Models appear in `docker model list`, but `docker model run` re-downloads

[krissetto]

I noticed this issue where some of my models that appear in docker model list are not recognized by docker model run and it tries to re-download them

OS: Ubuntu 25.10

$ docker model version
Docker Model Runner version v1.0.2-99-gca12aae1-dirty
Docker Engine Kind: Docker Engine

$ docker model status
Docker Model Runner is running

Status:
mlx: not installed
sglang: python3 not found in PATH
vllm: vLLM binary not found
diffusers: python3 not found in PATH
llama.cpp: running llama.cpp version: c55bce4

Screenshot

[ilopezluna]

Hello @krissetto that should be fixed by #617
I have just tried and it worked on:

MODEL_RUNNER_HOST=http://localhost:13434 docker model version
Docker Model Runner version v1.0.10-26-ga77d68229-dirty
Docker Engine Kind: Docker Engine (Manual Install)

MODEL_RUNNER_HOST=http://localhost:13434 docker model pull huggingface.co/unsloth/SmolLM3-3B-GGUF
b33563055168: Pull complete [==================================================>]   16.5kB/16.5kB
ee14de3d36e5: Pull complete [==================================================>]  1.915GB/1.915GB
b3e99c9e329e: Pull complete [==================================================>]  18.43kB/18.43kB
4de907d2d388: Pull complete [==================================================>]  1.915GB/1.915GB
Model pulled successfully
MODEL_RUNNER_HOST=http://localhost:13434 docker model run huggingface.co/unsloth/SmolLM3-3B-GGUF
> hello
Thinking:
Okay, the user just said "hello". I need to respond appropriately. 
[...]

MODEL_RUNNER_HOST=http://localhost:13434 docker model run hf.co/unsloth/SmolLM3-3B-GGUF
> hello
Thinking:
Okay, the user just
[...]

Could you update your local version of DMR and verify it works? 🙏