60

A few words on DS4

A relevant recent tweet from antirez: https://x.com/antirez/status/2054854124848415211

> Gentle reminder on how, in the recent DS4 fiesta, not just me but every other contributor found GPT 5.5 able to help immensely and Opus completely useless.

I've noticed the same for lower level squeezing-as-much-performance-as-possible code work.

12 minutes agominimaxir

I don't see an explanation of why they would make a model-specific inference engine vs just using llamacpp. There are already lots of people working on the llamacpp integration. This is a lot of effort spent on a single model which is likely to become obsolete when a different model comes out that does better. In some discussions, people are now making PRs against both the llamacpp branches and ds4... so it's taking a rare commodity (people investing development time in this model) and fragmenting it

21 minutes ago0xbadcafebee

I thought DeepSeek was closed-weights and proprietary? I wonder how it compares against Western open-weight models. The hugging face page contains the comparison only with proprietary models for some reason.

12 minutes agocodedokode

Just want to mention that I've been pulling down and using DwarfStar locally and it's incredible. I actually have it running on my personal macbook m4 max with 128gb of ram and I am running the server to share it through tailscale with my work laptop and just have pi running there.

The long context reasoning is something I haven't even seen in frontier models - I was running at 124k tokens earlier and it was still just buzzing along with no issues or fatigue.

I am amazed at how well it works, I'm using it right now for some pretty complex frontend work, and it is much much faster than, for example running a dense 27b or 31b model (like qwen or gemma) for me (The benefits of MoE) - but the long context capabilities have been what have been absolutely flooring me.

Super excited about this project and hope Antirez can keep himself from burning out - i've been following the repo pretty closely and there are a ton of PR's flooding in and it seems like he's had to do a lot of filtering out of slop code.

32 minutes agokamranjon

Is DS4 dwarf star 4 or deep seek 4?

24 minutes agole-mark

Just updated! Sorry I meant Dwarf Star - it's the only way I've actually managed to run DeepSeek flash on my local hardware

23 minutes agokamranjon

DwarfStar 4 is DeepSeek 4 (check the repo)

23 minutes agowolttam

This is great! I feel the same way about the deepseek v4 architecture for commodity hardware.

Also have enjoyed playing with https://huggingface.co/HuggingFaceTB/nanowhale-100m-base (but early days for me understanding this space)

an hour agobjconlan

Very cool! I had no idea that HF was doing this - I really love their small model experiments.

19 minutes agokamranjon

I got this running on a 128GB M5 the other day - pretty painless, model runs in about 80GB of RAM and it seemed to be very capable at writing code and tool execution.

an hour agosimonw

How’s the token throughput / response time?

39 minutes agoperfmode

Healthy!

  prefill: 30.91 t/s, generation: 29.58 t/s
From https://gist.github.com/simonw/31127f9025845c4c9b10c3e0d8612...
36 minutes agosimonw

I don't want to be a jerk but 31t/s prefill is basically unusable in an agentic situation. A mere 10k in context and you're sitting there for 5+ minutes before the first token is generated.

26 minutes agoxienze

if it's just the coding agent system prompt and tools, you can cache that

24 minutes agoaiscoming

Yeah the problem is that's just the start of the context. There's, you know, all the tool call results and file reads and stuff.