Hugging Face – Posts

Join the conversation

Join the community of Machine Learners and AI enthusiasts.

All HF Hub posts

posted an update 3 days ago

Post

14763

🇮🇳 Qwen3-4B Hindi Instruct v2 — a Hindi LLM that runs on your own machine
Most strong Hindi-capable models are either huge or cloud-only. I wanted one that's small enough to run locally but actually follows instructions in Hindi — so I fine-tuned Qwen3-4B on 10K Hindi instruction pairs and shipped it with a full GGUF quant ladder.
✅ Fine-tune (16-bit): huggingface.co/pankajpandey-dev/Qwen3-4B-Hindi-Instruct-v2
✅ GGUF (Q4/Q5/Q8): huggingface.co/pankajpandey-dev/Qwen3-4B-Hindi-Instruct-v2-GGUF
Runs in Ollama, llama.cpp, and LM Studio. The Q4_K_M is just 2.5 GB — fits comfortably on a laptop, CPU or GPU.
Part of my Hindi LLM Series — building openly-licensed Indic models for local and edge use. More coming (Gemma next). Feedback welcome 🙏
#Hindi #IndicNLP #GGUF #LocalLLM #Qwen

4 replies

FlameF0X

posted an update about 10 hours ago

Post

1950

MiniMax-M3 coming soon.
https://github.com/MiniMax-AI/MiniMax-M3

sergiopaniego

posted an update 1 day ago

Post

2785

new banger blog alert 🚨

@ariG23498 is starting a blog series about profiling in pytorch and part 1 just dropped

takes you from the simplest scenario to actually knowing what your gpu is doing. if you have never opened a profiler trace this is where you start

covers torch.profiler from scratch. reading tables and traces, overhead bound vs compute bound, the full dispatch chain from python to gpu kernels, and what torch.compile is actually fusing under the hood

