That model was trained in part utilizing their unreleased R1 "thinking" model. Today they've launched R1 itself, along with an entire household of brand-new models obtained from that base.
There's a great deal of stuff in the new release.
DeepSeek-R1-Zero seems the base model. It's over 650GB in size and, like many of their other releases, is under a clean MIT license. DeepSeek caution that "DeepSeek-R1-Zero encounters obstacles such as unlimited repetition, bad readability, and language blending." ... so they likewise released:
DeepSeek-R1-which "incorporates cold-start information before RL" and "attains efficiency comparable to OpenAI-o1 across math, code, and thinking jobs". That one is likewise MIT accredited, and is a comparable size.
I do not have the capability to run designs bigger than about 50GB (I have an M2 with 64GB of RAM), so neither of these 2 models are something I can easily play with myself. That's where the new distilled models are available in.
To support the research study neighborhood, forum.pinoo.com.tr we have open-sourced DeepSeek-R1-Zero, DeepSeek-R1, and six dense designs distilled from DeepSeek-R1 based on Llama and Qwen.
This is a remarkable flex! They have actually models based on Qwen 2.5 (14B, 32B, Math 1.5 B and wiki.vst.hs-furtwangen.de Math 7B) and morphomics.science Llama 3 (Llama-3.1 8B and Llama 3.3 70B Instruct).
Weirdly those Llama designs have an MIT license connected, which I'm uncertain works with the underlying Llama license. Qwen designs are Apache accredited so possibly MIT is OK?
(I also simply discovered the MIT license files state "Copyright (c) 2023 DeepSeek" so they might need to pay a little bit more attention to how they copied those in.)
Licensing aside, these distilled models are fascinating monsters.
Running DeepSeek-R1-Distill-Llama-8B-GGUF
Quantized variations are already beginning to appear. Up until now I have actually tried simply among those- unsloth/DeepSeek-R 1-Distill-Llama-8B-GGUF launched by Unsloth AI-and it's really fun to have fun with.
I'm running it utilizing the mix of Ollama, LLM and hb9lc.org the llm-ollama plugin.
First I brought and ran the design using Ollama itself:
This around 8.5 GB of design information and starts an interactive chat interface.
Once the model has actually been brought, LLM can speak to it as well. I choose using LLM for experiments because it logs everything to SQLite for later expedition.
I set up the plugin and ran it like this:
Each response starts with a ... pseudo-XML tag containing the chain of idea used to help produce the reaction. For "a joke about a pelican and a walrus who run a tea space together" that started like this:
Okay, so I need to come up with a joke about a pelican and wiki.snooze-hotelsoftware.de a walrus running a tea space together. Hmm, that's an interesting combination. Let me think of how these 2 characters may connect in a humorous scenario.
First, let's consider their attributes. Pelicans are known for their long beaks and Webbed feet, typically seen near the beach or water. Walruses have huge teeth, thick fur, and they're generally found in cooler climates, like icebergs or snowy areas. So, integrating these 2 into a tea space setting is currently an amusing image.
[...]
It then believed for 20 paragraphs before outputting the joke! Here's the complete records and here's the output at the end, shiapedia.1god.org after the tag:
The Joke:
A pelican and a walrus choose to open a quaint little tea room together. The walrus, with its big size, struggles to find comfortable chairs, so it rests on the table by accident, overturning the teapot. Meanwhile, opentx.cz the pelican, attempting to assist, utilizes its beak to position saucers on the table, causing a couple of spills.
After a series of humorous accidents, the walrus searches for and states with a grin, "This isn't so fishy anymore." The pelican smirks and remarks, "Maybe not, however we do have a great deal of krill in our tea!"
Explanation:
- Setup: Pelican and Walrus run an unlikely tea space. Their sizes lead to comical turmoil.
- Punchline: "This isn't so fishy anymore" is a pun on the situation becoming less absurd. The follow-up joke presents "krill," a play on words that ties into both marine life ("fish") and tea, making it amusing while fitting the style.
So yeah, the joke is horrible. But the procedure of getting there was such an intriguing insight into how these brand-new models work.
This was a fairly small 8B model. I'm anticipating experimenting with the Llama 70B variation, which isn't yet available in a GGUF I can keep up Ollama. Given the strength of Llama 3.3 70B-currently my preferred GPT-4 class model that I have actually worked on my own machine-I have high expectations.
Update 21st January 2025: I got this quantized version of that Llama 3.3 70B R1 distilled design working like this-a 34GB download:
Can it draw a pelican?
I attempted my traditional Generate an SVG of a pelican riding a bike timely too. It did refrain from doing very well:
It aimed to me like it got the order of the aspects wrong, so I followed up with:
the background wound up covering the remainder of the image
It thought some more and provided me this:
Just like the earlier joke, the chain of thought in the records was much more intriguing than the end outcome.
Other methods to try DeepSeek-R1
If you wish to attempt the model out without installing anything you can do so using chat.deepseek.com-you'll need to create an account (check in with Google, utilize an email address or provide a Chinese +86 contact number) and then select the "DeepThink" choice below the prompt input box.
DeepSeek offer the model through their API, utilizing an OpenAI-imitating endpoint. You can access that by means of LLM by dropping this into your extra-openai-models. yaml setup file:
Then run llm keys set deepseek and paste in your API key, then use llm -m deepseek-reasoner 'prompt' to run prompts.
This will not show you the reasoning tokens, regretfully. Those are provided by the API (example here) but LLM doesn't yet have a method to display them.