HackerTrans
TopNewTrendsCommentsPastAskShowJobs

waldfee

no profile record

comments

waldfee
·3 tahun yang lalu·discuss
i am only dabbling in this space myself, so can't answer everything. all the formats i mentioned are for a quantized version of the original model. basically a lower resolution version, with the associated precision loss. e.g. original model weights are in f16, the gptq version is in int4. a big difference in size but often an acceptable loss of quality. using quants is basically a tradeoff between quality and "can i run it?".

examples of original models are llama(2), mistral, xwin. they are not directly related to any quantized versions. quants are mostly done by third parties (e.g. thebloke[1]).

using a full model for inference requires pretty beefy hardware. most inference on consumer hardware is done with quantized versions for that reason.

[1] https://huggingface.co/TheBloke
waldfee
·3 tahun yang lalu·discuss
i don't think such a guide exists. this space is moving pretty fast. a short rundown

quantized model formats:

- GGML: used with llama.cpp, outdated, support is dropped or will be soon. cpu+gpu inference

- GGUF: "new version" of the GGML file format, used with llama.cpp. cpu+gpu inference. offers 2-8bit quantization

- GPTQ: pure gpu inference, used with AutoGPTQ, exllama, exllamav2, offers only 4 bit quantization

- EXL2: pure gpu inference, used with exllamav2, offers 2-8bit quantization

here[1] is a nice overview of VRAM usage vs perplexity of different quant levels (with the example of a 70b model in exl2 format)

[1] https://old.reddit.com/r/LocalLLaMA/comments/178tzps/updated...
waldfee
·5 tahun yang lalu·discuss
there are modern products in this niche, and there is huge interest. the market certainly seems to be there (regium tried to defraud people for almost a million dollars i believe was the kickstarter sum before they got shut down).

there is squareoff [0], with new products currently in development (swap / neo)

then there was regium, an elaborate scam on kickstarter [1]

now there is phantom [2], which hopefully is not a scam. they at least posted some engineering details on hackaday [3]

squareoff has chess.com support, hopefully with lichess support coming (they are promising it, but has not yet happend). phantom claims working lichess support and to work on chess.com support

[0] https://squareoffnow.com

[1] https://www.chess.com/news/view/update-on-regium-chess

[2] https://www.kickstarter.com/projects/wondersubstance/phantom...

[3] https://hackaday.io/project/179268
waldfee
·5 tahun yang lalu·discuss
check out KoReader [0] from the F-Droid Store, imho it's way better than the built-in onyx reader app.

[0] https://koreader.rocks
waldfee
·6 tahun yang lalu·discuss
Sorry, can't help there as I don't need it, therefore never tried to get it to work. I think the microg subreddit [0] is your best bet for pointers

[0] https://old.reddit.com/r/MicroG/
waldfee
·6 tahun yang lalu·discuss
I have used these images for more than a year now, runs perfectly fine. Use Aurora Store (from F-Droid) if you need any Google Play apps installed.

Be careful if you rely on SafetyNet, seems to be a pain to get working correctly.

If you use the GCM registration, but push notifications just dont work, try this [0], worked for me

[0] https://github.com/microg/GmsCore/issues/226#issuecomment-26...