find it here: https://huggingface.co/blog/torch-profiler

evalstate

posted an update about 2 hours ago

Post

Hugging Face MCP Server v0.3.17
~~~~~~~~~~~~~~~~~~~~~~~~~~~~

SEP-2640 "Skills Over MCP" support added (early access)

RiverRider

posted an update about 4 hours ago

Post

This is not the end of words. It is the end of pretending their meanings are determined.

Meaning Forks. SRT detects it.

Paste any text to identify contested terms

RiverRider/srt-introspect

Try any prompt (attached link) to see exactly what an LLM is thinking at every meaningful step of its answer

RiverRider/srt-introspect

Repository

https://github.com/space-bacon/SRT

Paper

https://github.com/space-bacon/SRT/blob/main/paper_nla.md

Explainer

https://github.com/space-bacon/SRT/blob/main/docs/EXPLAINERS.md

ovi054

posted an update about 6 hours ago

Post

Qwen Image Edit 2511 Fast + LoRA ⚡

ovi054/Qwen-Image-Edit-2511-LoRA

QIE-2511 is an image editing model with integrated LoRA capabilities. You can add any custom LoRA to generate and edit images within this Space.

👉 Try it now: ovi054/Qwen-Image-Edit-2511-LoRA

kanaria007

posted an update about 6 hours ago

Post

✅ Article highlight: *Deployment & Rollback Governance for Learning Worlds* (art-60-169, v0.1)

TL;DR:
This article argues that deployment is the highest-risk moment in a learning world.

Training produces a new policy. Deployment turns that policy into an institution inside the world. So rollout cannot be treated like a casual model swap. It needs deploy-gate contracts, canaries, phased rollout, kill-switches, rollback receipts, and explicit non-interference rules that stop “better learning” from silently rewriting world reality.

Read:
kanaria007/agi-structural-intelligence-protocols

Why it matters:
• treats deployment as governed change, not routine ops
• prevents silent reality drift when a newly trained policy changes world outcomes
• binds rollout to safety envelopes, evaluation validity, performance SLOs, and canon boundaries
• makes rollback and emergency stop part of the formal operating contract

What’s inside:
• a *model deploy gate contract* that defines when a learned policy may enter the world
• canary and phased rollout as explicit governed stages
• kill-switch and rollback receipts for emergency containment
• non-interference audits so training and deployment do not rewrite canon or governance outcomes
• appeal and publication boundaries for claims like “we deployed safely” or “we rolled back successfully”

Key idea:
Do not say:

*“we trained a better model, so we deployed it.”*

Say:

*“this policy entered the world under this deploy gate, this rollout stage, these envelope and SLO checks, these rollback guarantees, and these receipts.”*

That is how deployment becomes governance with receipts.

evijit

posted an update 1 day ago

Post

153

Weekend mini project! Since commentary on AI is inherently interdisciplinary, we connected the observations in the Pope's encyclical with decades of scholarship in Responsible AI and Ethics research and created an interactive space with these annotations!

Work with @IJ-Reynolds , @yjernite , and @meg

Lots to unpack. We started with 105 annotations. Please submit pull requests for more that we may have missed!

society-ethics/annotated-encyclical

lbourdois

posted an update 1 day ago

Post

193

New blog post!
An introduction to a little-known but highly effective model reduction method: 𝗧𝗿𝗶𝗺𝗺𝗶𝗻𝗴✂️
We show how to reduce model size (we went up to 87.24% reduction) while preserving its performance.

We applied this technique to 16 different model families across several modalities to illustrate that it works on any architecture (as long as the embedding layer is the last one of the model) and on any modality involving text.
From these 16 families, we generated over 𝟱,𝟱𝟬𝟬 𝗺𝗼𝗻𝗼𝗹𝗶𝗻𝗴𝘂𝗮𝗹 𝗺𝗼𝗱𝗲𝗹𝘀 𝗶𝗻 𝟭𝟮𝟰 𝗱𝗶𝗳𝗳𝗲𝗿𝗲𝗻𝘁 𝗹𝗮𝗻𝗴𝘂𝗮𝗴𝗲𝘀 🌍

Key takeaways from our experiments:
1️⃣ Trimming does not require a GPU. Our models were obtained on a CPU.
2️⃣ This method scales up to at least 4B parameters (we did not test beyond that).
3️⃣ Trimmed model is smaller than the original while preserving its performance. If you observe a slight performance drop, just fine-tuned to recover or even surpass the original performance.
4️⃣ For an equivalent compute budget, it is better to trim then fine-tune rather than fine-tuning the original model. Since the model is smaller, you can run more epochs/show more data and get in fine a better model than the original.
5️⃣ Trimming is a competitive alternative to distillation and quantization. E.g. we obtained our alternative to DistilBERT in 9 minutes on CPU vs. 90 hours of GPU for the latter.
6️⃣ Trimming could generate reasoning traces in the language of the trimmed model. This could be an alternative to generating traces in English and then translating them into the desired language.

And many other things (such as how much data are needed, the impact of the database used, the order in which it should be done, etc.) are available in the blogpost!

Blogpost: https://huggingface.co/blog/lbourdois/introduction-to-trimming
Models: alphaedge-ai/Trimming_models_search

1 reply

RakshitAralimatti

posted an update 1 day ago

Post

Reading engineering and research blogs from OpenAI, Anthropic, DeepMind, Meta and others has genuinely leveled up my understanding of AI systems and helped me in my day-to-day work. But keeping track of 20+ sites manually is a pain.

So I built AI Blogs Tracker — a Streamlit app that scrapes the actual blog listing pages (not search) of 20+ top AI companies and surfaces titles, dates, and links in one clean feed. Filter by source, by date, star posts to a reading list, or add your own custom sources.

One click. ~30 seconds. Everything in one place.

🔗 GitHub link - https://github.com/rakshit2020/Tech-Blogs-Tracker-of-Top-AI-Companies-Agent

1 reply

Recently active users