Granite Switch 4.1: Chat, Decisions, and Speech from One Endpoint IBM released Granite Switch 4.1, with Granite-Switch-4.1-8B and -30B models that combine a Decision aLoRA, a Guardian aLoRA, and the Granite-Speech-5.0-Turbo-CTC ASR model into a single hosted endpoint at roughly the VRAM footprint of the base model alone. IBM said the Guardian aLoRA on average outperforms Jev 1.13, Microsoft-Decision-1, GPT-Luna-Decisions, and Clef-Flash on safety-related decision tasks, and that Granite-Speech-5.0-Turbo-CTC is the fastest ASR model on the OpenASR Leaderboard. The adapters are built as cache-efficient Activated LoRAs, so the KV cache from a base model generation can be reused for subsequent adapter-based decisions. Automatic Speech Recognition • 0.5B • Updated • 59.1k • 78 https://huggingface.co/ibm-granite/granite-speech-5.0-470m-turboctc Granite Switch 4.1: Chat, Decisions, and Speech from One Endpoint Enterprise Article https://huggingface.co/blog Granite Switch https://github.com/generative-computing/granite-switch project enables a modular approach to model development. Now Granite Switch 4.1 models, armed with a Decision adapter and the fastest ASR model on the OpenASR leaderboard, can move efficiently between natural-language responses, general and specialized decisions, and speech-to-text tasks—all from one hosted endpoint. Specifically, Granite-Switch-4.1-8B https://huggingface.co/ibm-granite/granite-switch-4.1-8b and -30B https://huggingface.co/ibm-granite/granite-switch-4.1-30b models are configured to include: 1. A new Decision aLoRA https://huggingface.co/ibm-granite/granitelib-core-r1.0/blob/main/decision/README.md for general-purpose decision-making e.g. v1/systemone , 2. A specialized Guardian aLoRA https://huggingface.co/ibm-granite/granitelib-guardian-r1.0/blob/main/guardian-core/README.md that, on average, outperforms Jev 1.13, Microsoft-Decision-1, GPT-Luna-Decisions, and Clef-Flash on safety-related decision tasks, and 3. Granite-Speech-5.0-Turbo-CTC https://huggingface.co/ibm-granite/granite-speech-5.0-470m-turboctc , the fastest ASR model on the OpenASR Leaderboard https://huggingface.co/spaces/hf-audio/open asr leaderboard , embedded for use in transcription tasks. All three of these technologies along with the underlying base model are composed into one model, taking up approximately the same amount of VRAM on a GPU as the base model would by itself. That means more capabilities with fewer GPUs. And because the Switch adapters were built as cache-efficient Activated LoRAs https://research.ibm.com/blog/inference-friendly-aloras-lora , the KV cache created from a base model generation can be reused for any subsequent adapter-based decisions. Combined with our team's recent vLLM contributions https://github.com/generative-computing/granite-switch/blob/main/docs/SWITCH KERNELS.md to speed up low rank adapter inferencing, as well as a co-located ASR encoder, Granite Switch 4.1 models can switch between a chat request and a decision request and a spoken request, with blazing fast speeds. Granite-switch-4.1-8b can transcribe audio, generate summaries, screen for harmful content, and test 15+ generic flags—all running on a single endpoint in real time. Granite Libraries: The OG Decision Models Decision models aren’t new. Our previous release of Granite Libraries https://huggingface.co/collections/ibm-granite/granite-libraries features a repository of specialized “decision model” adapters and other specialized task adapters that can be composed into custom Granite Switch models. And while Jev and other closed decision models focus on generality, we’ve found that many of our specialized granitelib adapters can outperform generalists on specialized classification tasks. Compare granitelib adapters https://huggingface.co/spaces/ibm-granite/granitelib-evals with generalist decision models when evaluated on specialized classification tasks. Because adapters share the base model's KV cache, adding specialists alongside a generalist decision adapter costs almost nothing. You get broad coverage from the generalist and higher accuracy from specialists on the tasks you care about. In a “Race” between Granite-Switch-4.1-30B frozen base Granite + 4 specialized aLoRAs and Jev-27B frozen base Qwen model + 1 generic decision LoRA , Granite-Switch-4.1-30B was able to complete a real world, 5-step RAG workflow much faster than its competitor. Compose and Extend Your Own Granite Switch Model Granite-Switch-4.1-8B and -30B are released with “batteries included”. They are Granite Switch models that we compiled with a predefined set of helpful adapters. But you don’t need to use our version. Anyone can compose their own Granite Switch model https://generative-computing.github.io/granite-switch/adapter catalog.html , mixing and matching adapters from Granite Libraries to best meet their needs. Moving forward, we see tremendous opportunity for the community to compose and extend Granite Switch models with custom and third party adapters. We plan to release an Adapter SDK for training adapters in Mellea https://docs.mellea.ai/advanced/lora-and-alora-adapters , along side other tutorials and Granite Switch support. These adapters could be specialized classification adapters to augment decision-based workflows or adapters for specialized language tasks. You could even use custom adapters to teach a 3B parameter version of Granite to play the classic game Doom. Similarly to Jev, our custom 3B parameter Granite Switch model made split second decisions to navigate the Doom game, but was able to do so at 60 - 100 decisions / second, compared to Jev’s reported 10 decisions / second. 1 footnote-1 Unlike Jev, this Granite Switch model can also narrate and respond to your questions while it is playing. Resources for Getting Started - Granite Switch 4.1 Model Cards https://huggingface.co/collections/ibm-granite/granite-switch - Speech-to-Decisions Demo https://huggingface.co/spaces/ibm-granite/speech-to-decisions - Granite Switch GitHub Repo https://github.com/generative-computing/granite-switch - Original Granite Switch and Granite Libraries Release Blog https://research.ibm.com/blog/granite-libraries-project-switch - Create your own aLoRA v LoRA race on Colab https://colab.research.google.com/github/generative-computing/granite-switch/blob/main/tutorials/notebooks/alora vs lora race.ipynb . Footnotes