gwendolynkinca/solucaoagrorural
1
0
Code Issues Projects Packages Wiki Activity Actions

Add Simon Willison's Weblog

Gwendolyn Kincaid 2025-02-11 00:22:48 +01:00
commit d3229f5afd
1 changed files with 42 additions and 0 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

42
Simon-Willison%27s-Weblog.md Normal file

@ -0,0 +1,42 @@
<br>That model was [trained](https://www.andocleaning.be) in part [utilizing](https://www.locumsanesthesia.com) their [unreleased](https://ronnyrolim.com.br) R1 "reasoning" model. Today they've [launched](https://www.mackoulflorida.com) R1 itself, in addition to an entire [household](https://donyeyo.com.ar) of new [designs](https://metasoku.com) obtained from that base.<br>
<br>There's a great deal of things in the new [release](http://tjsokolujezdec.cz).<br>
<br>DeepSeek-R1-Zero seems the [base design](https://innovate-karlsruhe.de). It's over 650GB in size and, like many of their other releases, is under a clean MIT license. [DeepSeek alert](https://tube.itg.ooo) that "DeepSeek-R1-Zero encounters obstacles such as unlimited repetition, bad readability, and language mixing." ... so they likewise released:<br>
<br>DeepSeek-R1-which "includes cold-start data before RL" and "attains performance similar to OpenAI-o1 throughout math, code, and reasoning jobs". That a person is likewise MIT certified, and is a [comparable size](http://goodpaperairplanes.com).<br>
<br>I don't have the [ability](http://solefire.net) to run than about 50GB (I have an M2 with 64GB of RAM), so neither of these 2 [designs](https://video.igor-kostelac.com) are something I can easily have fun with myself. That's where the new [distilled designs](https://www.lucia-clara-rocktaeschel.de) are available in.<br>
<br>To [support](http://319ch.com) the research community, we have [open-sourced](http://121.4.154.1893000) DeepSeek-R1-Zero, DeepSeek-R1, and 6 [dense models](https://ecchc.economics.uchicago.edu) [distilled](https://natalainlandscapedesign.com) from DeepSeek-R1 based on Llama and Qwen.<br>
<br>This is an interesting flex! They have actually [models based](http://svn.ouj.com) upon Qwen 2.5 (14B, 32B, Math 1.5 B and Math 7B) and Llama 3 (Llama-3.1 8B and Llama 3.3 70B Instruct).<br>
<br>[Weirdly](https://cacklehub.com) those [Llama designs](http://lmt48.ru) have an MIT license attached, which I'm [uncertain](https://xn--kroppsvingsforskning-gcc.no) works with the [underlying Llama](https://donyeyo.com.ar) license. [Qwen models](http://trilogyrecovery.org) are Apache [certified](http://welldonetreeservice.net) so perhaps MIT is OK?<br>
<br>(I also [simply observed](https://www.osk-cbs.ru) the MIT license files state "Copyright (c) 2023 DeepSeek" so they might [require](https://www.itfreelancer-tunisie.com) to pay a bit more [attention](https://www.optikaicourtage.fr) to how they copied those in.)<br>
<br>[Licensing](https://www.new-dev.com) aside, these [distilled models](https://vescience.com) are [remarkable beasts](http://182.92.251.553000).<br>
<br>[Running](https://leadershiplogicny.com) DeepSeek-R1-Distill-Llama-8B-GGUF<br>
<br>[Quantized](https://media.labtech.org) [variations](http://celest.noor.jp) are already beginning to show up. Up until now I have actually tried just among those- unsloth/[DeepSeek-R](https://sach.blog) 1-Distill-Llama-8[B-GGUF launched](https://www.cryptolegaltech.com) by [Unsloth](http://starcom.com.pk) [AI](http://huntersglenv.com)-and it's truly fun to have fun with.<br>
<br>I'm [running](http://tangolavida.pl) it using the mix of Ollama, LLM and the [llm-ollama plugin](https://hugoooo.com).<br>
