1 5 10

Mengqi Li

Kullpar

ElementQi

AI & ML interests

None yet

Recent Activity

liked a dataset about 18 hours ago

open-thoughts/OpenThoughts3-1.2M

liked a model 4 days ago

ggml-org/MiniCPM-V-4.6-GGUF

liked a model 4 days ago

openbmb/MiniCPM-V-4.6

View all activity

Organizations

liked a dataset about 18 hours ago

open-thoughts/OpenThoughts3-1.2M

Viewer • Updated Jun 9, 2025 • 1.2M • 15.3k • 227

liked 2 models 4 days ago

ggml-org/MiniCPM-V-4.6-GGUF

Image-Text-to-Text • 0.8B • Updated 5 days ago • 11.9k • 22

openbmb/MiniCPM-V-4.6

Image-Text-to-Text • 1B • Updated about 2 hours ago • 56.5k • 660

liked a model 2 months ago

Jackrong/Qwen3.5-27B-Claude-4.6-Opus-Reasoning-Distilled

Image-Text-to-Text • 28B • Updated Apr 6 • 226k • • 2.84k

reacted to evalstate's post with 🤗 3 months ago

Post

3734

Hugging Face MCP Server v0.3.2
~~~~~~~~~~~~~~~~~~~~~~~~~~~~

- Replace model_search and dataset_search with combined hub_repo_search tool.
- Less distracting description for hf_doc_search
- model_search and dataset_search tool calls will still function (plan to remove next release).

4 replies

liked a model 3 months ago

moonshotai/Kimi-K2.5

Image-Text-to-Text • 1.1T • Updated 17 days ago • 1.76M • • 2.78k

reacted to andito's post with 🔥 11 months ago

Post

4103

🧠👁️ Can AI visualize solutions?

Humans often solve visual problems by sketching ideas in our minds. What if Vision-Language Models (VLMs) could do something similar, not by generating full images, but by using internal “mental sketches”?

That’s the idea behind Mirage, a new framework that empowers VLMs to reason using latent visual tokens. Instead of just thinking in words, Mirage mixes in abstract visual representations that help the model solve complex tasks.

These aren't photorealistic images. They're compact, internal representations optimized purely to support reasoning.

🔧 Mirage is trained in two phases:

1) Grounding: It learns to produce latent tokens anchored in real images.
2) Refinement: The model drops the images and learns to generate visual tokens on its own.

📈 And yes, it works!
On challenging benchmarks like Visual Spatial Planning, Jigsaw puzzles, and Spatial Attention Tasks, Mirage clearly outperforms GPT-4o and other strong baselines.
Smart sketches > empty words.

By mimicking the way humans visualize solutions, Mirage gives AI a new kind of imagination, one that’s faster, more efficient, and more human-like.
Kudos to the teams at UMass Amherst and MIT behind this exciting work.
Check the paper: Machine Mental Imagery: Empower Multimodal Reasoning with Latent Visual Tokens (2506.17218)

4 replies

reacted to yeonseok-zeticai's post with 👍🔥 11 months ago

Post

3321

Hi everyone,

I’ve been running small language models (SLLMs) directly on smartphones — completely offline, with no cloud backend or server API calls.

I wanted to share:
1. ⚡ Tokens/sec performance across several SLLMs
2. 🤖 Observations on hardware utilization (where the workload actually runs)
3. 📏 Trade-offs between model size, latency, and feasibility for mobile apps

There are reports for below models
- QWEN3 0.6B
- NVIDIA/Nemotron QWEN 1.5B
- SimpleScaling S1
- TinyLlama
- Unsloth tuned Llama 3.2 1B
- Naver HyperClova 0.5B

📜Comparable Benchmark reports (no cloud, all on-device):
I’d really value your thoughts on:
- Creative ideas to further optimize inference under these hardware constraints
- Other compact LLMs worth testing on-device
- Experiences you’ve had trying to deploy LLMs at the edge

If there’s interest, I’m happy to share more details on the test setup, hardware specs, or the tooling we used for these comparisons.

Thanks for taking a look, and you can build your own through at "https://mlange.zetic.ai"!