Backend gets 500 from Ollama that works fine

[lambdaofgod]

Describe the bug

I get

perplexica-backend-1   | error: Unhandled Rejection at: [object Promise], reason: Error: Ollama call failed with status code 500: llama runner process no longer running: -1

To Reproduce
Launch Perplexica with OLLAMA = "http://host.docker.internal:11434" in config, launch ollama and configure to use it

Additional context

Perplexica sees ollama, because I can choose it in settings, it also correctly detects the model that is running.

I checked ollama with

curl http://localhost:11434/api/generate -d '{
  "model": "aya",
  "prompt":"Why is the sky blue?"
}'

And it works fine.

[ItzCrazyKns]

It seems to work fine for me, maybe checkout ollama/ollama#3774