LocalLLaMA

localllama@sh.itjust.works

PostsComments

BeefAndPoultry@lemmus.orgEnglish · 5 小时前

llama.cpp in progress pull request for smart caching of MoE experts, 16% to 35% TPS boost for my RTX 2080

2

14

llama.cpp in progress pull request for smart caching of MoE experts, 16% to 35% TPS boost for my RTX 2080

BeefAndPoultry@lemmus.orgEnglish · 5 小时前

2

SirDimples@programming.devEnglish · 2 天前

GLM-5.3 Imminent drop

7

28

GLM-5.3 Imminent drop

SirDimples@programming.devEnglish · 2 天前

7

BeefAndPoultry@lemmus.orgEnglish · 2 天前

Qwen 3.8 Max (2.4T-a95b) and 27B open weights being released next week

110

Qwen 3.8 Max (2.4T-a95b) and 27B open weights being released next week

BeefAndPoultry@lemmus.orgEnglish · 2 天前

BeefAndPoultry@lemmus.orgEnglish · 1 天前

preserve_thinking fix for Gemma 4 templates

0

13

preserve_thinking fix for Gemma 4 templates

BeefAndPoultry@lemmus.orgEnglish · 1 天前

0

BeefAndPoultry@lemmus.orgEnglish · 5 天前

unsloth/DeepSeek-V4-Flash-0731-GGUF · Hugging Face

7

22

unsloth/DeepSeek-V4-Flash-0731-GGUF · Hugging Face

BeefAndPoultry@lemmus.orgEnglish · 5 天前

7

BeefAndPoultry@lemmus.orgEnglish · 6 天前

How to run Qwen 35b-a3b on 4GB to 8GB of VRAM (with 24+GB system RAM)

5

47

How to run Qwen 35b-a3b on 4GB to 8GB of VRAM (with 24+GB system RAM)

BeefAndPoultry@lemmus.orgEnglish · 6 天前

5

leanleft@lemmy.mlEnglish · 5 天前

model: deepseek-moe-16b

0

7

model: deepseek-moe-16b

leanleft@lemmy.mlEnglish · 5 天前

0

wartzarlab@lemmy.1095.meEnglish · 8 天前

Field note: measuring your context budget before a run, not after

2

0

Field note: measuring your context budget before a run, not after

wartzarlab@lemmy.1095.meEnglish · 8 天前

2

ikt@aussie.zoneEnglish · 9 天前

Kimi-K3 Releases on HuggingFace ~7 hours from this post

1

21

Kimi-K3 Releases on HuggingFace ~7 hours from this post

ikt@aussie.zoneEnglish · 9 天前

1

troed@fedia.io · 11 天前

llama rpc exists and works really well

7

24

llama rpc exists and works really well

troed@fedia.io · 11 天前

7

leanleft@lemmy.mlEnglish · 11 天前

how many models have you collected locally?

7

6

how many models have you collected locally?

leanleft@lemmy.mlEnglish · 11 天前

7

SuspiciousCarrot78@aussie.zoneEnglish · 13 天前

Grug (caveman compression) of Qwen3.6-35 and 27B is good!

8

26

Grug (caveman compression) of Qwen3.6-35 and 27B is good!

SuspiciousCarrot78@aussie.zoneEnglish · 13 天前

8

ikt@aussie.zoneEnglish · 14 天前

Kimi K3 is competitive with Fable; Kimi K3 + Fable is SoTA.

0

14

Kimi K3 is competitive with Fable; Kimi K3 + Fable is SoTA.

ikt@aussie.zoneEnglish · 14 天前

0

ikt@aussie.zoneEnglish · 14 天前

Vibecode Benchmarks

1

2

Vibecode Benchmarks

ikt@aussie.zoneEnglish · 14 天前

1

troed@fedia.io · 20 天前

Qwen 27B Q4 at usable speed on 16GB VRAM

5

31

Qwen 27B Q4 at usable speed on 16GB VRAM

troed@fedia.io · 20 天前

5

Shifu@lemmy.worldEnglish · 20 天前

Spent two weeks on a kernel that benchmarked 29x faster. End to end it's maybe 6-10%, and it's not even wired in yet.

