deepset Cloud Features
Here's what deepset Cloud enterprise ML/NLP platform offers
Custom NLP Pipelines
- Question answering
- Vector-based semantic search
- Hybrid retrieval
- Named entity recognition
- Query classification
- Pipeline templates
Prototypes & Feedback
- Build shareable prototypes
- Collect user feedback early
- Analyze feedback qualitatively and quantitatively
Latest Models
- Use deepset’s QA models with millions of downloads
- Use GPT-3 models
- Use Cohere Embedding models
- Use any model from Hugging Face Hub
- Upload private models
Scalable Cloud Document Store
- Store 100M+ documents with meta data
- Vector database
- KNN search
- Metadata filtering
Experiment Tracking
- Quickly configure experiments
- Run on auto-provisioned cloud infrastructure
- Compare experiment runs
- Traditional and model-based evaluation metrics
- Easy access to predictions and qualitative error analysis
- Tagging and adding notes
Deployment & Monitoring
- 1-Click deployments on auto-scaling GPU or CPU infrastructure
- Real-time updates on indexing status of your pipelines
- Monitor requests, usage and latency
API & Integrations
- Powerful REST API
- Hugging Face integration
Data
- Data management via UI
- Data management via API
- Upload any pdf or docx file
Notebooks
- Jupyter notebooks on GPU
- Jupyter notebooks on CPU
- SDK to connect directly to deepset Cloud from notebooks
Build NLP applications fast
From data to API-driven NLP backend services in days.