refactor: 修改模型架构

This commit is contained in:
2026-04-16 18:57:21 +08:00
parent cb93f8219e
commit a38334f4ce
17 changed files with 540 additions and 176 deletions
+2 -2
View File
@@ -115,9 +115,9 @@ export_dataset(
### 2. 分词并缓存为 HDF5
```python
from pipeline import BpeTokenizer, ProcessorFactory, cache_jsonl
from pipeline import AutoTokenizer, ProcessorFactory, cache_jsonl
tokenizer = BpeTokenizer("tokenizer.json")
tokenizer = AutoTokenizer("tokenizer.json")
processor = ProcessorFactory.create("pt", tokenizer)
cache_jsonl(