GLM-5.2 is live. #1 throughput on OpenRouter, pay-per-token on FriendliAI. Try it today ➜
Yvthyvq
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
Yvthyvq
Model Tree
Yvthyvq/Liujgoj-Cantonese-gemma-4-12B-it-cpt
Input Modalities
Output Modalities
Supported Functionality
本模型係基於 Yvthyvq/Liujgoj-Cantonese-gemma-4-12B-it-cpt 進行 SFT (Supervised Fine-Tuning) 嘅粵語羅馬字對話模型。
透過 CPT + SFT 兩階段訓練，我哋成功將 Gemma 4 12B 打造成一個能夠 聽懂粵語指令、理解羅馬字、進行多任務對話 嘅粵語 AI 助手。
| 標籤 | 用途 |
|---|---|
<ipa-lexicon_BLOCK> </ipa-lexicon_BLOCK> | IPA 詞條區塊 |
<ipa-syllable_BLOCK> </ipa-syllable_BLOCK> | IPA 音節區塊 |
<lexicon-ipa_BLOCK> </lexicon-ipa_BLOCK> | 詞條轉 IPA 區塊 |
<lexicon-syllable_BLOCK> </lexicon-syllable_BLOCK> | 詞條轉音節區塊 |
<syllable-ipa_BLOCK> </syllable-ipa_BLOCK> | 音節轉 IPA 區塊 |
python
from transformers import AutoTokenizer, AutoModelForCausalLMimport torchmodel_path = "Yvthyvq/Liujgoj-Cantonese-gemma-4-12B-it-sft"tokenizer = AutoTokenizer.from_pretrained(model_path)model = AutoModelForCausalLM.from_pretrained(model_path,torch_dtype=torch.bfloat16,device_map="auto")---## 數據格式所有數據採用統一的 messages 格式：json{"messages": [{"role": "user", "content": "..."},{"role": "assistant", "content": "..."}]}---🙏 致謝Google 團隊開發並開源 Gemma 系列模型Hugging Face 提供 Transformers 同 PEFT 框架所有粵語數據貢獻者
<syllable-lexicon_BLOCK> </syllable-lexicon_BLOCK> | 音節轉詞條區塊 |
<SCENE> </SCENE> | 對話場景區塊 |
<LEX_BLOCK> </LEX_BLOCK> | 詞條解釋區塊 |