sentence-transformers
English
embeddings
hyperbolic-geometry
poincare-ball
21-dimensional
scbe-aethermoore
sacred-tongues
Instructions to use issdandavis/phdm-21d-embedding with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- sentence-transformers
How to use issdandavis/phdm-21d-embedding with sentence-transformers:
from sentence_transformers import SentenceTransformer model = SentenceTransformer("issdandavis/phdm-21d-embedding") sentences = [ "The weather is lovely today.", "It's so sunny outside!", "He drove to the stadium." ] embeddings = model.encode(sentences) similarities = model.similarity(embeddings, embeddings) print(similarities.shape) # [3, 3] - Notebooks
- Google Colab
- Kaggle
The First Agent Product Surface Is a Local MCP
#5
by issdandavis - opened
Most agent failures do not start with reasoning. They start with orientation drift. A model opens a large repo, guesses what matters, and then explains or executes against stale context.
The practical SCBE change this week is a local MCP surface for repo orientation:
scbe_system_manifestscbe_command_catalogscbe_reference_lookupscbe_tokenizer_health
Those tools move repo identity, command lookup, and canonical document retrieval out of prompt memory and into a queryable operator surface.
That works well with the existing SCBE terminal control plane described in docs/TERMINAL_OPS_QUICKSTART.md.
The cleaner split is:
- MCP for orientation
- CLI for execution
That is the product boundary I trust more for multi-agent systems.
Sources
mcp/scbe-server/README.mdmcp/scbe-server/server.mjsdocs/map-room/scbe_source_roots.mddocs/TERMINAL_OPS_QUICKSTART.mdskills/scbe-mcp-systems/SKILL.mdCLAUDE.md