PyPI Stats

Search

All packages
Top packages

Track packages

squish-ai


PyPI page
Home page
Author: None
License: BUSL-1.1
Summary: Local LLM inference server for Apple Silicon. Block-level paged KV cache for long-context workloads. 5.4× faster end-to-end on 4K-token prompts vs Ollama, less RAM, INT3 support for Qwen3. OpenAI-compatible API.
Latest version: 9.34.14
Required dependencies: fastapi | huggingface-hub | mlx | mlx-lm | mlx-vlm | numpy | orjson | rich | safetensors | sse-starlette | transformers | uvicorn | zstandard
Optional dependencies: accelerate | cryptography | datasets | hnswlib | httpx | llmlingua | lm-eval | maturin | mkdocs-exclude | mkdocs-material | mypy | nltk | pytest | pytest-asyncio | pytest-cov | pytest-timeout | rouge_score | ruff | sacrebleu | sqlite-vec | torch | xgrammar

Downloads last day: 18
Downloads last week: 723
Downloads last month: 2,346