serving llama3 does not work

[lambdaofgod]

What is the issue?

I am able to run llama 3 (ollama run llama3) but when I try to run the server I get

{"error":"model 'llama3' not found, try pulling it first"}

This is in spite of ollama list detecting the model.

Specifically I ran

curl $LLAMA_URL -d '{
  "model": "llama3",      
  "messages": [
    { "role": "user", "content": "why is the sky blue?" }
  ]
}'

OS

Linux

GPU

No response

CPU

No response

Ollama version

0.1.32

[EverThingy]

Can you try adding a tag to the model parameter (by default it's latest)

curl $LLAMA_URL -d '{
  "model": "llama3:latest",      
  "messages": [
    { "role": "user", "content": "why is the sky blue?" }
  ]
}'

[lambdaofgod]

That's weird, that was the first thing I tried and it didn't work. Must have been fixed by an update.
Thanks!

[sridvijay]

I'm still running into the same issue. I've tried llama3 as well as llama3:latest and still get the same error:

model 'llama3' not found, try pulling it first.

When I run ollama list, I do see the model:

ollama list
NAME            ID              SIZE    MODIFIED    
llama3:latest   a6990ed6be41    4.7 GB  5 hours ago

I've tried reinstalling ollama as well, but no luck either :/

This is running on Ubuntu 20.04 running ollama 0.1.32

EDIT:

I solved it by running curl -X POST http://localhost:[port]/api/pull -d '{"model":"llama3"}'.

Not sure what I got wrong here, as I thought doing ollama pull llama3 in the command line would result in the same thing, but I guess the models are stored separately if you want to serve? Not sure, but it works!

[lambdaofgod]

@sridvijay this basically solves it, thanks! I've updated FAQ accordingly in #3936

[m-aliabbas]

same error. not solved