Based on the image, there is a person standing outdoors in front of a boat. The individual is wearing a brown hoodie with "GANDER MTN." written on it, blue jeans, and brown shoes. They are holding two large fish, one in each hand, by their mouths. The fish appear to be salmon, given their size and distinctive coloration. The background includes a boat on a trailer and some grassy areas.
Javed Alam PRO
Javedalam
AI & ML interests
Ai user, tester, and developer of open source ai based applications.
Recent Activity
liked a model about 7 hours ago
Comfy-Org/MiniMax-Music-3 liked a model about 7 hours ago
MiniMaxAI/MiniMax-Music3 liked a Space 1 day ago
NodeLinker/Qwen-3.8-27B-H200Organizations
replied to their post 1 day ago
replied to their post 1 day ago
The image shows a man standing outdoors with a group of people behind him. The background features colorful, striped mountains with hues of red, green, yellow, and purple. The sky is overcast with clouds, and there appears to be mist or low-hanging clouds around the mountain peaks. The man in the foreground is wearing a dark jacket over a gray T-shirt that has text and graphics on it, which include the words "Western Pennsylvania Mushroom Club." He has short, light-colored hair and is wearing glasses. The overall setting suggests a hiking or outdoor gathering in a scenic, possibly volcanic, landscape.
posted an update 1 day ago
Post
1222
A 3B Vision Model Running at 13 Tokens/Second — on a Phone
AI is moving from the cloud into your pocket.
I just tested Liquid AI's newly released LFM2.5-VL-3B locally on a Samsung Galaxy S26. The Q8_0 model ran through llama.cpp with Vulkan acceleration on the Qualcomm Adreno 840 GPU. I uploaded a real webpage screenshot from another device and asked the model to describe it.
It worked—and generated at 13.11 tokens per second.
That's significant for a 3B-class vision-language model running entirely on a phone. In my testing, LFM2.5-VL-3B is the most capable local vision model I've run on the S26 so far.
Liquid AI is building specifically for this world. Its focus is edge AI: capable models designed to run locally on phones, laptops, vehicles and other resource-constrained hardware instead of depending entirely on cloud inference. Its LFM family targets low memory usage, fast inference and deployment across CPUs, GPUs and NPUs.
LFM2.5-VL-3B brings vision into that strategy. It is an open-weight multimodal model capable of image understanding, OCR, document extraction, visual question answering and spatial reasoning. Liquid provides the weights publicly, including an official GGUF release for llama.cpp.
This is where open-weight AI gets interesting. A phone can now hold the model, process the image and generate useful multimodal responses locally at interactive speed.
No cloud GPU. No remote inference API. The phone is the AI computer.[LFM2.5-VL-3B — Hugging Face]( LiquidAI/LFM2.5-VL-3B)
[LFM2.5-VL-3B-GGUF — Hugging Face]( LiquidAI/LFM2.5-VL-3B-GGUF)
AI is moving from the cloud into your pocket.
I just tested Liquid AI's newly released LFM2.5-VL-3B locally on a Samsung Galaxy S26. The Q8_0 model ran through llama.cpp with Vulkan acceleration on the Qualcomm Adreno 840 GPU. I uploaded a real webpage screenshot from another device and asked the model to describe it.
It worked—and generated at 13.11 tokens per second.
That's significant for a 3B-class vision-language model running entirely on a phone. In my testing, LFM2.5-VL-3B is the most capable local vision model I've run on the S26 so far.
Liquid AI is building specifically for this world. Its focus is edge AI: capable models designed to run locally on phones, laptops, vehicles and other resource-constrained hardware instead of depending entirely on cloud inference. Its LFM family targets low memory usage, fast inference and deployment across CPUs, GPUs and NPUs.
LFM2.5-VL-3B brings vision into that strategy. It is an open-weight multimodal model capable of image understanding, OCR, document extraction, visual question answering and spatial reasoning. Liquid provides the weights publicly, including an official GGUF release for llama.cpp.