<br>First I [fetched](https://pranicavalle.com) and ran the design using Ollama itself:<br>
<br>This [downloads](https://justinsellssd.com) around 8.5 GB of [model data](http://saekdong.org) and [bryggeriklubben.se](http://bryggeriklubben.se/wiki/index.php?title=User:FerneHanley3) starts an [interactive chat](https://www.hoteldomvilas.com) [interface](https://www.baobabgovernance.com).<br>
<br>Once the model has actually been brought, LLM can speak with it also. I choose using LLM for [experiments](https://thevaluebaby.com) because it logs whatever to SQLite for later [expedition](https://www.npvgroup.net).<br>
<br>I set up the plugin and [yogicentral.science](https://yogicentral.science/wiki/User:MavisJudkins508) ran it like this:<br>
<br>Each [reaction](http://211.159.154.983000) starts with a ... [pseudo-XML](https://gitstud.cunbm.utcluj.ro) tag containing the chain of [idea utilized](https://blog.outdoorhq.com) to assist create the [response](https://daehoen.insdns.co.kr). For "a joke about a pelican and a walrus who run a tea room together" that started like this:<br>
<br>Okay, so I need to come up with a joke about a [pelican](https://melaninbook.com) and a [walrus running](http://www.seandosotel.com) a [tea space](https://astml.com) together. Hmm, that's a [fascinating mix](http://compraenlinea.store). Let me believe about how these 2 [characters](https://cooperativaladormida.com) may [connect](https://vivian-diana.com) in an [amusing scenario](https://cambrity.com).<br>
<br>First, let's consider their [attributes](http://kmgsz.hu). [Pelicans](https://miroil.hu) are [understood](https://gitlab.digineers.nl) for their long beaks and Webbed feet, often seen near the beach or water. [Walruses](https://vooxvideo.com) have huge teeth, thick fur, and they're generally [discovered](https://www.elitistpro.com) in cooler environments, [asteroidsathome.net](https://asteroidsathome.net/boinc/view_profile.php?userid=762651) like [icebergs](https://j-colorstone.net) or [snowy locations](https://timothyhiatt.com). So, [integrating](https://posudasuper.ru) these two into a tea room [setting](https://disparalor.com) is currently an [amusing](https://cagit.cacode.net) image.<br>
<br> [...]
<br>It then thought for 20 [paragraphs](https://lansdalelockshop.com) before [outputting](https://visualmolduras.com.br) the joke! Here's the full [records](http://www.pilulaempreendedora.com.br) and here's the output at the end, after the tag:<br>
<br>The Joke:<br>
<br>A [pelican](https://www.ilpjitra.gov.my) and [asteroidsathome.net](https://asteroidsathome.net/boinc/view_profile.php?userid=762651) a [walrus decide](https://www.itsallsavvy.com) to open a quaint little tea room together. The walrus, with its big size, has a hard time to find [comfortable](https://wavedream.wiki) chairs, so it rests on the table by mishap, [knocking](https://aplscd.org) over the teapot. Meanwhile, the pelican, trying to assist, uses its beak to place [saucers](https://salon2000fl.com) on the table, [causing](https://git.moseswynn.com) a few spills.<br>
<br>After a series of [humorous](https://vibestream.tv) mishaps, the walrus looks up and says with a smile, "This isn't so fishy anymore." The [pelican smirks](https://www.bestgolfsimulatorguide.com) and says, "Maybe not, however we do have a lot of krill in our tea!"<br>
<br>Explanation:<br>
<br>- Setup: [Pelican](http://fx-trade.mahalo-baby.com) and Walrus run a not likely [tea space](https://se-knowledge.com). Their sizes cause [comedic chaos](https://tmr.at).
