ehsanulhaque92/multimodal-prescriptive-pdm
Final Deploy v18: Correctly pull Ollama model during build
Final Deploy v17: Increase sleep time for Ollama startup
Final Deploy v16: Run Ollama inside the container
Final Deploy v15: Switch to gpt2 for maximum compatibility
Final Deploy v14: Explicitly pass token to InferenceClient
Final Deploy Fix: Switch to direct InferenceClient to bypass LangChain bugs
Final Deploy v13: Switch to direct InferenceClient API call
Final Deploy v12: Correct HuggingFaceEndpoint parameters
Final Deploy v11: Revert to HuggingFaceEndpoint with stable model
Final Deploy v10: Specify task for HuggingFaceHub
Final Deploy v9: Switch to HuggingFaceHub and flan-t5-large
Final Deploy Fix: Switch to Gemma-2b-it model for HF API compatibility
Final Deploy Fix: Make ChromaDB path configurable
Final Deploy v8: Convert Path objects to strings for ChromaDB
Final Deploy v7: Pre-download embedding model in Docker build
Final Fix v3: Correct typo in sentence-transformer model name
Fix: Set writable cache folder for sentence-transformers
Final Fix v3: Retrain models in clean environment for version consistency
Final Fix v2: Align scikit-learn versions and retrain models
Final Fix: Retrain all models for deployment compatibility and update notebooks
Final Fix: Add trained models and configure LFS
Final Deploy v5: Add git lfs pull to Dockerfile
Final Deploy v4: Pin lightgbm version and add libstdc++6
Final Deploy v3: Add Hugging Face metadata to README
Initial commit: Rebuild of complete RAG application