BTW, one last thing on my wishlist (in addition to notebook mode) is prompt caching/scrolling.
I realized that the base exllamav2 backend in ooba (and not the HF hack) doesn’t cache prompts, so prompt processing with 50K+ context takes well over a minute on my 3090. I don’t know if that’s also the case in exui, as I did not try a mega context prompt in my quick exui test.
Well, it depends on the model and stuff, and how you get to that 50k+ context. If it’s a single prompt, as in “Please summarize this novel: …” that’s going to take however long it takes. But if the model’s context length is 8k, say, then ExUI is only ever going to do prompt processing on up to 8k tokens, and it will maintain a pointer that advances in steps (the configurable “chunk size”).
So when you reach the end of the model’s native context, it skips ahead e.g. 512 tokens and then you’ll only have full context ingestion again after a total 512 tokens of added context. As for that, though, you should never experience over a minute of processing time on a 3090. I don’t know of a model that fits in a 3090 and takes that much time to inference on. Unless you’re running into the NVIDIA swapping “feature” because the model doesn’t actually fit on the GPU.
BTW, one last thing on my wishlist (in addition to notebook mode) is prompt caching/scrolling.
I realized that the base exllamav2 backend in ooba (and not the HF hack) doesn’t cache prompts, so prompt processing with 50K+ context takes well over a minute on my 3090. I don’t know if that’s also the case in exui, as I did not try a mega context prompt in my quick exui test.
Well, it depends on the model and stuff, and how you get to that 50k+ context. If it’s a single prompt, as in “Please summarize this novel: …” that’s going to take however long it takes. But if the model’s context length is 8k, say, then ExUI is only ever going to do prompt processing on up to 8k tokens, and it will maintain a pointer that advances in steps (the configurable “chunk size”).
So when you reach the end of the model’s native context, it skips ahead e.g. 512 tokens and then you’ll only have full context ingestion again after a total 512 tokens of added context. As for that, though, you should never experience over a minute of processing time on a 3090. I don’t know of a model that fits in a 3090 and takes that much time to inference on. Unless you’re running into the NVIDIA swapping “feature” because the model doesn’t actually fit on the GPU.
Yi-34B-200K is the base model I’m using. Specifically the Capybara/Tess tunes.
I can squeeze 63K context on it at 3.5bpw. Its actually surprisingly good at continuing a full context story, referencing details throughout and such.
Anyway I am on linux, so no gpu swap like windows. I am indeed using it in a chat/novel style chat, so the context does scroll and get cached in ooba.