XV

xVITA Embed 300M zhTW

XvitaGGUF by xCloudinfoUpdated 2026-10-26
Download on Hugging Face
Parameters308M
Context8K
Layers24
Architecturegemma-embedding
Smallest quant0.3 GB
Quick answer

xVITA Embed 300M zhTW needs about 1.0 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 0.3 GB (Q8_0) to 0.6 GB (F16).

xVITA Embed 300M zhTW – 300M‑parameter embedding model for Traditional Chinese

xVITA Embed 300M zhTW is a 0.308 billion‑parameter gemma‑embedding model. It supports up to 8192 context tokens and is available in Q8_0 and F16 quantizations. The model is optimized for paragraph‑level and full‑document retrieval in Traditional Chinese.

Good at

  • Paragraph‑level RAG
  • Full‑document retrieval
  • High cosine similarity with safetensors

Limits

  • No vision support
  • License: gemma
  • Requires llama.cpp or ollama

Best for

  • Traditional Chinese document search
  • Embedding generation for RAG pipelines
  • Vector similarity tasks

Who it’s for. Developers building retrieval‑augmented systems in Traditional Chinese. Researchers needing a lightweight embedding model with proven cosine similarity performance.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q8_00.3 GB 1.0 GB
Fits Get file →
Q8_00.3 GB 1.0 GB
Fits Get file →
F160.6 GB 1.3 GB
Fits Get file →
F160.6 GB 1.3 GB
Fits Get file →

For 8 GB VRAM, use Q8_0 (file size 0.33 GB, estimated memory 1 GB at 8k context). If more precision is required and 8 GB VRAM is available, F16 (file size 0.62 GB, estimated memory 1.3 GB) can be used.

What GPU do you need for xVITA Embed 300M zhTW?

These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) 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 F16 (0.6 GB file, about 1.3 GB in use).

Which xVITA Embed 300M zhTW 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 Q8_0 at 0.3 GB, with the quality trade-offs that come with fewer bits. 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

Architecturegemma-embedding
Parameters308M
Layers24
Trained context8,192 tokens
KV cache (fp16)24 KB per token

Source

Quantized byxCloudinfo
FamilyXvita
Licensegemma
Downloads (HF)620
Last updated2026-10-26

Frequently asked questions about xVITA Embed 300M zhTW

How much VRAM does xVITA Embed 300M zhTW need?

About 1.0 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 xVITA Embed 300M zhTW on an 8 GB GPU?

Yes, with F16: the 0.6 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 xVITA Embed 300M zhTW GGUF?

It depends on the quant: Q8_0 is 0.3 GB, Q8_0 is 0.3 GB, F16 is 0.6 GB, F16 is 0.6 GB.

What context length does xVITA Embed 300M zhTW support?

The GGUF header lists a trained context length of 8,192 tokens (8K). The KV cache grows with the context you use: about 0.2 GB at 8K and 0.2 GB at 8K.

Where can I download xVITA Embed 300M zhTW GGUF?

The files are on Hugging Face in the xCloudinfo/xVITA-Embed-300M-zhTW-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 xVITA Embed 300M zhTW 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 xVITA Embed 300M zhTW on 8 GB VRAM?

Yes, the Q8_0 quantization requires approximately 1 GB of VRAM at 8k context, well within 8 GB.

Does this model support English text?

The model is tailored for Traditional Chinese; English support is not guaranteed.

What prompt format does llama.cpp expect?

Use the EmbeddingGemma format: query: `task: search result | query: {your question}` and document: `title: none | text: {document content}`.

Similar models

Browse all →

More from xCloudinfo

All by xCloudinfo →

Other tiny models

All tiny models →