This is where open-weight AI gets interesting. A phone can now hold the model, process the image and generate useful multimodal responses locally at interactive speed.
No cloud GPU. No remote inference API. The phone is the AI computer.[LFM2.5-VL-3B — Hugging Face]( LiquidAI/LFM2.5-VL-3B)
[LFM2.5-VL-3B-GGUF — Hugging Face]( LiquidAI/LFM2.5-VL-3B-GGUF)
- 2 replies
reacted to OzTianlu's post with 🔥 6 months ago
Post
1757
Scaling UP in Kai! 🌊
NoesisLab/Kai-3B-Instruct
Introducing NoesisLab/Kai-3B-Instruct What happens when you force a 3B model to reason entirely in its latent space ?
Meet Kai-3B, our latest industrial-grade reasoning model fine-tuned using the Adaptive Dual Search (ADS) algorithm.
GSM8K (0-shot, Direct Answer): 39.27% 🤯 (Llama-2-7B is ~14.6%)
HumanEval (Pass@1): 39.02% 💻 (Overtakes Gemma-2-2B's 30%)
MMLU (5-shot): 53.62% 📚 (Crushing the 50% barrier)
ARC-Challenge: 51.88%🎯
PIQA: 77.53%
HellaSwag: 69.53%
Kai-3B proves that reasoning density doesn't strictly require parameter bloat or verbose generation. It acts as a perfect, cold-blooded Agent action-engine—ideal for JSON routing, SWE-bench patch generation, and anywhere you need absolute structured certainty without token waste.
NoesisLab/Kai-3B-Instruct
Introducing NoesisLab/Kai-3B-Instruct What happens when you force a 3B model to reason entirely in its latent space ?
Meet Kai-3B, our latest industrial-grade reasoning model fine-tuned using the Adaptive Dual Search (ADS) algorithm.
GSM8K (0-shot, Direct Answer): 39.27% 🤯 (Llama-2-7B is ~14.6%)
HumanEval (Pass@1): 39.02% 💻 (Overtakes Gemma-2-2B's 30%)
MMLU (5-shot): 53.62% 📚 (Crushing the 50% barrier)
ARC-Challenge: 51.88%🎯
PIQA: 77.53%
HellaSwag: 69.53%
Kai-3B proves that reasoning density doesn't strictly require parameter bloat or verbose generation. It acts as a perfect, cold-blooded Agent action-engine—ideal for JSON routing, SWE-bench patch generation, and anywhere you need absolute structured certainty without token waste.
- 2 replies
posted an update 6 months ago
Post
367
When an AI Model Solves College-Level Math and Physics — On a Phone
This morning I came across a model called Nanbeige4.1-3B, and what began as simple curiosity quickly became something more significant.
I loaded an already 4-bit quantized version and ran it locally on a phone. No GPU, no cloud support, no hidden infrastructure — just a compact reasoning model operating entirely at the edge.
I started with classical mechanics: acceleration, force, friction on an incline. The model worked through them cleanly and correctly. Then I stepped into calculus and gave it a differential equation. It immediately recognized the structure, chose the proper method, carried the mathematics through without confusion, and verified the result.
It did not behave like a model trying to sound intelligent. It behaved like a system trained to solve problems.
And it was doing this on a phone.
For a long time, we have associated serious reasoning in AI with massive models and enormous compute. Capability was supposed to live inside data centers. Bigger models were expected to mean smarter systems.
But watching Nanbeige4.1-3B handle college-level math and physics forces a rethink of that assumption. Intelligence is not only expanding — it is compressing. Better training and sharper reasoning alignment are allowing smaller models to operate far beyond what their size once suggested.
When structured problem-solving runs locally on pocket hardware, the implications are larger than they first appear. Experimentation becomes personal. Engineers can explore ideas without waiting on infrastructure. Students can access serious analytical capability from a device they already carry. Builders are no longer required to send every complex task into the cloud.
