11–20 of 96 posts
From all the experimentation I've done, 7B parameter models just don't seem to be able to produce useful output reliably enough for my use cases. What use cases do people have for these smaller LLM's?
From all the experimentation I've done, 7B parameter models just don't seem to be able to produce useful output reliably enough for my use cases. What use cases do people have for these smaller LLM's?
The main use case is that it's probably the only size consumers can run on their personal devices. If you don't want your data going into an external platform like OpenAI it's the only solution even if it's not very usuable.
when will the llm race peak? have we peaked already?
From all the experimentation I've done, 7B parameter models just don't seem to be able to produce useful output reliably enough for my use cases. What use cases do people have for these smaller LLM's?
Bert had lots of use cases and this one is supposedly stronger model.
Earlier quoted context omitted.
Bert had lots of use cases and this one is supposedly stronger model.
I've used BERT in a number of production apps, it feels like a very apples to oranges comparison given how the AI landscape has changed since BERT's release.
Earlier quoted context omitted.
The main use case is that it's probably the only size consumers can run on their personal devices. If you don't want your data going into an external platform like OpenAI it's the only solution even if it's not very usuable.
These are constraints but not a use case.
From all the experimentation I've done, 7B parameter models just don't seem to be able to produce useful output reliably enough for my use cases. What use cases do people have for these smaller LLM's?
If you're relying on prompting for the 7B models IMO you're gonna have a bad time — they're mostly toys at that: interesting output but not consistently useful. But finetuning gets better results, and it's cheap to finetune.
Earlier quoted context omitted.
These are constraints but not a use case.
Isn't it obvious they're referring to the use case of using a model given those constraints?