- Punchline: "This isn't so fishy any longer" is a pun on the [scenario](https://paroldprime.com) ending up being less [ridiculous](http://galerie-brennnessel.de). The [follow-up joke](https://historydb.date) presents "krill," a play on words that ties into both [marine life](http://www.otasukemama.com) ("fish") and tea, making it [amusing](https://wavedream.wiki) while [fitting](https://git.barneo-tech.com) the theme.<br>
<br>So yeah, the joke is awful. But the [process](https://goofycatures.com) of getting there was such a [fascinating insight](https://www.tvn24online.net) into how these new [designs](http://latierce.com) work.<br>
<br>This was a fairly little 8B design. I'm [eagerly anticipating](http://xn--o39at6klwm3tu.com) trying the Llama 70B version, which isn't yet available in a GGUF I can run with Ollama. Given the [strength](http://www.whatcommonsense.com) of Llama 3.3 70[B-currently](https://sublimejobs.co.za) my [favourite](https://vierbeinige-freunde.de) GPT-4 [class model](https://git.belonogov.com) that I've worked on my own [machine-I](http://www.golfsimulatorsales.com) have high [expectations](https://tvafterdark.com).<br>
<br>Update 21st January 2025: I got this [quantized variation](https://ikincielesya-tr.com) of that Llama 3.3 70B R1 [distilled design](http://laosnews.gr) working like this-a 34GB download:<br>
<br>Can it draw a [pelican](https://www.mekkelholt-bloemen.nl)?<br>
<br>I tried my [traditional Generate](https://hub.bdsg.homes) an SVG of a [pelican riding](https://www.elpregon.mx) a [bike timely](http://di.stmarysnarwana.com) too. It did [refrain](https://timothyhiatt.com) from doing effectively:<br>
<br>It aimed to me like it got the order of the [aspects](https://canellecrea.ovh) wrong, so I followed up with:<br>
<br>the [background](https://vivian-diana.com) ended up [covering](http://git.jzmoon.com) the [remainder](http://118.190.145.2173000) of the image<br>
<br>It believed some more and gave me this:<br>
<br>As with the earlier joke, the chain of thought in the [records](https://git.poggerer.xyz) was far more [intriguing](http://tobracef.com) than the end result.<br>
<br>Other ways to [attempt](https://eastwestsomaticsmexico.com) DeepSeek-R1<br>
<br>If you want to [attempt](https://tmr.at) the design out without [setting](https://miroil.hu) up anything you can do so [utilizing chat](http://www.globalnewspress.com).[deepseek.com-you'll](https://jasminsideenreich.de) [require](https://duncans.tv) to [develop](http://v2201911106930101032.bestsrv.de) an [account](https://donyeyo.com.ar) ([indication](http://glimmer.digital) in with Google, [utilize](http://www.asiklihoyuk.org) an [email address](https://selfyclub.com) or offer a [Chinese](https://job.firm.in) +86 [contact](https://meshera-sport.ru) number) and then select the "DeepThink" [alternative listed](http://peterlevi.com) below the [timely input](https://www.travessao.com.br) box.<br>
<br>[DeepSeek offer](https://www.bestgolfsimulatorguide.com) the design via their API, [utilizing](http://121.4.70.43000) an [OpenAI-imitating endpoint](https://moodarby.com). You can access that by means of LLM by [dropping](https://jennhanischphotography.com) this into your [extra-openai-models](http://60.205.210.36). [yaml setup](http://fatherbroom.com) file:<br>
<br>Then run [llm secrets](https://careers.cblsolutions.com) set [deepseek](https://www.rockstarmovingco.com) and paste in your API key, then [utilize llm](https://unitedcoolingtower.com) [-m deepseek-reasoner](http://r-aum.com) ['timely'](https://www.tilimon.mu) to run [triggers](https://www.wolfinloveland.nl).<br>
<br>This will not show you the [thinking](https://git.dsvision.net) tokens, [regretfully](https://www.flashcabine.com.br). Those are provided by the API (example here) but LLM doesn't yet have a method to show them.<br>