5

16

Spent two weeks on a kernel that benchmarked 29x faster. End to end it's maybe 6-10%, and it's not even wired in yet.

Shifu@lemmy.worldEnglish · 20 天前

5

catch88@sh.itjust.worksEnglish · 22 天前

I gave my local LLM agent a face — an avatar frontend with an "emotion-beat" output contract (open-source)

sh.itjust.works

0

I gave my local LLM agent a face — an avatar frontend with an "emotion-beat" output contract (open-source)

sh.itjust.works

catch88@sh.itjust.worksEnglish · 22 天前

troed@fedia.io · 23 天前

My llama-server suddenly started error 400 on the chat template - this fixed it

0

6

My llama-server suddenly started error 400 on the chat template - this fixed it

troed@fedia.io · 23 天前

0

RainbowsInc@slrpnk.netEnglish · 26 天前

Major Performance Drop

6

9

Major Performance Drop

RainbowsInc@slrpnk.netEnglish · 26 天前

6

troed@fedia.io · 29 天前

GPU bifurcation - options

4

5

GPU bifurcation - options

troed@fedia.io · 29 天前

4

LocalLLaMA@sh.itjust.works

localllama@sh.itjust.works

You are not logged in. However you can subscribe from another Fediverse account, for example Lemmy or Mastodon. To do this, paste the following into the search field of your instance: !localllama@sh.itjust.works

Welcome to LocalLLaMA! Here we discuss running and developing machine learning models at home. Lets explore cutting edge open source neural network technology together.

Get support from the community! Ask questions, share prompts, discuss benchmarks, get hyped at the latest and greatest model releases! Enjoy talking about our awesome hobby.

As ambassadors of the self-hosting machine learning community, we strive to support each other and share our enthusiasm in a positive constructive way.

Rules:

Rule 1 - No harassment or personal character attacks of community members. I.E no namecalling, no generalizing entire groups of people that make up our community, no baseless personal insults.

Rule 2 - No comparing artificial intelligence/machine learning models to cryptocurrency. I.E no comparing the usefulness of models to that of NFTs, no comparing the resource usage required to train a model is anything close to maintaining a blockchain/ mining for crypto, no implying its just a fad/bubble that will leave people with nothing of value when it burst.

Rule 3 - No comparing artificial intelligence/machine learning to simple text prediction algorithms. I.E statements such as "llms are basically just simple text predictions like what your phone keyboard autocorrect uses, and they’re still using the same algorithms since <over 10 years ago>.

Rule 4 - No implying that models are devoid of purpose or potential for enriching peoples lives.

Visibility: Public

This community can be federated to other instances and be posted/commented in by their users.

56 users / day
226 users / week
389 users / month
855 users / 6 months
48 local subscribers
4.99K subscribers
464 Posts
2.39K Comments
Modlog

sh.itjust.works

A bilingual (EN/FR) general-purpose instance located in eastern Canada! Powered by 99% renewable energy! Everyone is welcome eh.

Useful Links

Lemmy Getting Started Guide
Official Lemmy Support
Search communities across all lemmy instance
sh.itjust.works meta discussion community: !main@sh.itjust.works
sh.itjust.works Matrix chat: https://matrix.to/#/#sh.itjust.works:matrix.org

Donations

Rules:

Be respectful. Everyone should feel welcome here.
No bigotry - including racism, sexism, ableism, homophobia, transphobia, or xenophobia.
No Ads / Spamming.
No pornography.

Règles :

Soyez respectueux. Tout le monde doit se sentir le bienvenu ici.
Pas de bigoterie - y compris le racisme, le sexisme, le capacitisme, l’homophobie, la transphobie ou la xénophobie.
Pas de publicités / Pas de spam.
Pas de pornographie.

Fediseer

Other UI options (more to come)

https://oldsh.itjust.works/

Monitoring Services

lemmy-meter.info