this post was submitted on 23 Nov 2023
1 points (100.0% liked)

LocalLLaMA

1 readers
1 users here now

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

founded 10 months ago
MODERATORS
communick@poweruser.forum
1
what is the best 7b right now ? (alien.top)
submitted 10 months ago by GasBond@alien.top to c/localllama@poweruser.forum
40 comments fedilink hide all child comments
 
  1. for coding
  2. for generating stories, writing email, poems etc.
  3. good overall
  4. etc.
you are viewing a single comment's thread
view the rest of the comments
[–] akumaburn@alien.top 1 points 10 months ago

For all of the above, Tess-XS-v1.0 (Here's the updated GGUF https://huggingface.co/TheBloke/Tess-XS-v1.1-GGUF) . Nothing else I've tested at the same parameter size is quite as good, though intel's neural-chat (https://huggingface.co/TheBloke/neural-chat-7B-v3-1-GGUF) comes close. Yi-6B is unimpressive, and consistently outperformed by mistral based fine-tunes in my actual testing (Despite performing extremely well on the benchmarks). Yi-34B is in a class of its own but you asked for a 7B size model..

For stories thespis-mistral-7b (https://huggingface.co/TheBloke/Thespis-Mistral-7B-v0.6-GGUF) can be better if you're looking for NSFW.

If you're willing to step up a bit, the newly released Orca 2 13B (https://huggingface.co/TheBloke/Orca-2-13B-GGUF) drastically outperforms the above in all but NSFW content (and even then it punches well). The license isn't great however..