Update

Signed-off-by: Michael Yuan <[email protected]>

docs/llamaedge_vs_ollama.md

@@ -10,7 +10,6 @@ choose LlamaEdge over them?
 * LlamaEdge is very small. The entire runtime and application is only 30MB. That is about 1/3 of the nearest competitor.
 * LlamaEdge does not need root or sudo permissions. It does not install or run any daemon on your system. Hence LlamaEdge can be easily embedded into your own app.
 * LlamaEdge apps are cross-platform. A single binary file can run on all supported OSes, CPUs, and GPUs. That also makes it simple to embed LlamaEdge in your apps.
-* LlamaEdge works with model files you download from Huggingface. There is no need for a special download hub.
 * Through Docker integration, an LlamaEdge container combines model files, configurations, and runtime into a single package ensuring compatibility and portability over time. All from the Docker Hub you already use.
 * LlamaEdge supports alternative runtimes beyond llama.cpp to achieve the most optimal performance for your model and hardware.
 * LlamaEdge already supports multimodal vision models. It will soon support speech-to-text and text-to-image models through as OpenAI-compatible APIs.