What makes moments like this easy to miss is that they rarely arrive with fanfare. There is no dramatic announceme
The model responses are here
https://fate-stingray-0b3.notion.site/AI-model-Nanbeige4-1-3B-304
This morning I came across a model called Nanbeige4.1-3B, and what began as simple curiosity quickly became something more significant.
I loaded an already 4-bit quantized version and ran it locally on a phone. No GPU, no cloud support, no hidden infrastructure — just a compact reasoning model operating entirely at the edge.
I started with classical mechanics: acceleration, force, friction on an incline. The model worked through them cleanly and correctly. Then I stepped into calculus and gave it a differential equation. It immediately recognized the structure, chose the proper method, carried the mathematics through without confusion, and verified the result.
It did not behave like a model trying to sound intelligent. It behaved like a system trained to solve problems.
And it was doing this on a phone.
For a long time, we have associated serious reasoning in AI with massive models and enormous compute. Capability was supposed to live inside data centers. Bigger models were expected to mean smarter systems.
But watching Nanbeige4.1-3B handle college-level math and physics forces a rethink of that assumption. Intelligence is not only expanding — it is compressing. Better training and sharper reasoning alignment are allowing smaller models to operate far beyond what their size once suggested.
When structured problem-solving runs locally on pocket hardware, the implications are larger than they first appear. Experimentation becomes personal. Engineers can explore ideas without waiting on infrastructure. Students can access serious analytical capability from a device they already carry. Builders are no longer required to send every complex task into the cloud.
What makes moments like this easy to miss is that they rarely arrive with fanfare. There is no dramatic announceme
The model responses are here
https://fate-stingray-0b3.notion.site/AI-model-Nanbeige4-1-3B-304
- 4 replies
posted an update 6 months ago
Post
2173
GLM-OCR: A Tiny 0.9B-Parameter Model That Punches Far Above Its Weight
Released today by Z.ai, GLM-OCR is a compact vision-language model designed specifically for document understanding. At just 0.9 billion parameters, it belongs to a new generation of lightweight AI systems proving that raw model size is no longer the only path to high performance.
Despite its small footprint, GLM-OCR posts exceptionally strong results across major document benchmarks. It scores 94.6 on OmniDocBench, 94.0 on OCRBench, and an impressive 96.5 on UniMERNet for formula recognition—numbers that place it alongside, and in some cases ahead of, significantly larger specialized OCR models. The takeaway is clear: efficiency is rapidly becoming a defining feature of modern AI design.
Developed by Z.ai, a research group focused on advancing multimodal foundation models, GLM-OCR reflects a broader shift toward highly optimized architectures that deliver serious capability without requiring massive compute resources.
In practical testing, the model ran successfully in Google Colab on an NVIDIA L4 GPU, demonstrating that advanced document AI is no longer restricted to large research clusters. Engineers, researchers, and developers can now deploy high-quality OCR workflows from relatively accessible hardware.
GLM-OCR signals an important trend in artificial intelligence: smaller, purpose-built models are beginning to rival heavyweight systems while being dramatically easier to run. For anyone working with scanned documents, PDFs, or structured text extraction, this release is a strong indicator of where efficient multimodal AI is heading next.
The google colab notebook for the model
https://colab.research.google.com/drive/1SiXjxPdb-7UJWhtAjPrMZYqPJhyLk9Rc?usp=sharing
The Huggingface model page
zai-org/GLM-OCR
Released today by Z.ai, GLM-OCR is a compact vision-language model designed specifically for document understanding. At just 0.9 billion parameters, it belongs to a new generation of lightweight AI systems proving that raw model size is no longer the only path to high performance.
Despite its small footprint, GLM-OCR posts exceptionally strong results across major document benchmarks. It scores 94.6 on OmniDocBench, 94.0 on OCRBench, and an impressive 96.5 on UniMERNet for formula recognition—numbers that place it alongside, and in some cases ahead of, significantly larger specialized OCR models. The takeaway is clear: efficiency is rapidly becoming a defining feature of modern AI design.
