[HN Gopher] Run LLaMA and Alpaca on your computer
       ___________________________________________________________________
        
       Run LLaMA and Alpaca on your computer
        
       Author : kristianpaul
       Score  : 136 points
       Date   : 2023-04-05 17:03 UTC (5 hours ago)
        
 (HTM) web link (github.com)
 (TXT) w3m dump (github.com)
        
       | ConanRus wrote:
       | [dead]
        
       | nunobrito wrote:
       | Nice, thanks for making this available.
        
       | kevviiinn wrote:
       | Has anyone been able to get this to download the 65B model? It
       | keeps failing for Mr but 30B downloads just fine
        
       | toastal wrote:
       | Linked library has a DCMA takedown. Host this stuff outside of US
       | corporate control.
        
         | shaky-carrousel wrote:
         | If the model weights are able to be copyrighted, then they are
         | also able to be a derivative work of something copyrighted, as
         | the model is based on copyrighted works. So either the weights
         | are infringing copyright, which means they don't belong to
         | Meta, or the weights can't be copyrighted.
         | 
         | Either way, that DMCA is worthless.
        
           | hbn wrote:
           | I'm not sure the logic follows.
           | 
           | Music can be copyrighted, and it can also be made from
           | samples from other music that's copyrighted. But sampling
           | still happens without infringing copyright.
           | 
           | I'd say the small handful of bits flipped in a model from
           | training on some text, or piece of code, or an image is even
           | less copyrighted information brought over than a music sample
           | or borrowed/referenced melody.
        
           | demoride wrote:
           | Your philosophical argument is interesting, but what the op
           | was saying was one of the linked repos used in this repo is
           | inaccessible due to DMCA: https://github.com/shawwn/llama-dl
           | 
           | So while what you say may be true the DMCA seems to have
           | worth for these orgs because they can get code removed by the
           | host, who is uninterested in litigating, and the repo owner
           | likely is even less capable of litigating the DMCA.
           | 
           | Unfortunately as a tool of fear and legal gridlock DMCA has
           | shown itself to be very useful to those with ill intent.
        
           | nl wrote:
           | I'd suggest all the internet lawyers on this thread read
           | "Foundation models and Fair Use":
           | https://arxiv.org/abs/2303.15715
           | 
           | TL;DR: There's a good fair use argument but it isn't a given
           | that fair use always applies.
        
           | halfeatenscone wrote:
           | I don't think this follows. Calling it a derivative work
           | already feels like a stretch, but even granting that framing,
           | the use is clearly transformative and therefore likely to be
           | considered fair use in the US.
           | 
           | I actually wrote a Wikipedia article on the intersection of
           | copyright law and deep learning models the other day (https:/
           | /en.wikipedia.org/wiki/Artificial_intelligence_and_co...). I
           | was hoping to include a section on the copyrightability of
           | model weights, but was sadly able to find 0 coverage in
           | reliable sources.
        
             | nl wrote:
             | This should include https://arxiv.org/abs/2303.15715
        
         | sliken wrote:
         | Or just post a magnet link to a torrent.
        
           | knodi123 wrote:
           | Downloading alpaca weights actually does use a torrent now!
           | But it seems a little.... leechy... to use a fire-and-forget
           | CLI torrent client to grab a dozens of gigabytes file and
           | then quit.
        
       | Dwedit wrote:
       | llama.cpp is simpler to get working than this is.
        
         | mastazi wrote:
         | How so?
         | 
         | how is this:                   npx dalai llama 7B         npx
         | dalai serve
         | 
         | harder than:                   # build this repo         git
         | clone https://github.com/ggerganov/llama.cpp         cd
         | llama.cpp         make              #For Windows and CMake, use
         | the following command instead:         cd
         | <path_to_llama_folder>         mkdir build         cd build
         | cmake ..         cmake --build . --config Release
         | # obtain the original LLaMA model weights and place them in
         | ./models         ls ./models         65B 30B 13B 7B
         | tokenizer_checklist.chk tokenizer.model              # install
         | Python dependencies         python3 -m pip install torch numpy
         | sentencepiece              # convert the 7B model to ggml FP16
         | format         python3 convert-pth-to-ggml.py models/7B/ 1
         | # quantize the model to 4-bits (using method 2 = q4_0)
         | ./quantize ./models/7B/ggml-model-f16.bin ./models/7B/ggml-
         | model-q4_0.bin 2              # run the inference
         | ./main -m ./models/7B/ggml-model-q4_0.bin -n 128
        
       | seydor wrote:
       | I hope we will see a revival of native desktop development with
       | AI. There is really no reason for bloated JS containers, we can
       | ask AI to rewrite it in C++.
        
         | dist-epoch wrote:
         | There will be a single app in the future: a generative AI app
         | which will synthesize pixels on the screen depending on what
         | you are doing.
        
           | knodi123 wrote:
           | That will be absurdly wasteful of power. My calculator runs
           | using the power of ambient light in a comfortably lit room.
           | There are some things that an LLM is well-suited for, but
           | c'mon. Unless electricity and hardware becomes free, I just
           | don't see your prediction coming true.
        
         | brucethemoose2 wrote:
         | One might call C++ bloat. Why not C, or maybe even a stricter C
         | subset more amenable to compiler optimization.
        
         | suby wrote:
         | I hope so too. Plus, in my experience with using GPT4 to
         | generate code, it's more effective at writing C++ than it is at
         | writing Python. Something about annotating the types helps it
         | with getting things right.
        
           | hbn wrote:
           | > Something about annotating the types helps it with getting
           | things right.
           | 
           | Same rule holds true for human programmers!
        
       | nickthegreek wrote:
       | For some reason my install does not have the model selector at
       | the top. Real shame, thats the feature I wanted!
        
       | tmaly wrote:
       | The windows install says: When installing Visual Studio, make
       | sure to check the 3 options as highlighted below:
       | 
       | Python development Node.js development Desktop development with
       | C++
       | 
       | What if you already have visual studio installed and you did not
       | check all of these?
        
         | muricula wrote:
         | It's been a little while but iirc, you can rerun the visual
         | studio installer and check those boxes and it will update your
         | install.
        
       | henry_viii wrote:
       | Previous discussion:
       | 
       | https://news.ycombinator.com/item?id=35127020
        
       ___________________________________________________________________
       (page generated 2023-04-05 23:01 UTC)