🤝 Open to Collab

Muhammad Umair

umair894

7 222 316

AI & ML interests

Multimodal Reidentification | Feature Upscaling | Cross-modal alignment | robust generalization | PhD UESTC

Recent Activity

liked a model about 23 hours ago

baidu/Unlimited-OCR

liked a Space 1 day ago

FINAL-Bench/POCKET-26B-CPU

upvoted a paper 1 day ago

SpatialCLI: Learning to Reason With Spatial Tools, Then Without Them

View all activity

Organizations

liked a model about 23 hours ago

baidu/Unlimited-OCR

Image-Text-to-Text • 3B • Updated 3 days ago • 2.46M • 3.68k

liked a Space 1 day ago

POCKET-26B vs Bonsai · CPU (Gemma4-based)

⚔

BONSAI vs POCKET — 26B Gemma4 MoE out-runs a 27B on CPU

upvoted 2 papers 1 day ago

SpatialCLI: Learning to Reason With Spatial Tools, Then Without Them

Paper • 2607.27703 • Published 2 days ago • 21

Metis: Memory Foundation Model

Paper • 2607.26760 • Published 3 days ago • 250

reacted to appvoid's post with 🔥 1 day ago

Post

1368

A Small Model is All You Need. Meet palmer-006 (90M)

After 3 years of experiments, we are finally releasing our flagship tiny model: **palmer-006**.

If you are building for edge hardware, SBCs (Raspberry Pi, etc.), or low-power devices, this is for you. Inspired by Andrej Karpathy's idea of a self-contained "cognitive core," we wanted to see how much power we could pack into a sub-100M parameter footprint.

🧠 **How we "Palmerized" it:**
We believe in starting our experiments with the absolute strongest baseline possible.
1. Light fine-tuning on highly curated data
2. Model merging
3. Another light fine-tuning round
4. Adjusted Mamba for maximum token speed ⚡️

⚠️ *Note: This is a foundational language model. It has not been instruction-tuned yet!*

Also, since this needs instruction tuning next to become a chat assistant—**what dataset would you recommend we use for the instruct tune?**

---
🔗 **Quick Links & Info:**

* **License:** Open for research, education, hobby, and modification! (For commercial use/hosted APIs, shoot an email to nosoyhackercodigo@gmail.com. *PS: Donators can claim a free commercial license!*)

* **Attribution:** Built using AI tech from the Technology Innovation Institute (TII).

Can't wait to see what you build at the edge. Let me know your prompt completions below! 👇

appvoid/palmer-006

10 replies

liked a Space 2 days ago

Edge Inference Benchmarks

🚀

On-Device benchmarks across devices and models.

reacted to JonnaMat's post with 🔥 2 days ago

Post

1499

🐆💨 How small can an LLM go and still run on Jetson? We benchmarked + released the whole suite on NVIDIA - LLMs, VLMs, vision, speech, embeddings: embedl/Edge-Inference-Benchmarks

What use cases do you have for those models?

6 replies

upvoted a collection 2 days ago

Edge-lm

Collection

Extremely compressed LLMs for deployment on edge devices • 9 items • Updated 11 days ago • 1

reacted to danielhanchen's post with ❤️ 2 days ago

Post

3549

Kimi K3 can now be run locally! ✨

The 1-bit model retains ~78.9% accuracy after we shrunk it from 1.56TB to 594GB (-62% size).

Run on a Mac Studio connected with 128GB RAM device. Kimi K3 is the strongest open model to date.

GGUF: unsloth/Kimi-K3-GGUF
Guide: https://unsloth.ai/docs/models/kimi-k3

5 replies

upvoted a paper 2 days ago

TurboVLA: Real-Time Vision-Language-Action Model at 32 Hz on an RTX 4090 with <1 GB VRAM

Paper • 2607.27205 • Published 3 days ago • 125

upvoted a paper 3 days ago

JarvisHub: An Open Harness for Canvas-Native Multimodal Creative Agents

Paper • 2607.23588 • Published 6 days ago • 121

upvoted a paper 8 days ago

ReferTrack: Referring Then Tracking for Embodied Visual Tracking

Paper • 2607.20061 • Published 10 days ago • 53

reacted to SeaWolf-AI's post with 👍 9 days ago

Post

4020

📱 POCKET — a 35-billion-parameter model that runs on your iPhone, and on your PC with no GPU

We're releasing POCKET, VIDRAFT's flagship Darwin-36B-Opus compressed for on-device use. No fork, no CUDA, no cloud — it runs on stock llama.cpp. It's a sparse Mixture-of-Experts model (256 experts, only 8 active per token), so the file can be large while the work per token stays small. That's what lets a 35B model run on a phone, and generate fast on a CPU with no graphics card.

Measured (POCKET-35B IQ1_M vs Bonsai-27B Q1_0):
• CPU generate (Xeon, 16 threads): 27.0 vs 10.1 tok/s → 2.69× faster
• GPU generate (H100): 197 vs 89 tok/s → 2.22× faster
• GPU prompt processing (H100): 753 vs 1816 → 0.41× (Bonsai wins this one — MoE prefill wakes every expert, so sparsity stops helping there. We say so.)
• Quality (HellaSwag, 400 q): 61.0% vs 60.0% → a tie (confidence intervals overlap)

On a real consumer laptop — MacBook M3 Pro (18 GB) — POCKET wins every axis, prompt processing included:
• Metal generate: 25.4 vs 12.8 → 1.99×
• CPU generate: 13.8 vs 4.4 → 3.13×
• Metal prompt: 240.7 vs 73.4 → 3.28×

One more quiet fact: the same-size, quality-oriented rival Ternary-Bonsai-27B (7.2 GB) fails to load in upstream llama.cpp at all — it needs the PrismML fork. POCKET runs on the tools you already have: LM Studio, Ollama, PocketPal, MLX.

📖 Full story (tech, measurements, recipes): https://huggingface.co/blog/FINAL-Bench/pocket

Models:
📦 POCKET-35B-GGUF (PC / server, no GPU): FINAL-Bench/POCKET-35B-GGUF
🇰🇷 POCKET-KR-GGUF (Android): FINAL-Bench/POCKET-KR-GGUF
🍎 POCKET-KR-MLX (iPhone / Mac): FINAL-Bench/POCKET-KR-MLX
🌍 POCKET-EN-GGUF (English phone / PC): FINAL-Bench/POCKET-EN-GGUF
🖥️ Live demo (answering on a CPU, no GPU): FINAL-Bench/POCKET-35B-CPU
📚 Collection: FINAL-Bench/pocket-models-6a618ee5d23eafb7e185a5c6