this post was submitted on 09 Nov 2023
1 points (100.0% liked)

LocalLLaMA

3 readers
1 users here now

Community to discuss about Llama, the family of large language models created by Meta AI.

founded 1 year ago
MODERATORS
 

chat.lmsys.org is great. It has the best open source models, and it let's you control temperature and other parameters.

However, they have a limit on the message length that I can send to the LLM, something like 400 words, although the model supports much longer messages.

Do you happen to know alternatives that allow longer messages?

Thanks in advance!

top 9 comments
sorted by: hot top controversial new old
[–] ortegaalfredo@alien.top 1 points 1 year ago

LLMs on neuroengine.ai should support way more than 400 words. Don't know exactly the limit.

[–] openLLM4All@alien.top 1 points 1 year ago

I'm not sure what the limit is on Text Generation UI which is fully local.

I don't think infermatic.ai has a limit either.

[–] mcmoose1900@alien.top 1 points 1 year ago

Technically the Horde doesn't have a limit, but most hosts are running 4K-8K context models:

https://lite.koboldai.net/

[–] cwl1907@alien.top 1 points 1 year ago (3 children)

Hey I'm one of the maintainers of chat.lmsys.org. We previously set this limit to avoid heavy compute but we are considering to increase it. how long is your input typically?

[–] ammar-@alien.top 1 points 1 year ago

And is there a plan to provide paid APIs for the available models that we can use programmatically, like OpenAI API?

[–] ammar-@alien.top 1 points 1 year ago

Hey, thank you for replying! 400 words per message are good for most uses, but sometimes I need longer messages like 600 words for example. It's great that chat.lmsys.org has many great models and they get updated all the time, so it would be great to be able to use these models with longer messages. Thanks!

[–] ammar-@alien.top 1 points 11 months ago (1 children)

Hey, I have a question please. Are you using the original models or quantized versions of them?

[–] cwl1907@alien.top 1 points 11 months ago (1 children)

All are original and no quantization at all, for fair comparison.

[–] ammar-@alien.top 1 points 11 months ago