NE

Raj 0

NemotronGGUF by aungzaythantUpdated 2026-10-12
Download on Hugging Face
Parameters637M
Context—
Layers—
Architectureasr
Smallest quant0.7 GB
Quick answer

Raj 0 needs about 1.3 GB of memory at Q8_0 (a 0.7 GB file) with an 8K context window, so it fits on a 4 GB GPU.

Raj‑0: Burmese ASR model based on Nemotron 3.5

Raj‑0 is a 0.637 billion‑parameter Automatic Speech Recognition model fine‑tuned for Burmese. It extends NVIDIA’s Nemotron 3.5 ASR with language‑ID conditioning to support 40 language‑locales. The model is released under Apache 2.0 and can run locally without cloud GPU.

Good at

  • Multilingual transcription
  • Low‑latency streaming
  • Burmese speech recognition

Limits

  • No vision support
  • Requires 8 k context for full memory estimate
  • Only Q8_0 quant available

Best for

  • Transcribing Burmese audio files
  • Live Burmese voice‑assistant demos
  • On‑device transcription on mid‑range phones

Who it’s for. Developers building Burmese language voice applications or researchers needing an open‑source ASR model. The model is suitable for on‑device inference and low‑cost production deployments.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q8_00.7 GB 1.3 GB
Fits Get file →

The Q8_0 quant weighs 0.72 GB and requires about 1.3 GB of VRAM at 8 k context. This is suitable for GPUs with 4 GB or more. No other quantizations are provided, so use Q8_0 for all deployments.

What GPU do you need for Raj 0?

These figures add the file size, an fp16 KV cache at an 8K context window and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.

  • 4 GB VRAM: the best quant that fits is Q8_0 (0.7 GB file, about 1.3 GB in use).

Which Raj 0 quant should you choose?

Q8_0 (0.7 GB) is the middle option in this repository. Close to lossless. It costs roughly twice the memory of a 4-bit file, so it suits cases where you have plenty of VRAM.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architectureasr
Parameters637M

Source

Quantized byaungzaythant
FamilyNemotron
Licenseapache-2.0
Downloads (HF)843
Last updated2026-10-12

Frequently asked questions about Raj 0

How much VRAM does Raj 0 need?

About 1.3 GB at Q8_0 with an 8K context window. That is the 0.7 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.

Can I run Raj 0 on an 8 GB GPU?

Yes, with Q8_0: the 0.7 GB file needs about 1.3 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of Raj 0 GGUF?

It depends on the quant: Q8_0 is 0.7 GB.

Where can I download Raj 0 GGUF?

The files are on Hugging Face in the aungzaythant/Raj-0 repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

What license does Raj 0 use?

Hugging Face lists the license as apache-2.0. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.

Can I run Raj‑0 on 8 GB VRAM?

Yes, the Q8_0 quant needs about 1.3 GB of VRAM at 8 k context, well within an 8 GB GPU.

Does Raj‑0 support English?

It is fine‑tuned for Burmese; English words may appear phonetically but accuracy is limited.

Is Raj‑0 suitable for real‑time streaming?

Yes, it uses a cache‑aware FastConformer‑RNNT architecture designed for low‑latency streaming.

Other GGUFs of nemotron-3.5-asr-streaming-0.6b

Similar models

Browse all →

More from aungzaythant

All by aungzaythant →

More Nemotron models

All Nemotron →

Other tiny models

All tiny models →