Building on HF

57 3 57

simonko

simonko912

Simonko-912

AI & ML interests

guy with low end hardware learning about llm's and learning about finetuning my hardware: intel xeon E5v4 32gb ddr4 ram amd rx 5700 8gb vram (Hf allows to set from 6600 lol) 512gb ssd friends with some people from mradermacher contact: dc: simonko_11015 github: Simonko-912

Recent Activity

new activity about 10 hours ago

mradermacher/Qwen3-4B-Islamic-Arabic-GGUF:🚩 Report: Copyright infringement

new activity about 10 hours ago

mradermacher/Llama3.3-8B-Instruct-Thinking-Heretic-Uncensored-Claude-4.5-Opus-High-Reasoning-i1-GGUF:Unable to use with Ollama

reacted to BlueNipples's post with 👀 1 day ago

Good news, llama.cpp seems to be close to supporting MTP on qwen models. Bad news, every single gguf will have to be redone when it is.

View all activity

Organizations

New activity in mradermacher/Qwen3-4B-Islamic-Arabic-GGUF about 10 hours ago

🚩 Report: Copyright infringement

#1 opened about 10 hours ago by

NightPrince

New activity in mradermacher/Llama3.3-8B-Instruct-Thinking-Heretic-Uncensored-Claude-4.5-Opus-High-Reasoning-i1-GGUF about 10 hours ago

Unable to use with Ollama

#1 opened about 10 hours ago by

thearties

reacted to BlueNipples's post with 👀 1 day ago

Post

2777

Good news, llama.cpp seems to be close to supporting MTP on qwen models. Bad news, every single gguf will have to be redone when it is.

1 reply

New activity in social-post-explorers/README 2 days ago

question on how to join and etc

#49 opened 2 days ago by

simonko912

New activity in Guilherme34/Firefly-V3 3 days ago

new (BETTER) wrapper

#3 opened 5 days ago by

simonko912

New activity in mradermacher/BabyHercules-4x150M-GGUF 3 days ago

Interdependent Infrastructure Failure Analysis 2039

349

#9 opened 4 months ago by

mradermacher

New activity in mradermacher/model_requests 4 days ago

https://huggingface.co/Michael-Kozu/Deimos-A4

#2319 opened 4 days ago by

Michael-Kozu

https://huggingface.co/yemeni-ai-lab/gemma-4-e2b-yemeni-arabic-assistant-merged

#2318 opened 4 days ago by

yemenarchive

https://huggingface.co/DQN-Labs/dqncode2-preview2

#2317 opened 4 days ago by

DQN-Labs

New activity in mradermacher/Octen-Embedding-4B-GGUF 4 days ago

GGUF request Octen/Octen-Embedding-0.6B

#1 opened 4 days ago by

rafaelkallis

New activity in mradermacher/model_requests 4 days ago

HuggingFaceTB/nanowhale-100m

#2316 opened 4 days ago by

erenyeager-1

reacted to salma-remyx's post with 🔥 5 days ago

Post

5072

SciCrafter measured something AI practitioners have intuited: frontier agents are improving at executing inside well-framed problems, but lag at framing the problem in the first place.

GPT-5.2, Gemini-3-Pro, and Claude Opus 4.5 all plateaued near 26% on a new Minecraft benchmark for probing AI capabilities in the discovery-to-application loop.

So the authors ran targeted interventions:
* Hints about what to investigate doubled performance.
* A structured experimentation template added 7-14 more points.
* Structured consolidation beat free-form summaries by 6 points.
* Curriculum context beat independent task-solving.

These interventions helped the agent frame what’s worth investigating, and structure what gets learned so it compounds. The bottleneck for AI in scientific workflows is upstream of execution.

Their findings are congruent with the design patterns we've adopted at Remyx AI to help AI teams close the development loop scientifically.

Agents work well inside structured loops, but they perform poorly when tasked with creating the structure. Instrumenting your scientific workflows offers greater leverage than scaling compute with a less informed search.

In the work of building production AI systems, teams are flying through execution. The bigger challenge is identifying which experiments moved which production outcome, or what to try next.

One of the more interesting results I found this week by tracking work in AI for scientific workflows using Remyx: https://engine.remyx.ai/papers/d8f23b9b-b14b-4ada-b44e-ccfc221c06b4