AI & ML interests
None defined yet.
Recent Activity
View all activity
以 gpt-oss 為基座的台灣繁體中文對話模型,120B 與 20B 兩種規模。Traditional Chinese chat, gpt-oss based.
船舶、航運、無人船(USV/ROV/AUV)與海事法規的領域模型。Maritime domain models covering vessels, shipping and regulations.
台灣華語語音線:語音辨識(Whisper 微調,繁體中文輸出保證)與語音合成。Taiwanese Mandarin speech: ASR (Whisper fine-tunes with guaranteed Traditional Chinese output) and TTS.
以 gpt-oss-20b 為基座的程式碼生成模型,體積小、可在單卡或邊緣裝置部署。
以 Gemma 為基座、針對台灣繁體中文蒸餾微調的通用對話模型。Gemma-based models distilled for Taiwanese Traditional Chinese.
-
xCloudinfo/Gemma-4-26B-A4B-TAIDE-zhTW
Image-Text-to-Text • 26B • Updated • 28 -
xCloudinfo/Gemma-4-26B-A4B-TAIDE-zhTW-GGUF
Image-Text-to-Text • 25B • Updated • 352 • 1 -
xCloudinfo/Gemma-4-26B-A4B-TAIDE-zhTW-abliterated-GGUF
Image-Text-to-Text • 25B • Updated • 372 -
xCloudinfo/Gemma-3-TAIDE-12b-Chat-GGUF
Text Generation • 13B • Updated • 1
移除殘差流拒答方向(refusal direction)的模型變體,供對齊研究與紅隊測試,非production用途。
-
xCloudinfo/gpt-oss-120b-Uncensored-xCloud
Text Generation • 2B • Updated • 98 -
xCloudinfo/gpt-oss-120b-Uncensored-xCloud-GGUF
Text Generation • 117B • Updated • 119 -
xCloudinfo/gpt-oss-20b-Uncensored-xCloud
Text Generation • 22B • Updated • 20 -
xCloudinfo/gpt-oss-20b-Uncensored-xCloud-GGUF
Text Generation • 21B • Updated • 266
非自研模型的量化轉存,方便地端部署。Quantized re-uploads of third-party models for on-premise deployment.
繁體中文檢索兩段:嵌入(300M 768 維、1B 2048 維可截)與重排序,供 RAG 使用。Traditional Chinese embedding and reranker for RAG.
繁體中文文件與台灣表單的 OCR 模型線,涵蓋印刷與手寫。Traditional Chinese document OCR, print and handwriting.
-
xCloudinfo/Gemma-4-E4B-xReadable-zhTW-OCR-v1
Image-Text-to-Text • 8B • Updated • 4.61k -
xCloudinfo/Llama-xReadable-11B-zhTW-OCR-v3
Image-Text-to-Text • 11B • Updated • 32 -
xCloudinfo/Llama-xReadable-11B-zhTW-OCR-v2
Image-Text-to-Text • 11B • Updated • 9 -
xCloudinfo/Llama-xReadable-11B-zhTW-OCR-v1
Image-Text-to-Text • 11B • Updated • 14
能把整個專案寫完的程式模型線,支援多檔案編輯、建置、除錯迴圈與工具呼叫。Coding agent models for end-to-end project completion.
移除殘差流拒答方向(refusal direction)的模型變體,供對齊研究與紅隊測試,非production用途。
-
xCloudinfo/gpt-oss-120b-Uncensored-xCloud
Text Generation • 2B • Updated • 98 -
xCloudinfo/gpt-oss-120b-Uncensored-xCloud-GGUF
Text Generation • 117B • Updated • 119 -
xCloudinfo/gpt-oss-20b-Uncensored-xCloud
Text Generation • 22B • Updated • 20 -
xCloudinfo/gpt-oss-20b-Uncensored-xCloud-GGUF
Text Generation • 21B • Updated • 266
以 gpt-oss 為基座的台灣繁體中文對話模型,120B 與 20B 兩種規模。Traditional Chinese chat, gpt-oss based.
非自研模型的量化轉存,方便地端部署。Quantized re-uploads of third-party models for on-premise deployment.
船舶、航運、無人船(USV/ROV/AUV)與海事法規的領域模型。Maritime domain models covering vessels, shipping and regulations.
繁體中文檢索兩段:嵌入(300M 768 維、1B 2048 維可截)與重排序,供 RAG 使用。Traditional Chinese embedding and reranker for RAG.
台灣華語語音線:語音辨識(Whisper 微調,繁體中文輸出保證)與語音合成。Taiwanese Mandarin speech: ASR (Whisper fine-tunes with guaranteed Traditional Chinese output) and TTS.
繁體中文文件與台灣表單的 OCR 模型線,涵蓋印刷與手寫。Traditional Chinese document OCR, print and handwriting.
-
xCloudinfo/Gemma-4-E4B-xReadable-zhTW-OCR-v1
Image-Text-to-Text • 8B • Updated • 4.61k -
xCloudinfo/Llama-xReadable-11B-zhTW-OCR-v3
Image-Text-to-Text • 11B • Updated • 32 -
xCloudinfo/Llama-xReadable-11B-zhTW-OCR-v2
Image-Text-to-Text • 11B • Updated • 9 -
xCloudinfo/Llama-xReadable-11B-zhTW-OCR-v1
Image-Text-to-Text • 11B • Updated • 14
以 gpt-oss-20b 為基座的程式碼生成模型,體積小、可在單卡或邊緣裝置部署。
能把整個專案寫完的程式模型線,支援多檔案編輯、建置、除錯迴圈與工具呼叫。Coding agent models for end-to-end project completion.
以 Gemma 為基座、針對台灣繁體中文蒸餾微調的通用對話模型。Gemma-based models distilled for Taiwanese Traditional Chinese.
-
xCloudinfo/Gemma-4-26B-A4B-TAIDE-zhTW
Image-Text-to-Text • 26B • Updated • 28 -
xCloudinfo/Gemma-4-26B-A4B-TAIDE-zhTW-GGUF
Image-Text-to-Text • 25B • Updated • 352 • 1 -
xCloudinfo/Gemma-4-26B-A4B-TAIDE-zhTW-abliterated-GGUF
Image-Text-to-Text • 25B • Updated • 372 -
xCloudinfo/Gemma-3-TAIDE-12b-Chat-GGUF
Text Generation • 13B • Updated • 1