Commit Graph
17 Commits
Author SHA1 Message Date
ViperEkura fc47319240 refactor: simplify BaseFactory and separate ModelFactory from AutoModel
- Extract _resolve_base_type and _validate_component as module-level helpers
- Replace ForwardRef._evaluate private API with eval in module namespace
- Remove broad except Exception in __init_subclass__, _component_base always set
- Replace direct _entries mutation in strategy.py with register() call form
- Remove dead TOKENIZER_CLASSES registry from AutoTokenizer
- Extract ModelFactory(BaseFactory[nn.Module]) as pure factory
- AutoModel now inherits only nn.Module, no factory state
- Move @AutoModel.register to @ModelFactory.register in transformer.py and encoder.py
2026-07-30 09:38:20 +08:00
ViperEkura 0b0693a0a2 fix: make ChatTemplate picklable for spawn multiprocessing
- Add __getstate__/__setstate__ to drop cached _compiled Jinja2 template
- Jinja2 Template.root_render_func is a dynamic closure unpicklable by reference
- cached_property rebuilds the template lazily on first render after unpickle
2026-07-29 13:24:13 +08:00
ViperEkura 958df58f9d refactor: unify tokenizer encode and apply_chat_template for batch support
- encode(str) single-thread, encode(List[str]) Rust parallel encode_batch
- apply_chat_template accepts single Messages or List[Messages] for batch
- add Message/Messages type aliases at module level
2026-07-20 17:49:26 +08:00
ViperEkura 17127f8b3c fix: make tokenizer picklable for spawn multiprocessing
- ChatTemplate: defer Jinja2 compilation to cached_property, exclude compiled template from __getstate__ (its dynamic root function has __module__=None and falls back to __main__, breaking pickle)
- AutoTokenizer: bypass __getattr__ for underscore-prefixed attrs to prevent infinite recursion during unpickle when __dict__ is empty
2026-07-19 11:59:55 +08:00
ViperEkura 01ce1fb9e3 refactor : Pipeline 去除去重,ids 重命名为 sequence,泛型透传
- 移除 Pipeline 内置去重逻辑及 dedup_signature 工具函数
- 删除 ProcessingConfig.deduplicate 字段
- builder 返回 'sequence' 替代 'ids',与 dataset 层统一
- pipeline 纯透传,泛型处理任意 key 补齐默认值
2026-05-31 15:14:27 +08:00
ViperEkura 14f83cbdac perf : 预编译 Jinja2 Template,避免每次 render 重新构建 2026-05-31 14:50:16 +08:00
ViperEkura 8f1b32f2b6 fix: 移除多余 request 参数并增强 tokenizer 健壮性
- 路由和 _get_engine 不再需要 request 参数,直接引用模块级 app
- from_pretrained 增加文件完整性校验,缺 tokenizer.json 则抛 FileNotFoundError
- 移除 from_pretrained 中未使用的 **kwargs
2026-05-17 12:52:18 +08:00
ViperEkura 283bcaf2ff fix: 修复 CLI 参数缺失/重复、device_ids 越界、generate 参数名不一致、scheduler 时序、非流式截断等 bug
- train.py: 补上 --batch_size、--grpo_clip_eps,删除 3 处重复 --group_size
- generate.py: --model_dir 改为 --param_path 对齐 README
- automodel.py: from_pretrained 新增 strict 参数(默认 True)
- parallel/setup.py: 修复 device_ids 索引越界
- train_callback.py: scheduler.step() 移至 on_step_end
- test_train_strategy.py: 测试中补 optimizer.step()
- engine.py: 非流式改为循环等待所有任务完成,补 remove_task 清理
- scheduler.py: Task 添加 _pages_freed 标志,杜绝双重释放
- trainer.py: accumulation_steps=0 时 clamp 为 1
- tokenizer.py: save_pretrained 添加 _tokenizer is None 检查
- benchmark.py: 修复 ModelConfig 过时 import 路径
- inference/__init__.py: 修复 stale docstring
2026-05-09 14:36:42 +08:00
ViperEkura bbeaff4c60 refactor: 精简推理引擎代码,优化参数传递规范 2026-04-09 14:17:48 +08:00
ViperEkura ab5e207f42 feat: 增加缓存处理 2026-04-08 20:54:14 +08:00
ViperEkura ace8f6ee68 chore: 优化未使用的模块 2026-04-06 09:54:17 +08:00
ViperEkura a57a16430d fix: 修复tokenizer存储的问题 2026-04-06 09:36:29 +08:00
ViperEkura 3f67e53088 fix: 修复tokenizer 参数问题 2026-04-06 09:22:46 +08:00
ViperEkura 39766aa1dc chore: 修改类名,优化导入顺序 2026-04-05 22:27:57 +08:00
ViperEkura d2b36cc85d fix: 修复特殊token 的问题 2026-04-05 20:09:47 +08:00
ViperEkura fc278d17ab feat: 实现模型动态注册机制 2026-04-05 19:38:12 +08:00
ViperEkura c94a246c71 chore: 重命名目录 2026-04-04 17:03:22 +08:00