kvizas/parakeet-tdt-0.6b-diar-head

AI model by kvizas

Open original source ↗

From the publisher

Model card & documentation

Source preview

Read the publisher’s intended use, setup instructions, evaluations and limitations. The original model card is the source of truth.

License. Everything in this repository — the head weights (head.pt), diarize.py and the configuration — is released under CC-BY-4.0, the same licence as the base model nvidia/parakeet-tdt-0.6b-v2, which this repository does not redistribute: diarize.py downloads it from NVIDIA at run time. This is a speaker diarization system built entirely on top of a frozen ASR model. NVIDIA's Parakeet-TDT-0.6B-v2 is loaded, switched to limited-context attention so it can take a whole meeting in one pass, and then never touched again: no weights change, no gradients flow into it, and its transcription quality is unaffected. The only thing that is read out of it is the activation of Conformer layer 5 — one 1024-dimensional vector every 80 ms. A small trained transformer (15.2 M parameters, 4 layers, d=512) reads that sequence and emits, per frame, four independent "is speaker k talking right now" probabilities. Because the probabilities are independent, overlapped speech is…

kvizas/parakeet-tdt-0.6b-diar-head on Hugging Face A short preview, not the full document.

Read the full model card ↗ · Preview checked 2026-09-21T07:50:40.988Z

Inside the original model card — Document outline
  1. Parakeet-L5 diarization head
  2. What this is
  3. Quick start
  4. How it works
  5. Evaluation
  6. Limitations
  7. Code and paper

Headings are captured from the source. Links open the publisher’s document, not a locally hosted copy.

Documentation belongs to its respective authors. Reported project/model license: cc-by-4.0. A listing is not a grant of reuse or training rights. Confirm the document’s own terms at the source.

Model overview

kvizas/parakeet-tdt-0.6b-diar-head is a task-unspecified model repository published by kvizas on Hugging Face. The source reports the nemo library.

This page summarizes Hub metadata. For intended use, training data, evaluation results and limitations, consult the original model card.

Model facts

Task
Not reported
Library
nemo
Recent downloads (30 days)
0
Cumulative likes
0
Hugging Face trending score
0
Reported safetensors parameters
Not reported
Architecture
Not reported
License
cc-by-4.0
Access
Not gated by Hugging Face
Created
2026-09-21T06:12:44.000Z
Last modified
2026-09-21T06:13:22.000Z

Compatibility and lineage

Reported languages: en

Reported base models: nvidia/parakeet-tdt-0.6b-v2

Parameter count is not a RAM/VRAM requirement. Check precision, quantization, context length and runtime compatibility in the model card; no hardware or API-cost claim is inferred here.

Use and evaluate

Check license terms and access requirements first. Review the model card’s documented loading instructions, evaluations and safety limitations. Benchmark scores are not imported by this directory, and popularity is not an accuracy ranking.

No model weights are downloaded or executed by AltAPIs. Never enable remote model code without reviewing it.

Model card and usage instructions

Model files and configuration

Community discussion

Source and freshness

Source: Hugging Face Hub. Metadata observed 2026-09-21T06:36:48.800Z. Daily imports are snapshots, not real-time monitoring.

Popularity and source listings do not establish security, suitability, licensing rights or benchmark performance.

Open original source

Related models