MMLForge Search models, datasets, tasks… Start free
Home › Models › Image Feature Extraction › ibm-granite/granite-speech-4.1-2b-nar
G

ibm-granite/granite-speech-4.1-2b-nar

Image Feature Extraction·ibm-granite· 178.9K· 57
transformers apache-2.0 Image Feature Extraction 2.3B params

Model Summary: Granite-Speech-4.1-2B-NAR is a non-autoregressive (NAR) speech recognition model that formulates ASR as conditional transcript editing. Instead of decoding tokens one at a time, it edits a CTC hypothesis in a single forward pass using a bidirectional LLM, achieving competitive accuracy with faster inference than autoregressive alternatives. The model is based on the NLE (Non-autoreg

Open in MLForge Sign up free Desktop app Source ↗
# pull & run locally
pip install mlforge-sdk && mlforge pull ibm-granite/granite-speech-4.1-2b-nar

Model details

Task
Image Feature Extraction
Provider
ibm-granite
Framework
transformers
Parameters
2.3B
Size
26 GB
License
apache-2.0
Downloads
178.9K
Likes
57
Paper
arXiv:2603.08397
Updated
2026-06-18

About ibm-granite/granite-speech-4.1-2b-nar

Model Summary: Granite-Speech-4.1-2B-NAR is a non-autoregressive (NAR) speech recognition model that formulates ASR as conditional transcript editing. Instead of decoding tokens one at a time, it edits a CTC hypothesis in a single forward pass using a bidirectional LLM, achieving competitive accuracy with faster inference than autoregressive alternatives. The model is based on the NLE (Non-autoregressive LLM-based Editing) architecture described in this paper.

Related Image Feature Extraction

D facebook/dinov2-small Image Feature Extraction ·22.1M params 2.9M 67 🤗 HF V google/vit-base-patch16-224-in21k Image Feature Extraction ·86.4M params 1.8M 411 🤗 HF V timm/vit_small_patch14_reg4_dinov2.lvd142m Image Feature Extraction 1.5M 7 🤗 HF N nomic-ai/nomic-embed-vision-v1.5 Image Feature Extraction ·92.9M params 1.3M 220 🤗 HF D facebook/dinov2-base Image Feature Extraction ·86.6M params 1.3M 181 🤗 HF D facebook/dinov2-large Image Feature Extraction ·304.4M params 970.9K 113 🤗 HF

Browse all Image Feature Extraction models →