Downloads · 30 days
942
29% of all-time downloads
xCloudinfo/Gemma-4-26B-A4B-TAIDE-zhTW-GGUF
Gemma-4-26B-A4B-TAIDE-zhTW-GGUF is a image-text-to-text model from xCloudinfo. Use it for the image-text-to-text task on the model card, and read the license before you ship it in a product. The card lists the license as apache-2.0.
云碩科技 · xCloudinfo · 系列:繁中在地化 · TAIDE zh-TW
Downloads · 30 days
942
29% of all-time downloads
All-time downloads
3.2K
Public
Repo size
128 GB
Likes
1
Public
Click a slice to open those files.
.gguf111 GB · 100%
From the Hugging Face model README
云碩科技 · xCloudinfo · 系列:繁中在地化 · TAIDE zh-TW
xCloudinfo/Gemma-4-26B-A4B-TAIDE-zhTW 的 GGUF(llama.cpp / Ollama) 版本,含視覺投影器(mmproj),可圖文對話。六階量化齊全,依裝置VRAM/CPU記憶體選用。
| 檔案 | 量化 | 大小 | 用途 |
|---|---|---|---|
model-Q8_0.gguf | Q8_0 | 26 GB | 幾乎無損,資源充足時首選 |
model-Q6_K.gguf | Q6_K | 22 GB | 高品質,效能/精度平衡 |
model-Q5_K_M.gguf | Q5_K_M | 18 GB | 品質與體積折衷 |
model-Q4_K_M.gguf | Q4_K_M | 16 GB | 部署首選,一般裝置最常用 |
model-IQ4_XS.gguf | IQ4_XS | 13 GB | 較小體積,帶imatrix校準 |
model-IQ2_M.gguf | IQ2_M | 9.7 GB | 極限壓縮,記憶體受限裝置 |
imatrix.dat | — | 55 MB | 量化用重要性矩陣(IQ系列必須) |
mmproj-Gemma-4-26B-A4B-TAIDE-zhTW-f16.gguf | f16 | 1.2 GB | 視覺投影器(vision) |
llama-server -m model-Q4_K_M.gguf \
--mmproj mmproj-Gemma-4-26B-A4B-TAIDE-zhTW-f16.gguf -c 4096 -ngl 99
純文字可省略
--mmproj;把model-Q4_K_M.gguf換成其他階的檔名即可切換量化等級。
google/gemma-4-26B-A4B-it,Apache-2.0 + Google Gemma 4 License(含Prohibited Use Policy)。taide/Gemma-3-TAIDE-12b-Chat 蒸餾;須遵守 TAIDE模型授權條款 並保留本聲明。云碩科技(xCloudinfo Corp. Limited)專注於地端私有化人工智慧解決方案。xVITA是云碩自研模型的品牌名稱。
由 云碩科技xCloudinfo於自有AI算力資源池製作。
其他模型請見 https://hf.co/xCloudinfo。
xCloudinfo (云碩科技) · Series: Traditional Chinese localization · TAIDE zh-TW
The GGUF (llama.cpp / Ollama) version of xCloudinfo/Gemma-4-26B-A4B-TAIDE-zhTW, including the vision projector (mmproj) for image-and-text conversation. All six quantization tiers are provided; choose one according to your device's VRAM / CPU memory.
| File | Quantization | Size | Use |
|---|---|---|---|
model-Q8_0.gguf | Q8_0 | 26 GB | Nearly lossless; first choice when resources are ample |
model-Q6_K.gguf | Q6_K | 22 GB | High quality; balance of performance and precision |
model-Q5_K_M.gguf | Q5_K_M | 18 GB | Compromise between quality and size |
model-Q4_K_M.gguf | Q4_K_M | 16 GB | First choice for deployment; most commonly used on typical devices |
model-IQ4_XS.gguf | IQ4_XS | 13 GB | Smaller size, with imatrix calibration |
model-IQ2_M.gguf | IQ2_M | 9.7 GB | Extreme compression, for memory-constrained devices |
imatrix.dat | — | 55 MB | Importance matrix used for quantization (required for the IQ series) |
mmproj-Gemma-4-26B-A4B-TAIDE-zhTW-f16.gguf | f16 | 1.2 GB | Vision projector (vision) |
llama-server -m model-Q4_K_M.gguf \
--mmproj mmproj-Gemma-4-26B-A4B-TAIDE-zhTW-f16.gguf -c 4096 -ngl 99
For text only,
--mmprojcan be omitted; replacemodel-Q4_K_M.ggufwith another tier's file name to switch the quantization level.
google/gemma-4-26B-A4B-it, Apache-2.0 + Google Gemma 4 License (including the Prohibited Use Policy).taide/Gemma-3-TAIDE-12b-Chat; the TAIDE model license terms must be followed and this notice retained.xCloudinfo Corp. Limited (云碩科技) builds on-premises, private AI solutions. xVITA is xCloudinfo's in-house model brand.
Made by xCloudinfo (云碩科技) on its own AI compute resource pool.
More models: https://hf.co/xCloudinfo