GE

t5gemma-b-b-ul2 tokenizer (Gemma BPE)

GemmaGGUF by thepatchUpdated 2026-08-25
Download on Hugging Face
Parameters282M
Context—
Layers—
Architecturesa3-tokenizer
Smallest quant0.0 GB
Quick answer

t5gemma-b-b-ul2 tokenizer (Gemma BPE) needs about 0.8 GB of memory at Q8_0 (a 0.3 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 4 quants, from 0.0 GB to 1.1 GB (F32).

T5Gemma-b-b-ul2 Encoder + Tokenizer for sa3.cpp

The T5Gemma-b-b-ul2 encoder is a frozen encoder-only model used by Stable Audio 3 to embed text prompts. It is available in multiple precisions and is shared across all SA3 variants. The encoder is validated against the PyTorch reference with cosine similarity ~1.0.

Good at

  • Fast text encoding for audio generation
  • Low memory footprint in quantized form
  • Compatible with sa3.cpp variants

Limits

  • No conditioner included
  • No vision support
  • Requires Gemma license terms

Best for

  • Embedding prompts for Stable Audio 3
  • Preprocessing text for audio models
  • Low‑resource text encoding tasks

Who it’s for. Developers building or extending Stable Audio 3 models. Users who need a lightweight, shared encoder for multiple SA3 variants. Those who want to experiment with different precisions for memory trade‑offs.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
—0.0 GB 0.5 GB
Fits Get file →
Q8_00.3 GB 0.8 GB
Fits Get file →
F160.6 GB 1.1 GB
Fits Get file →
F321.1 GB 1.8 GB
Fits Get file →

For 8 GB VRAM, use Q8_0 (0.3 GB, 0.8 GB memory). For tighter budgets, F16 (0.56 GB, 1.1 GB memory) is a good balance. F32 (1.13 GB, 1.8 GB memory) offers the highest fidelity but needs more memory.

What GPU do you need for t5gemma-b-b-ul2 tokenizer (Gemma BPE)?

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 F32 (1.1 GB file, about 1.8 GB in use).

Which t5gemma-b-b-ul2 tokenizer (Gemma BPE) quant should you choose?

Q8_0 (0.3 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.

If memory is tight, the smallest file is the first one in the table at 0.0 GB, with the quality trade-offs that come with fewer bits.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturesa3-tokenizer
Parameters282M

Source

Quantized bythepatch
FamilyGemma
Licensegemma
Downloads (HF)4,358
Last updated2026-08-25

Frequently asked questions about t5gemma-b-b-ul2 tokenizer (Gemma BPE)

How much VRAM does t5gemma-b-b-ul2 tokenizer (Gemma BPE) need?

About 0.8 GB at Q8_0 with an 8K context window. That is the 0.3 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 t5gemma-b-b-ul2 tokenizer (Gemma BPE) on an 8 GB GPU?

Yes, with F32: the 1.1 GB file needs about 1.8 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of t5gemma-b-b-ul2 tokenizer (Gemma BPE) GGUF?

It depends on the quant: file is 0.0 GB, Q8_0 is 0.3 GB, F16 is 0.6 GB, F32 is 1.1 GB.

Where can I download t5gemma-b-b-ul2 tokenizer (Gemma BPE) GGUF?

The files are on Hugging Face in the thepatch/t5gemma-b-b-ul2-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

What license does t5gemma-b-b-ul2 tokenizer (Gemma BPE) use?

Hugging Face lists the license as gemma. 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 this encoder on 4 GB VRAM?

The smallest quantized version (Q8_0) uses 0.8 GB memory at 8k context, so 4 GB VRAM is sufficient for inference.

Does this model support vision inputs?

No, it is an encoder-only text model with no vision capability.

What license applies to this encoder?

It is released under the Gemma Terms of Use, including Section 3.2 restrictions.

Other GGUFs of t5gemma-b-b-ul2

Similar models

Browse all →

More from thepatch

All by thepatch →

More Gemma models

All Gemma →

Other tiny models

All tiny models →