Developed by Z.ai, a research group focused on advancing multimodal foundation models, GLM-OCR reflects a broader shift toward highly optimized architectures that deliver serious capability without requiring massive compute resources.
In practical testing, the model ran successfully in Google Colab on an NVIDIA L4 GPU, demonstrating that advanced document AI is no longer restricted to large research clusters. Engineers, researchers, and developers can now deploy high-quality OCR workflows from relatively accessible hardware.
GLM-OCR signals an important trend in artificial intelligence: smaller, purpose-built models are beginning to rival heavyweight systems while being dramatically easier to run. For anyone working with scanned documents, PDFs, or structured text extraction, this release is a strong indicator of where efficient multimodal AI is heading next.
The google colab notebook for the model
https://colab.research.google.com/drive/1SiXjxPdb-7UJWhtAjPrMZYqPJhyLk9Rc?usp=sharing
The Huggingface model page
zai-org/GLM-OCR
posted an update 7 months ago
Post
3022
KittenTTS Nano — Tiny, Expressive, Practical
KittenTTS Nano is a lightweight, CPU-only text-to-speech model designed to prove that natural, expressive voices don’t require massive cloud stacks or GPUs. At roughly ~15M parameters, it runs fast on modest hardware, supports multiple expressive voices, and exposes simple controls for pacing and tone. This makes it ideal for edge devices, demos, and anyone who wants full control over TTS without latency, lock-in, or infrastructure overhead.
Try it here
Javedalam/KittenTTS
The model page
KittenML/kitten-tts-nano-0.2
KittenTTS Nano is a lightweight, CPU-only text-to-speech model designed to prove that natural, expressive voices don’t require massive cloud stacks or GPUs. At roughly ~15M parameters, it runs fast on modest hardware, supports multiple expressive voices, and exposes simple controls for pacing and tone. This makes it ideal for edge devices, demos, and anyone who wants full control over TTS without latency, lock-in, or infrastructure overhead.
Try it here
Javedalam/KittenTTS
The model page
KittenML/kitten-tts-nano-0.2
- 2 replies
posted an update 8 months ago
Post
270
Testing LFM2-2.6B-Exp on Differential Equations
I recently tested LFM2-2.6B-Exp, an experimental language model developed by Liquid AI, to see how well it handles differential equations in a practical, step-by-step setting.
LFM2-2.6B-Exp is notable for how it was trained: it is an RL-first experimental checkpoint, built without supervised fine-tuning warm-up or distillation. Reinforcement learning was applied sequentially, starting with instruction following and later expanding to knowledge and math. This makes it a particularly interesting model to evaluate beyond benchmark scores.
In hands-on testing, the model performed surprisingly well for its size on standard undergraduate-level differential equations—first-order ODEs, second-order linear equations with constant coefficients, and nonhomogeneous problems using undetermined coefficients. It followed instructions closely and produced clear, structured solution steps.
However, the model showed limitations on more subtle methods, such as Laplace transforms with time shifting and variation of parameters, where maintaining mathematical invariants matters more than following a familiar template. In these cases, answers often looked correct structurally but failed under careful verification. This behavior is consistent with an RL-first training approach: strong at producing expected answer forms, but not always robust on deeper theoretical details.
Liquid AI, the company behind this model, is strongly focused on edge AI, developing efficient models designed for deployment outside large data-center environments. Their model lineup spans from very small models (millions of paramet
Model google notebook can be accessed here
https://colab.research.google.com/drive/1QH9d97oc68VJd0xe4vAbvHArQxpk4Ism?usp=sharing
Full detailed article here
https://fate-stingray-0b3.notion.site/Hands-On-with-LFM2-2-6B-Exp-Testing-an-RL-First-Model-on-Differential-Equations-2d43b975deec809e8b05c52652cfb500
I recently tested LFM2-2.6B-Exp, an experimental language model developed by Liquid AI, to see how well it handles differential equations in a practical, step-by-step setting.
