Train virtual-shard anime parser

Train augmented anime filename parser

Organize parser modules and tools

Clean stale local scripts

Remove structural parser rule assists

Document training lineage

Train thin anime filename parser

Remove stale diagnostics reports

Add inference performance benchmark

Polish Hugging Face repository docs

Clean special code parsing

Fix GM-Team bilingual title parsing

Improve anime filename parser model

Add parser diagnostics and inference debugging

Add Codex Colab training workflow

Add agent repository guidelines

Mirror character vocab in model repo

Add character-token DMHY training path

colab: onnx导出改为非阻塞+补全onnxscript

colab: 加上onnx/onnxruntime依赖

修复colab_train.py shell引号嵌套导致的Syntax error

重建词表：从632K dmhy_weak.jsonl统计频率取top8000，覆盖96.2%

重新训练模型，补充训练结果

提高训练速度并使用GPU进行训练

完成整个数据集的整理

Document AniFileBERT maintenance workflow

Add AniFileBERT model and training project

initial commit