Hugging Face
Search models, datasets, users...
Ctrl K
Models
Datasets
Spaces
Buckets
Docs
Enterprise
Pricing
Website
Tasks
HuggingChat
Collections
Languages
Organizations
Community
Blog
Posts
Daily Papers
Hardware
Learn
Discord
Forum
GitHub
Solutions
Team & Enterprise
Hugging Face PRO
Enterprise Support
Inference Providers
Inference Endpoints
Storage Buckets
Log In
Sign Up
🔄
In a Training Loop
5.2
TFLOPS
Alexandros Liapatis
PRO
alexliap
10
9
69
Follow
cristop's profile picture
NovalisStark's profile picture
hmb's profile picture
9 followers
·
11 following
alexliap
alexandros-liapatis
AI & ML interests
Generative AI + Traditional ML
Recent Activity
liked
a model
5 days ago
higurashi-ai/Haru-4B
updated
a dataset
16 days ago
alexliap/greek-cc-manifests
replied
to
savoji
's
post
18 days ago
LFM3-1.7B from @LiquidAI is the #1 trending model on the Hub (855 likes). It takes about 1 GB at 4-bit, and in the vendor's own table it beats Qwen3.5 9B on several agent benchmarks. The edge decision moves from "are small models good enough?" to "which architecture and runtime do I pick?" One 1.7B backbone now ships as text, VL-2B and Omni-2B checkpoints with open weights. 🔍 Architecture and speed • 24 layers: 13 gated convolutions (kernel 7), 5 sliding-window attention (256 tokens), 6 global attention, chosen by a search over 14 device and backend targets • llama.cpp Q4_0 at 32K context: 1.8 to 2.8x faster decode than the next model on 3 devices • Full 128K context needs 1.6 GB of KV cache vs 4.3 GB (Qwen3.5-4B) and 5.6 GB (MiniCPM5-2B) 📊 Liquid AI's table • Average 49.37 vs 47.24 for Qwen3.5 9B; BFCL V4 71.39 vs 62.83; WebArena-Verified 38.42 vs 28.94 • Qwen3.5 9B still wins SWE-bench Verified (51.20 vs 43.40) and GPQA Diamond (80.40 vs 63.37) 🛠️ Try it • GGUF (174,177 downloads last month) runs in llama.cpp b11960+, Ollama and LM Studio; MLX 4-bit for Apple Silicon; SGLang for GPUs • llama-cli -hf LiquidAI/LFM3-1.7B-GGUF:Q4_K_M --temp 0.1 --top-k 50 --repeat-penalty 1.1 • Omni-2B runs in the browser in the WebGPU Space ⚠️ Caveats • The numbers are vendor-reported, with no independent reproduction in the sources. The card says it is not for agentic coding or complex multi-step reasoning. https://huggingface.co/LiquidAI/LFM3-1.7B https://huggingface.co/LiquidAI/LFM3-1.7B-GGUF https://huggingface.co/LiquidAI/LFM3-Omni-2B https://huggingface.co/spaces/LiquidAI/LFM3-Omni-2B-WebGPU Has a sub-2B model held up on your own tool-calling harness? • Medium: https://medium.com/p/05507e4e81c8 • Substack: https://sajjadpakdamansavoji.substack.com/p/liquid-ais-lfm3-turns-the-small-model • X: https://x.com/saavoji/status/2117307514177941672 • Website: https://sajjadpsavoji.github.io/blog#liquid-ais-lfm3
View all activity
Organizations
None yet
alexliap
's activity
All
Models
Datasets
Spaces
Buckets
Papers
Collections
Community
Posts
Upvotes
Likes
Articles
New activity in
disco-eth/WorldSpeech
about 1 month ago
Diff el_gr vs el-cy
2
#6 opened about 2 months ago by
alexliap
New activity in
AxiomicLabs/Open_SLM_Leaderboard
2 months ago
Add typakos-140M-it benchmark results
1
#149 opened 2 months ago by
alexliap
New activity in
alexliap/greek-cc
2 months ago
[bot] Conversion to Parquet
#1 opened 2 months ago by
parquet-converter
New activity in
openbmb/UltraData-Math
2 months ago
Languages
2
#33 opened 2 months ago by
alexliap
New activity in
nvidia/nemotron-3.5-asr-streaming-0.6b
5 months ago
Lora funetuning
#10 opened 5 months ago by
alexliap
New activity in
alexliap/greek-synth-v1
6 months ago
[bot] Conversion to Parquet
#1 opened 6 months ago by
parquet-converter
New activity in
ilsp/llama-krikri-8b-ag-mg-qlora
6 months ago
Dataset sources
2
#1 opened 6 months ago by
alexliap
New activity in
shigureui/Qwen3-VL-Embedding-2B-FP8
9 months ago
Script for recreating the quantization
#1 opened 9 months ago by
alexliap
New activity in
alexliap/high-quality-gr-text
10 months ago
[bot] Conversion to Parquet
#1 opened 10 months ago by
parquet-converter
New activity in
callanwu/WebWalkerQA
about 1 year ago
Benchmark
#3 opened about 1 year ago by
alexliap
Load more