This website requires JavaScript.
Explore
Help
Register
Sign In
ViperEkura
0 Followers
·
0 Following
Joined on
2026-04-02
Repositories
8
Projects
Packages
Public Activity
Starred Repositories
ViperEkura
pushed tag
v1.3.5
to
ViperEkura/AstrAI
2026-05-19 21:23:29 +08:00
ViperEkura
pushed to
main
at
ViperEkura/AstrAI
2026-05-19 21:23:04 +08:00
785d65436c
fix: 修复 to_dict list 类型丢失与 OpenAI stop 参数失效
64be81b7b3
feat: ProgressBarCallback 支持日志行输出到 stdout
45479b5731
feat: metric 参数通过 TrainConfig 传递
e0a3337c22
docs: 更新视频链接
812238060b
fix: docker-compose UID/GID 添加默认值,修复 docker.sh logs 命令
Compare 6 commits »
ViperEkura
pushed tag
v1.3.6
to
ViperEkura/AstrAI
2026-05-19 21:23:04 +08:00
ViperEkura
pushed to
main
at
ViperEkura/video-promo
2026-05-19 16:08:22 +08:00
d471cfa276
add hook.py: terminal demo + claim + logo reveal opening scene (~7s)
ViperEkura
pushed to
main
at
ViperEkura/postgraduate-prep
2026-05-18 21:57:50 +08:00
4e51345415
feat: 新增王道CN Ch3数据链路层笔记(组帧/CRC/流量控制/MAC/以太网/PPP/HDLC/交换机)
ViperEkura
pushed to
main
at
ViperEkura/postgraduate-prep
2026-05-18 21:54:21 +08:00
8e89ff2764
feat: 增加错题04-06(变限积分求导/|x-t|奇偶性/√x²绝对值陷阱)
ViperEkura
pushed to
main
at
ViperEkura/video-promo
2026-05-18 15:51:06 +08:00
12d587aa92
refactor continuous_batching: 4-lane pipeline (PENDING/PREFILL/DECODE/FINISHED), remove bogus Trans arrow, Refill=admission per AstrAI arch
ViperEkura
pushed to
main
at
ViperEkura/AstrAI
2026-05-17 21:04:09 +08:00
6c8533f1d2
docs: 修正文档中类名/字段名与代码不一致之处
ViperEkura
pushed to
main
at
ViperEkura/AstrAI
2026-05-17 20:30:52 +08:00
d0e3464663
docs: 修正文档中类名/字段名与代码不一致之处
2c2697390d
feat: 新增 GradientCheckpointingCallback
7621f05d3f
docs: AdamW beta 默认值改为 (0.9, 0.95)
Compare 3 commits »
ViperEkura
pushed to
main
at
ViperEkura/AstrAI
2026-05-17 16:45:35 +08:00
10ebd7211f
feat: 新增 Muon 优化器
42a391f0fb
feat: 训练中新增验证循环
97c7ac0f4f
refactor: Transformer更名为AutoRegressiveLM并新增EmbeddingEncoder
8f1b32f2b6
fix: 移除多余 request 参数并增强 tokenizer 健壮性
c241a5dcef
refactor: 优化并行训练配置与启动管理
Compare 5 commits »
ViperEkura
pushed to
main
at
ViperEkura/AstrAI
2026-05-17 12:04:04 +08:00
44dab27fdc
feat: 数据集加载时校验必填字段
a44fd22a99
fix: 修复训练与模型参数传递问题
8a11a7d444
fix: 修复训练脚本两处参数传递问题
1d54491809
refactor: 改用递归子模块 init 替代统一 normal_(0.006)
Compare 4 commits »
ViperEkura
pushed to
main
at
ViperEkura/AstrAI
2026-05-17 10:31:56 +08:00
ad9f4d9cf6
refactor: generate_ar 改用流式输出并去除冗余注释
e1638a7ade
fix: 修正AdamW超参数默认值与文档示例
f91bfee33e
refactor: Config序列化统一BaseConfig基类
d7a7f570ed
refactor: 训练循环改为两重迭代并统一参数命名
7dea929788
refactor: checkpoint 按 HF 方式存独立 .pt 文件,callback 接管恢复
Compare 10 commits »
ViperEkura
pushed to
main
at
ViperEkura/AstrAI
2026-05-15 23:39:25 +08:00
3d12a03909
docs : 拆分文档并补充类图缺失类和关系线
ViperEkura
pushed to
main
at
ViperEkura/AstrAI
2026-05-15 22:56:13 +08:00
c169659611
docs: 修正 assets/docs/ 类图、数据流、参数文档及贡献指南
ViperEkura
pushed to
main
at
ViperEkura/AstrAI
2026-05-15 21:37:58 +08:00
e12f1a7ee5
feat: BaseModelConfig + DeepSeekMoE + 工厂模式替代 if/else
ef25efffa2
refactor: 拆分 module.py 为 components 子包
19532440b4
chore: 版本号升至 1.3.5
Compare 3 commits »
ViperEkura
pushed to
main
at
ViperEkura/AstrAI
2026-05-15 18:11:08 +08:00
9096e413c3
refactor: RotaryEmbedding 合并 cos/sin 为单一复数缓存
9d5e9fa6c4
perf: DDP 加 gradient_as_bucket_view/static_graph/broadcast_buffers,AdamW fused
08dde46778
fix: 修复训练循环 step/backward 顺序,重构为三重循环嵌套
Compare 3 commits »
ViperEkura
pushed to
main
at
ViperEkura/AstrAI
2026-05-14 21:38:17 +08:00
513f1f7826
perf: waiting_queue 改用 deque,pull_candidates 从 O(n²) 降到 O(1)
ViperEkura
pushed to
main
at
ViperEkura/AstrAI
2026-05-14 21:31:43 +08:00
e3382f6bb5
fix: 修复推理引擎 batch decode 中多项正确性与并发问题
ViperEkura
pushed to
main
at
ViperEkura/AstrAI
2026-05-14 21:27:36 +08:00
29b5717a38
fix: 修复推理引擎 batch decode 中多项正确性与并发问题
ViperEkura
pushed to
main
at
ViperEkura/AstrAI
2026-05-14 21:00:44 +08:00
f0339022c1
fix: batch 推理示例添加 chat template 和 system prompt
First
Previous
...
3
4
5
6
7
...
Next
Last