LFM2-2.6B-Exp is notable for how it was trained: it is an RL-first experimental checkpoint, built without supervised fine-tuning warm-up or distillation. Reinforcement learning was applied sequentially, starting with instruction following and later expanding to knowledge and math. This makes it a particularly interesting model to evaluate beyond benchmark scores.
In hands-on testing, the model performed surprisingly well for its size on standard undergraduate-level differential equations—first-order ODEs, second-order linear equations with constant coefficients, and nonhomogeneous problems using undetermined coefficients. It followed instructions closely and produced clear, structured solution steps.
However, the model showed limitations on more subtle methods, such as Laplace transforms with time shifting and variation of parameters, where maintaining mathematical invariants matters more than following a familiar template. In these cases, answers often looked correct structurally but failed under careful verification. This behavior is consistent with an RL-first training approach: strong at producing expected answer forms, but not always robust on deeper theoretical details.
Liquid AI, the company behind this model, is strongly focused on edge AI, developing efficient models designed for deployment outside large data-center environments. Their model lineup spans from very small models (millions of paramet
Model google notebook can be accessed here
https://colab.research.google.com/drive/1QH9d97oc68VJd0xe4vAbvHArQxpk4Ism?usp=sharing
Full detailed article here
https://fate-stingray-0b3.notion.site/Hands-On-with-LFM2-2-6B-Exp-Testing-an-RL-First-Model-on-Differential-Equations-2d43b975deec809e8b05c52652cfb500
reacted to merve's post with 👍 10 months ago
Post
14463
deepseek-ai/DeepSeek-OCR is out! 🔥 my take ⤵️
> pretty insane it can parse and re-render charts in HTML
> it uses CLIP and SAM features concatenated, so better grounding
> very efficient per vision tokens/performance ratio
> covers 100 languages
> pretty insane it can parse and re-render charts in HTML
> it uses CLIP and SAM features concatenated, so better grounding
> very efficient per vision tokens/performance ratio
> covers 100 languages
- 5 replies
posted an update 10 months ago
Post
433
DeepSeek-OCR is a new open-source, vision-language OCR model from DeepSeek-AI (the same lab behind the DeepSeek-V and DeepSeek-R series).
It’s built to read complex, real-world documents — screenshots, PDFs, forms, tables, and handwritten or noisy text — and output clean, structured Markdown.
---
⚙️ Core capabilities
Multimodal (Vision + Language):
Uses a hybrid vision encoder + causal text decoder to “see” layouts and generate text like a language model rather than just classifying characters.
Markdown output:
Instead of raw text, it structures output with Markdown syntax — headings, bullet lists, tables, and inline formatting — which makes the results ideal for direct use in notebooks or LLM pipelines.
PDF-aware:
Includes a built-in PDF runner that automatically slices pages into tiles, processes each region, and re-assembles multi-page outputs.
Adaptive tiling (“crop_mode”):
Automatically splits large pages into overlapping tiles for better recognition of dense, small fonts — the “Gundam mode” mentioned in their docs.
Vision backbone:
Based on DeepSeek-V2’s VL-encoder (≈3 B parameters) trained on massive document + scene-text corpora.
Handles resolutions up to 1280 × 1280 px and dynamically scales lower.
Language head:
Uses the same causal decoder family as DeepSeek-V2, fine-tuned for text reconstruction, so it can reason about table alignment, code blocks, and list structures.
Open and MIT-licensed:
Weights and inference code are fully open under the MIT license, allowing integration into other projects or retraining for domain-specific OCR.
---
🆕 What’s new about its approach
Traditional OCR (e.g., Tesseract, PaddleOCR) → detects and classifies glyphs.
