IN TODAY'S SIGNAL |
|
Read time: 3 min 45 sec |
|
🎖️ Top News
📌 Latitude
⚡️ Trending Signals
-
Microsoft's GraphRAG extracts structured data from unstructured text
-
InternLM 2.5 tops open LLM leaderboard for models under 12B
-
Eleven Labs adds voice isolation for clear audio in productions
-
Gemini now allows text, drawings, photos, emojis as prompts
-
Nomic's new LLM desktop app supports thousands of LLMs
🛠️ Top of Github
-
MindsDB: automate AI/ML workflows with data source integration
-
DeepSpeed: Microsoft’s 15x faster RLHF training for large models
-
Maestro: orchestrate subagents with Claude Opus, GPT, local LLMs
🧠 Tutorial
|
|
|
|
|
Enjoying this newsletter?
Please forward it to a friend or colleague. It helps us keep this content free. |
|
|
|
TOP NEWS |
|
Open-Source |
French AI Lab Announces an Open-Source GPT-4o Multimodal Alternative: Moshi |
|
⇧ 4790 Likes |
 |
What's New |
|
Kyutai Labs, a French AI lab that has raised $300M, is dedicated to open-sourcing every AI model it develops. Its latest release, Moshi, is a real-time multimodal foundation model that surpasses functionalities of OpenAI's GPT-4o, focusing on real-time audio interactions and emotional expressiveness.
Core Technical Specifications
- Base Model: Helium 7B, a 7-billion parameter language model.
- Training Data: Jointly trained on text and audio, including synthetic data.
- Latency: Achieves an end-to-end latency of 200 milliseconds.
- Compatibility: Operates on both MacBooks and consumer-grade GPUs.
Advanced Features and Capabilities
- Moshi processes two audio streams simultaneously, allowing it to listen and speak without modeling explicit speaker turns.
- Utilizes a two-channel input/output system, seamlessly integrating text tokens and audio codecs.
- Watermarking technology detects AI-generated audio, aiming to enhance the model's security features.
Accessibility and Innovation
- Kyutai plans to release Moshi as an open-source project, which will include the model, code, and associated research papers.
|
Access |
|
|
|
|
| TRY THE DEMO |
|
|
|
 |
Launchpad: A Revamped Container Experience for AI Engineers |
|
Launchpad is Latitude.sh's purpose-built platform for AI applications, a fast solution that lets you run everything from small databases to the largest AI models.
The platform has been constantly evolving based on the feedback from the Machine-Learning community, and now has many new features available to improve your container experience:
-
Different GPUs: NVIDIA's L40S (48 GB vRAM) and H100 (80 GB vRAM)
-
SSH support: SSH access for debugging and development
-
Filesystem volumes: Add persistent storage to multiple containers
-
Per-minute billing: Pay only for what you use
-
Blueprint library: Instantly launch containers from a community-built library
Create your free account today and deploy container-based applications on dedicated GPUs and CPUs in just a few seconds! |
| GET STARTED |
|
partner with us |
|
|
|
TRENDING SIGNALS |
|
Graphs |
|
|
|
⇧ 1790 Likes |
|
|
|
Open Source |
|
|
|
⇧ 248 Likes |
|
|
|
Voice |
|
|
|
⇧ 615 Likes |
|
|
|
Prompting |
|
|
|
⇧ 385 Likes |
|
|
|
Language Models |
|
|
|
⇧ 1000 Likes |
|
|
|
|
|
|
|
|
TOP OF GITHUB |
|
Web Apps |
|
|
|
MindsDB connects data sources with AI/ML frameworks and automates routine workflows between them. You can deploy, serve, and fine-tune models in real-time, utilizing data from databases, vector stores, or applications. |
|
☆ 24,810 |
|
|
|
Optimization |
|
|
|
Microsoft DeepSpeed is a deep learning optimization library that makes distributed training and inference easy, efficient, and effective. It empowers ChatGPT-like model training offering 15x speedup over SOTA RLHF. |
|
☆ 33,719 |
|
|
|
Agents |
|
|
|
Maestro helps you orchestrate subagents with Claude Opus, GPT, and local LLMs. Supports APIs like OpenAI and Cohere. Automates task breakdown, execution, refinement, and logging. Integrates with local servers. |
|
☆ 3521 |
|
|
|
|
|
|
|
|
TUTORIAL |
How to add Generative AI capabilities to Pandas |
|
PandasAI is a Python library that integrates Generative AI with Pandas, allowing developers to analyze complex data frames and create visualizations using natural language.
With PandasAI, you can clean data, impute missing values, generate new features, analyze, and manipulate data effortlessly. It enables both non-technical and technical users to interact with data in a more intuitive way, saving time and effort.
You can use it in Jupyter notebooks, Streamlit apps, or deploy it as a REST API. Install PandasAI using pip, set your API key, and start querying your data or generating charts with simple commands. |
import os
import pandas as pd
from pandasai import Agent
# Sample DataFrame
sales_by_country = pd.DataFrame({
"country": ["United States", "United Kingdom", "France",
"Germany", "Italy", "Spain", "Canada"],
"revenue": [5000, 3200, 2900, 4100, 2300, 2100, 2500]
})
# By default, unless you choose a different LLM, it will use BambooLLM.
# You can get your free API key signing up at https://pandabi.ai
os.environ["PANDASAI_API_KEY"] = "YOUR_API_KEY"
agent = Agent(sales_by_country)
agent.chat('Which are the top 5 countries by sales?')
|
| CHECK THE REPO |
|
|
|
LAST WEEK'S GREATEST HITS |
|
|
|
|
|
|