Introducing smolagents: simple agents that write actions in code.
•
505
Aymeric Roucher
m-ric
AI & ML interests
Leading Agents at Hugging Face 🤗
Recent Activity
updated
a Space
about 11 hours ago
m-ric/hf-model-downloads
liked
a Space
about 12 hours ago
NyxKrage/LLM-Model-VRAM-Calculator
upvoted
an
article
about 12 hours ago
The Large Language Model Course
Articles
Organizations
Posts
84
Post
707
𝗠𝗶𝗻𝗶𝗠𝗮𝘅'𝘀 𝗻𝗲𝘄 𝗠𝗼𝗘 𝗟𝗟𝗠 𝗿𝗲𝗮𝗰𝗵𝗲𝘀 𝗖𝗹𝗮𝘂𝗱𝗲-𝗦𝗼𝗻𝗻𝗲𝘁 𝗹𝗲𝘃𝗲𝗹 𝘄𝗶𝘁𝗵 𝟰𝗠 𝘁𝗼𝗸𝗲𝗻𝘀 𝗰𝗼𝗻𝘁𝗲𝘅𝘁 𝗹𝗲𝗻𝗴𝘁𝗵 💥
This work from Chinese startup @MiniMax-AI introduces a novel architecture that achieves state-of-the-art performance while handling context windows up to 4 million tokens - roughly 20x longer than current models. The key was combining lightning attention, mixture of experts (MoE), and a careful hybrid approach.
𝗞𝗲𝘆 𝗶𝗻𝘀𝗶𝗴𝗵𝘁𝘀:
🏗️ MoE with novel hybrid attention:
‣ Mixture of Experts with 456B total parameters (45.9B activated per token)
‣ Combines Lightning attention (linear complexity) for most layers and traditional softmax attention every 8 layers
🏆 Outperforms leading models across benchmarks while offering vastly longer context:
‣ Competitive with GPT-4/Claude-3.5-Sonnet on most tasks
‣ Can efficiently handle 4M token contexts (vs 256K for most other LLMs)
🔬 Technical innovations enable efficient scaling:
‣ Novel expert parallel and tensor parallel strategies cut communication overhead in half
‣ Improved linear attention sequence parallelism, multi-level padding and other optimizations achieve 75% GPU utilization (that's really high, generally utilization is around 50%)
🎯 Thorough training strategy:
‣ Careful data curation and quality control by using a smaller preliminary version of their LLM as a judge!
Overall, not only is the model impressive, but the technical paper is also really interesting! 📝
It has lots of insights including a great comparison showing how a 2B MoE (24B total) far outperforms a 7B model for the same amount of FLOPs.
Read it in full here 👉 MiniMax-01: Scaling Foundation Models with Lightning Attention (2501.08313)
Model here, allows commercial use <100M monthly users 👉 MiniMaxAI/MiniMax-Text-01
This work from Chinese startup @MiniMax-AI introduces a novel architecture that achieves state-of-the-art performance while handling context windows up to 4 million tokens - roughly 20x longer than current models. The key was combining lightning attention, mixture of experts (MoE), and a careful hybrid approach.
𝗞𝗲𝘆 𝗶𝗻𝘀𝗶𝗴𝗵𝘁𝘀:
🏗️ MoE with novel hybrid attention:
‣ Mixture of Experts with 456B total parameters (45.9B activated per token)
‣ Combines Lightning attention (linear complexity) for most layers and traditional softmax attention every 8 layers
🏆 Outperforms leading models across benchmarks while offering vastly longer context:
‣ Competitive with GPT-4/Claude-3.5-Sonnet on most tasks
‣ Can efficiently handle 4M token contexts (vs 256K for most other LLMs)
🔬 Technical innovations enable efficient scaling:
‣ Novel expert parallel and tensor parallel strategies cut communication overhead in half
‣ Improved linear attention sequence parallelism, multi-level padding and other optimizations achieve 75% GPU utilization (that's really high, generally utilization is around 50%)
🎯 Thorough training strategy:
‣ Careful data curation and quality control by using a smaller preliminary version of their LLM as a judge!
Overall, not only is the model impressive, but the technical paper is also really interesting! 📝
It has lots of insights including a great comparison showing how a 2B MoE (24B total) far outperforms a 7B model for the same amount of FLOPs.
Read it in full here 👉 MiniMax-01: Scaling Foundation Models with Lightning Attention (2501.08313)
Model here, allows commercial use <100M monthly users 👉 MiniMaxAI/MiniMax-Text-01
Post
2138
𝗪𝗲'𝘃𝗲 𝗷𝘂𝘀𝘁 𝗿𝗲𝗹𝗲𝗮𝘀𝗲𝗱 𝘀𝗺𝗼𝗹𝗮𝗴𝗲𝗻𝘁𝘀 𝘃𝟭.𝟯.𝟬 🚀, and it comes with a major feature: you can now log agent runs using OpenTelemetry to inspect them afterwards! 📊
This interactive format is IMO much easier to inspect big multi-step runs than endless console logs.
The setup is very easy, in a few lines of code.
Find a tutorial here 👉 https://huggingface.co/docs/smolagents/tutorials/inspect_runs
This interactive format is IMO much easier to inspect big multi-step runs than endless console logs.
The setup is very easy, in a few lines of code.
Find a tutorial here 👉 https://huggingface.co/docs/smolagents/tutorials/inspect_runs
Collections
14
-
Scaling Autoregressive Models for Content-Rich Text-to-Image Generation
Paper • 2206.10789 • Published • 4 -
Beyond Chinchilla-Optimal: Accounting for Inference in Language Model Scaling Laws
Paper • 2401.00448 • Published • 28 -
Training Compute-Optimal Large Language Models
Paper • 2203.15556 • Published • 10 -
Scaling Laws for Neural Language Models
Paper • 2001.08361 • Published • 7
spaces
27
pinned
Running
112
🤔📊
Agent Data Analyst
Need to analyze data? Let a Llama-3.1 agent do it for you!
pinned
Running
69
🗺️🏕️
AI Travel Planner
Plan your next vacation with the help of an AI!
pinned
Running
194
🍫👀
Chunk Visualizer
Pick a text splitter => visualize chunks. Great for RAG.
pinned
Running
on
Zero
125
✍👀
Beam Search Visualizer
View how beam search decoding works, in detail!
Running
3
🏢
Hf Model Downloads
Runtime error
🔎👀
Attribution Visualizer
Find which input tokens influence the output of your LLM.
models
3
datasets
18
m-ric/smol_agents_benchmark
Viewer
•
Updated
•
132
•
134
m-ric/agents_medium_benchmark_3
Viewer
•
Updated
•
132
•
23
m-ric/agents_medium_benchmark_2
Viewer
•
Updated
•
142
•
211
•
7
m-ric/agents_medium_benchmark
Viewer
•
Updated
•
172
•
92
•
3
m-ric/newsletter
Updated
•
30
m-ric/songs_july_24
Viewer
•
Updated
•
85.6k
•
34
m-ric/huggingface_doc_qa_eval
Viewer
•
Updated
•
65
•
232
•
9
m-ric/songs
Preview
•
Updated
•
10
m-ric/rag_scores_cookbook
Viewer
•
Updated
•
5
•
58
m-ric/images
Viewer
•
Updated
•
1
•
10