DeepSeek-OCR → interprets the entire document as a multimodal
Google colab notebook running deepseek OCR
https://colab.research.google.com/drive/1Fjzv3UYNoOt28HpM0RMUc8kG34EFgvuu?usp=sharing
The model url
deepseek-ai/DeepSeek-OCR
reacted to Xenova's post with 👍 12 months ago
Post
28831
Okay this is insane... WebGPU-accelerated semantic video tracking, powered by DINOv3 and Transformers.js! 🤯
Demo (+ source code): webml-community/DINOv3-video-tracking
This will revolutionize AI-powered video editors... which can now run 100% locally in your browser, no server inference required (costs $0)! 😍
How does it work? 🤔
1️⃣ Generate and cache image features for each frame
2️⃣ Create a list of embeddings for selected patch(es)
3️⃣ Compute cosine similarity between each patch and the selected patch(es)
4️⃣ Highlight those whose score is above some threshold
... et voilà! 🥳
You can also make selections across frames to improve temporal consistency! This is super useful if the object changes its appearance slightly throughout the video.
Excited to see what the community builds with it!
Demo (+ source code): webml-community/DINOv3-video-tracking
This will revolutionize AI-powered video editors... which can now run 100% locally in your browser, no server inference required (costs $0)! 😍
How does it work? 🤔
1️⃣ Generate and cache image features for each frame
2️⃣ Create a list of embeddings for selected patch(es)
3️⃣ Compute cosine similarity between each patch and the selected patch(es)
4️⃣ Highlight those whose score is above some threshold
... et voilà! 🥳
You can also make selections across frames to improve temporal consistency! This is super useful if the object changes its appearance slightly throughout the video.
Excited to see what the community builds with it!
- 4 replies
reacted to mlabonne's post with 🔥 12 months ago
Post
7027
Liquid just released two 450M and 1.6B param VLMs!
They're super fast and leverage SigLIP2 NaFlex encoders to handle native resolutions without distortion. It's ideal for on-device deployment in constrained environments like phones.
It's available today on Hugging Face, with an inference and a fine-tuning Colab notebooks.
LiquidAI/LFM2-VL-450M
LiquidAI/LFM2-VL-1.6B
They're super fast and leverage SigLIP2 NaFlex encoders to handle native resolutions without distortion. It's ideal for on-device deployment in constrained environments like phones.
It's available today on Hugging Face, with an inference and a fine-tuning Colab notebooks.
LiquidAI/LFM2-VL-450M
LiquidAI/LFM2-VL-1.6B
reacted to fdaudens's post with ❤️ about 2 years ago
Post
3454
Updated the Journalists on 🤗 community page:
- new text-to-speech tools collection https://huggingface.co/collections/JournalistsonHF/text-to-speech-6675c4dccdaa11e86928a15b
- additional leaderboards in the eval collection: TTS-AGI/TTS-Arena and https://huggingface.co/spaces/dylanebert/3d-arena
- new tools in the Text-Analysis collection: gokaygokay/Florence-2, pdf2dataset/pdf2dataset, cvachet/pdf-chatbot
- Xenova/realtime-whisper-webgpu in the Transcription collection
- radames/flash-sd3-taesd3 in the Image Tools collection
- Last but not least, okaris/omni-zero in the fun collection for zero-shot stylized portrait creation
Is there any tool you would like to see added?
Find all the curated tools here: https://huggingface.co/collections/JournalistsonHF/
- new text-to-speech tools collection https://huggingface.co/collections/JournalistsonHF/text-to-speech-6675c4dccdaa11e86928a15b
- additional leaderboards in the eval collection: TTS-AGI/TTS-Arena and https://huggingface.co/spaces/dylanebert/3d-arena
- new tools in the Text-Analysis collection: gokaygokay/Florence-2, pdf2dataset/pdf2dataset, cvachet/pdf-chatbot
- Xenova/realtime-whisper-webgpu in the Transcription collection
- radames/flash-sd3-taesd3 in the Image Tools collection
- Last but not least, okaris/omni-zero in the fun collection for zero-shot stylized portrait creation
Is there any tool you would like to see added?
Find all the curated tools here: https://huggingface.co/collections/JournalistsonHF/