Llama 3 model not working bug

[daknow]

Llama 3 model not working. Others Models working Llama 3 not.

[n8n]

It looks like your topic is missing some important information. Could you provide the following if applicable.

• n8n version:
• Database (default: SQLite):
• n8n EXECUTIONS_PROCESS setting (default: own, main):
• Running n8n via (Docker, npm, n8n cloud, desktop app):
• Operating system:

[daknow]

Anyone read this?

[Canato]

Can you try with Basic LLM Chain first?

[rods]

I have the same problem with node “Ai Agent”! Works fine for me using llama3 on “Basic LLM Chain”,

With mistral, the AI Agent works too!

I was watching the Ollama terminal and when I selected the llama3 model and test it, I the node didn’t do the request

n8n Version: 1.39.1
Win11 - running with npx - local
The other things are default

[Jon]

Are there any error messages in n8n?

[Morris_Lu]

yep you are right ,
AI Agent >> Fail to connect to Ollama Chat Model
BUT
Basic LLM Chain >> works perfect!! w/Ollama Chat Model
the same setting but result so different so strange!!!

@rods

I have the same problem with node “Ai Agent”! Works fine for me using llama3 on “Basic LLM Chain”,

With mistral, the AI Agent works too!

I was watching the Ollama terminal and when I selected the llama3 model and test it, I the node didn’t do the request

n8n Version: 1.39.1
Win11 - running with npx - local
The other things are default

[Morris_Lu]

@Jon

Are there any error messages in n8n?

NodeOperationError: Node does not have any credentials set!
Error: Node does not have any credentials set!

BUT it works on Basic LLM Chain,
so weird

[rods]

for me, stay in infinity loading

[Niskeletor]

Hi Jon and Morris_Lu,

I’ve been experiencing a similar issue while integrating n8n with the Llama 3 model. Here’s the context:

When I use the "AI Agent" node with Llama 3, it fails to connect to the Ollama Chat Model and doesn't perform the request.
However, when I use the "Basic LLM Chain" node with the same settings, it works perfectly fine with the Ollama Chat Model.
Using Mistral with the "AI Agent" node also works without any issues.

Here are the details of my setup:

n8n Version: 1.41
Running on docker on Debian
Other settings are default

Despite having the same settings, the results are inconsistent. The “AI Agent” node fails with Llama 3 but works with OpenAI, and the “Basic LLM Chain” works with Llama 3 without any issues.

I’ve noticed that when attempting to use Llama 3, the node either goes into an infinite loading state or loops through 10 iterations without success. On the other hand, models from OpenAI work flawlessly.

Is there a specific template or configuration required to work with Llama 3 models in n8n? Additionally, it would be valuable to assign vector memory and tools to these models. Any guidance or suggestions on resolving this issue would be greatly appreciated.

Thank you!

Nis