Skip Navigation

InitialsDiceBearhttps://github.com/dicebear/dicebearhttps://creativecommons.org/publicdomain/zero/1.0/„Initials” (https://github.com/dicebear/dicebear) by „DiceBear”, licensed under „CC0 1.0” (https://creativecommons.org/publicdomain/zero/1.0/)M

morrowind

@ morrowind @lemm.ee

Posts
10
Comments
15
Joined
1 yr. ago

  • Trump administration reportedly considers a US DeepSeek ban | TechCrunch

    Jump

  • Such dumbasses, even if this was a good strategy, they're still banning one company and let others (arguably more dangerous ones) go scot free

  • Societal rules

    Jump

  • Alright, I'm waiting on the youtube playlist

  • Dolphin: A Large-Scale Automatic Speech Recognition Model for Eastern Languages

    Jump

  • Technically it supports fewer languages than whisper, 40 vs 99

    The main problem isn't "bother", it's training data. You need hundreds of thousands of hours of high quality transcripts to train models like these and that just doesn't exist for like zulu or whatever

  • LocalLLaMA @sh.itjust.works
    morrowind @lemm.ee

    Dolphin: A Large-Scale Automatic Speech Recognition Model for Eastern Languages

    arxiv.org /abs/2503.20212

  • Sentence transformers v4

    Jump

  • I want to clarify something. Reranker is a general term that can refer to any model used for reranking. It is independent of implementation.

    What you refer to

    because reranker models look at the two pieces of content simultaneously and can be fine tuned to the domain in question. They shouldn't be used for the initial retrieval because the evaluation time is O(n²) as each combination of input

    Is a specific implementation known as CrossEncoder that is common for reranking models but not retrieval ones for the reasons you described. But you can also use any other architecture

  • Zoomers & Boomers are the same

    Jump

  • On god

  • LocalLLaMA @sh.itjust.works
    morrowind @lemm.ee

    Sentence transformers v4

  • Some updates on community changes and future goals (03-28-2025)

    Jump

  • Thumbnail looks a little odd when small. You may want to go for a more digital llama aesthetic

  • LocalLLaMA @sh.itjust.works
    morrowind @lemm.ee

    NotaGen: Advancing Musicality in Symbolic Music Generation with Large Language Model Training Paradigms

    electricalexis.github.io /notagen-demo/

  • StarVector - a foundation model for generating svgs

    Jump

  • autotracers can't generate svgs from text

  • StarVector - a foundation model for generating svgs

    Jump

  • Claude frequently draws svgs to illustrate things for me (I'm guessing it's in the prompt) but even though it's better at it than all the other models, it still kinda sucks. It's just fudamentally dumb task to do for a purely language model, similar to the arc-agi benchmark , just makes more sense for a vision model and trying to get an llm to do is a waste

  • LocalLLaMA @sh.itjust.works
    morrowind @lemm.ee

    StarVector - a foundation model for generating svgs

    huggingface.co /starvector/starvector-1b-im2svg

  • EXAONE Deep ━ Setting a New Standard for Reasoning AI - LG AI Research News

    Jump

  • what is the license? The link on hf just 404s

  • LocalLLaMA @sh.itjust.works
    morrowind @lemm.ee

    EXAONE Deep ━ Setting a New Standard for Reasoning AI - LG AI Research News

    www.lgresearch.ai /news/view

  • Sketch-of-Thought: Efficient LLM Reasoning with Adaptive Cognitive-Inspired Sketching

    Jump

  • Very similar to chain of draft but seems more thorough

  • LocalLLaMA @sh.itjust.works
    morrowind @lemm.ee

    Sketch-of-Thought: Efficient LLM Reasoning with Adaptive Cognitive-Inspired Sketching

    arxiv.org /abs/2503.05179
  • LocalLLaMA @sh.itjust.works
    morrowind @lemm.ee

    Sorting-Free GPU Kernels for LLM Sampling

    flashinfer.ai /2025/03/10/sampling.html

  • Reka Flash, open source 21B model comparable to QWQ 32B

    Jump
  • LocalLLaMA @sh.itjust.works
    morrowind @lemm.ee

    Reka Flash, open source 21B model comparable to QWQ 32B

  • Qwen/QwQ-32B · Hugging Face

    Jump

  • It matches R1 in the given benchmarks. R1 has 671B params (36 activated) while this only has 32

  • Qwen/QwQ-32B · Hugging Face

    Jump

  • insane, absolutely insane

  • LocalLLaMA @sh.itjust.works
    morrowind @lemm.ee

    Chain of Draft: Thinking Faster by Writing Less

    arxiv.org /abs/2502.18600
  • LocalLLaMA @sh.itjust.works
    morrowind @lemm.ee

    Atom of Thoughts (AOT): lifts gpt-4o-mini to 80.6% F1 on HotpotQA, surpassing o3-mini and DeepSeek-R1

    bsky.app /profile/sungkim.bsky.social/post/3ljgwfe3flk2h

  • Alibaba Releases Advanced Open Video Model, Immediately Becomes AI Porn Machine

    Jump

  • good luck trying to run a video model locally

    Unless you have top tier hardware