That model was trained in part using their unreleased R1 "thinking" design. Today they have actually released R1 itself, together with a whole household of new designs obtained from that base.
There's an entire lot of things in the brand-new release.
DeepSeek-R1-Zero appears to be the base design. It's over 650GB in size and, like the majority of their other releases, is under a clean MIT license. DeepSeek caution that "DeepSeek-R1-Zero experiences challenges such as limitless repeating, poor readability, and language mixing." ... so they likewise launched:
DeepSeek-R1-which "includes cold-start data before RL" and "attains performance similar to OpenAI-o1 across math, code, and reasoning jobs". That one is also MIT accredited, and is a similar size.
I do not have the capability to run designs larger than about 50GB (I have an M2 with 64GB of RAM), so neither of these 2 designs are something I can quickly have fun with myself. That's where the new distilled models are available in.
To support the research study community, we have open-sourced DeepSeek-R1-Zero, bytes-the-dust.com DeepSeek-R1, and six dense models distilled from DeepSeek-R1 based on Llama and Qwen.
This is a fascinating flex! They have actually designs based on 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).
Weirdly those Llama models have an MIT license connected, which I'm uncertain is compatible with the underlying Llama license. Qwen models are Apache certified so perhaps MIT is OK?
(I also just discovered the MIT license files say "Copyright (c) 2023 DeepSeek" so they might need to pay a little bit more attention to how they copied those in.)
Licensing aside, king-wifi.win these distilled designs are fascinating beasts.
Running DeepSeek-R1-Distill-Llama-8B-GGUF
Quantized versions are already starting to appear. Up until now I've attempted just among those- unsloth/DeepSeek-R 1-Distill-Llama-8B-GGUF released by Unsloth AI-and it's truly enjoyable to have fun with.
I'm running it utilizing the combination of Ollama, LLM and the llm-ollama plugin.
First I and king-wifi.win ran the model utilizing Ollama itself:
This downloads around 8.5 GB of model data and begins an interactive chat interface.
Once the design has been fetched, LLM can speak to it also. I choose using LLM for experiments since it logs whatever to SQLite for later exploration.
I installed the plugin and ran it like this:
Each action begins with a ... pseudo-XML tag containing the chain of idea used to help generate the action. For "a joke about a pelican and a walrus who run a tea space together" that began like this:
Okay, so I require to come up with a joke about a pelican and a walrus running a tea room together. Hmm, that's a fascinating combination. Let me consider how these two characters might communicate in a funny circumstance.
First, ratemywifey.com let's consider their qualities. Pelicans are understood 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 environments, like icebergs or snowy locations. So, combining these two into a tea room setting is already an amusing image.
[...]
It then thought for 20 paragraphs before outputting the joke! Here's the full transcript and here's the output at the end, after the tag:
The Joke:
A pelican and a walrus choose to open a quaint little tea room together. The walrus, with its huge size, has a hard time to find comfy chairs, so it sits on the table by accident, knocking over the teapot. Meanwhile, townshipmarket.co.za the pelican, attempting to assist, utilizes its beak to put saucers on the table, triggering a few spills.
After a series of comical mishaps, the walrus searches for and says with a smile, "This isn't so fishy any longer." The pelican smirks and says, "Maybe not, however we do have a great deal of krill in our tea!"
Explanation:
- Setup: Pelican and Walrus run a not likely tea room. Their sizes result in comedic mayhem.
- 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 theme.
So yeah, the joke is awful. But the process of arriving was such a fascinating insight into how these new designs work.
This was a fairly small 8B design. I'm looking forward to checking out 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've worked on my own machine-I have high expectations.
Update 21st January 2025: forum.pinoo.com.tr I got this quantized variation of that Llama 3.3 70B R1 distilled model working like this-a 34GB download:
Can it draw a pelican?
I tried my classic Generate an SVG of a pelican riding a bike timely too. It did refrain from doing really well:
It aimed to me like it got the order of the elements incorrect, so I followed up with:
the background ended up covering the remainder of the image
It believed some more and provided me this:
Just like the earlier joke, the chain of believed in the transcript was much more fascinating than completion outcome.
Other methods to attempt DeepSeek-R1
If you desire to attempt the model out without setting up anything at all you can do so utilizing chat.deepseek.com-you'll need to develop an account (check in with Google, utilize an email address or supply a Chinese +86 phone number) and then select the "DeepThink" alternative below the prompt input box.
DeepSeek offer the model through their API, using an OpenAI-imitating endpoint. You can access that through LLM by dropping this into your extra-openai-models. yaml configuration file:
Then run llm keys set deepseek and paste in your API key, then utilize llm -m deepseek-reasoner 'prompt' to run prompts.
This will not show you the reasoning tokens, championsleage.review sadly. Those are dished out by the API (example here) however LLM does not yet have a method to display them.