR-facebook-wav2vec2-large-960h-lv60-self
-11. https://huggingface.co/spaces/awacke1/ASR-facebook-wav2vec2-base-960h
-12. https://huggingface.co/spaces/awacke1/ASR-High-Accuracy-Test
-13. https://huggingface.co/spaces/awacke1/ASRGenerateStory
-14. https://huggingface.co/spaces/awacke1/TTS-STT-Blocks
-15. https://huggingface.co/spaces/awacke1/2-LiveASR
-16. https://huggingface.co/spaces/awacke1/CloneAnyVoice
-17. https://huggingface.co/spaces/awacke1/ASR-SOTA-NvidiaSTTMozilla
-18. https://huggingface.co/spaces/awacke1/ASRSpeechRecognition1
-19. https://huggingface.co/spaces/awacke1/1110-ASRLiveExample
-20. https://huggingface.co/spaces/awacke1/Z1-ASRLiveSpeechRecognition-GR
-21. https://huggingface.co/spaces/awacke1/PrivateASRWithMemory
-22. https://huggingface.co/spaces/awacke1/TimerASRLive
-
-# Best Voice Apps - HF:
-1. https://huggingface.co/spaces/BilalSardar/Voice-Cloning
-2. https://huggingface.co/spaces/RamAnanth1/chatGPT_voice
-3. https://huggingface.co/spaces/Voicemod/speech-synthesis-demo
-4. https://huggingface.co/spaces/ysharma/Voice-to-Youtube
-5. https://huggingface.co/spaces/ramkamal2000/voice-conversion-yourtts
-6. https://huggingface.co/spaces/RamAnanth1/co_chat_voice
-7. https://huggingface.co/spaces/ysharma/Voice-to-jokes
-8. https://huggingface.co/spaces/jayesh95/Voice-QA
-
-
-
-# Supervised Learning (SL) for ML and Reinforcement Learning with Human Feedback (RLHF):
-
-For human imitation we use reinforcement learning for fine tuning since feedback based on rewards shapes the quality of output where an agent completes a task and then observes a result.  SL works on ranks not responses so is good for modifying elements at the token level however RLHF is trained to estimate the quality of the response with cumulative rewards for coherent conversation.  RLHF considers context and coherence of entire conversation.  Supervised learning is used to teach the model initially where the model learns basic structure and content.  In the RLHF stage the model is refined with responses that represent improved accuracy.
-
-
-
-
-
-# Mermaid Model for Core NLP Tasks:
-
-```mermaid
-graph LR;
-    A[Reader]-->B[Classifier];
-    A-->C[Retriever];
-    A-->D[Summarizer];
-    B-->E[Ranker];
-    B-->F[Query Classifier];
-    D-->G[Generator];
-    F-->H[Question Generator];
-    H-->G;
-    I[File Converter]-->J[Preprocessor];
-    J-->A;
-    I-->C;
-    K[Snowflake]-->B;
-    L[Oracle]-->B;
-    M[Pandas CSV]-->A;
-    N[Index]-->C;
-    N-->E;
-    O[Query with Filters]-->F;
-    P[Evaluation]-->E;
-    P-->F;
-    Q[Retraining]-->B;
-    Q-->E;
-    R[Annotation]-->B;
-```
-
-# Core NLP Task Model for QA
-
-Tasks:
-1. Reader
-2. Summarizer
-3. Classifier
-4. Retriever
-5. Ranker
-6. Query Classifier
-7. Question Generator
-8. Generator
-
-Connectors:
-1. File Converter
-2. Preprocessor
-3. Snowflake
-4. Oracle
-5. Pandas CSV
-
-Supported Workflow:
-1. Index
-2. Query with Filters
-3. Evaluation
-4. Retraining
-5. Annotation
-
 # QA Model Spaces:
 
 QA use cases include QA, Semantic Document and FAQ Search.