Today at 3:30 PM, Subhashini Venugopalan will be at the #ICASSP2025 Google booth to discuss how to enable the open source Gemma LLM to recognize speech and to act as an ASR model. She'll also discuss how Gemma can be trained using RL to understand disordered speech.
OPEN SOURCE
-
Llama-4-Maverick tops smolagents leaderboard, beats DeepSeek-R1
By
–
Llama-4-Maverick is CRAZY GOOD to power agents It's now the top open model on smolagents LLM leaderboard, beating the much larger DeepSeek-R1!
Congrats @ThomasScialom and team! -
No lock-in: vendor flexibility and self-hosting freedom
By
–
no lock-in You can switch/mix with other providers. You can freely switch to self-hosting at any point. Does this make sense?
-
Open-Source Transparency Enables Faster Bug Detection and Fixes
By
–
The beauty of open-source is that you create more transparency & we can fix misunderstanding super quickly which seems to be happening here. Imagine all the similar occurrences that probably happened with proprietary APIs that we never even heard of!
-
Anthropic’s MCP Protocol Leaks Outside Company First Time
By
–
and here's the first "leak" of MCP outside Anthropic we discussed on the pod 🙂
-
Pretty-printing Jinja templates on the Hub
By
–
should we pretty-print Jinja templates on the Hub? Feels like it would make reading those pesky chat_templates files more pleasant to read WDYT?
-
Llama4: 3 new Open Source AIs with 10 million tokens
By
–
Boom → @MetaAI just released #Llama4, 3 new Open Source AIs with 10 million tokens. What is it worth and how to use Llama 4 for free? I tell you everything here https://
youtu.be/-Ns2-CfgXC4 -
Defining Openness in Machine Learning Systems: Fine Details
By
–
In case you care about the fine details of what "open" means in the context of ML systems. https://
arxiv.org/abs/2405.15802 -
Open-Reasoner-Zero: Scaling Reinforcement Learning on Base Models
By
–
Open-Reasoner-Zero: An Open Source Approach to Scaling Up Reinforcement Learning on the Base Model Hu et al.: https://
arxiv.org/abs/2503.24290 #ArtificialIntelligence #DeepLearning #MachineLearning -
Open Deep Search: Open-Source AI Search Framework Rivals GPT-4o
By
–
7). Open Deep Search Introduces Open Deep Search (ODS), an open-source search AI framework that rivals top proprietary systems like GPT-4o Search Preview and Perplexity Sonar.