Transcript
Building a hill-climbing machine: Launching seven new MAI models, Mustafa Suleyman, MAI, 2026.06
Paper/Blog Link My Issue
#Article #LanguageModel #AIAgents #TextToImageGeneration #Blog #Coding #Proprietary #TTS #ImageSynthesis #Author Thread-Post Issue Date: 2026-06-03 Comment
- MAI-Thinking-1: Building a Hill-Climbing Machine, Microsoft, 2026.06
元ポスト:
関連:
Voxtral transcribes at the speed of sound, Mistral AI, 2026.02
Paper/Blog Link My Issue
#Article #SpeechProcessing #Blog #MultiLingual #OpenWeight #Proprietary #AutomaticSpeechRecognition(ASR) #Realtime Issue Date: 2026-02-05 Comment
元ポスト:
Voxtral Mini Transcribe V2はproprietaryモデルでAPI利用のみ、Vostraal RealtimeはOpenWeightで公開
mistralai/Voxtral-Mini-4B-Realtime-2602:
https://huggingface.co/mistralai/Voxtral-Mini-4B-Realtime-2602
Vostral Mini Transcrive V2に対するVoxtral Realtimeの性能の比較。Voxtral Realtimeは遅延を調整可能なようで、遅延が大きければ大きいほど高い性能が出るが、リアルタイムに近づけば近づくほど性能はその分劣化する。
LFM2-2.6B-Transcript, LiquidAI, 2026.01
Paper/Blog Link My Issue
#Article #NLP #LanguageModel #OpenWeight #RecurrentModels Issue Date: 2026-01-09 Comment
関連:
- Introducing LFM2: The Fastest On-Device Foundation Models on the Market, LiquidAI, 2025.07