model-metadata

Convention based approach to MVC Model Validation and Model Metadata

DeepSeek Harness 插件：维护 llm-pi-ai 模型的准确上下文窗口，避免长会话被误判为上下文溢出。

Local dashboard for NVIDIA build.nvidia.com models with sortable metadata, model-card specs, live probes, and tool-calling detection.

Fetches and organizes Ollama cloud model information into clean tables showing sizes, capabilities, and direct links to models.

Unified model card schema for NANDA-compatible agent registries — covers LoRA adapters, edge ONNX, federated, and heuristic models with built-in validation and lifecycle tracking

🤖 Production-ready Python registry for OpenAI models with parameter validation, pricing data, and CLI tools. Supports GPT-5, Azure OpenAI, automated updates, and enterprise deployment.

Tools to inspect and maintain ollama models.

Offline LiteLLM pricing and model metadata diff CLI for AI gateway cost reviews

Open catalog of free-tier LLM providers, models, limits, capabilities, and quirks.

A portable visual identity and metadata language for AI models

Portable model checkpoint metadata and offline SHA-256 integrity verification — no W&B, MLflow, or tracking server.

Editable model metadata, pricing table, and budget routing notes for OpenAI-compatible gateways

Extract and normalize LLM model specifications from heterogeneous sources (config.json, GGUF/safetensors headers, LICENSE, tokenizer configs) into one unified, type-safe Pydantic schema.

DeepSeek Harness plugin to audit and correct llm-pi-ai model capability declarations by probing the endpoint itself: out-of-range maxTokens, contextWindow, reasoning levels, image input.

As ML models grow in complexity, structured, machine-readable metadata is crucial for reuse. We propose MRM3, a novel approach for storing ML model metadata integrated into a Neo4j knowledge graph.

Source-verified facts feed for commercial LLM APIs — every price, limit, and lifecycle fact carries a verbatim primary-source quote. 135 models, 16 providers, machine-readable.