Commit Graph

8114 Commits

Author SHA1 Message Date
JavaZero
6f25245ef7 fix: adjust chunking logic in _slice_tensor_to_shards for tensor distribution 2025-12-03 20:44:26 +08:00
i-robot
487251c166 !7791 【master】【bugfix】修复build_optim把type、used_fused这些信息pop掉了,导致后面代码无法获取到这些信息
Merge pull request !7791 from JavaZero/fix_muon_auto_callback
2025-12-03 11:38:02 +00:00
i-robot
d63b217451 !7777 【master】【bugfix】修复FlashAttention里面的reducemax没有配置shard
Merge pull request !7777 from JavaZero/fix_reduce_max
2025-12-03 08:11:14 +00:00
i-robot
53d6451e85 !7734 【master】增加load_checkpoint_utils.py以及run_check.py的测试用例
Merge pull request !7734 from AAA碧根果批发赵少/testcase
2025-12-03 07:59:10 +00:00
JavaZero
a039ef294f ensure config is copied in build_optim 2025-12-03 11:59:01 +08:00
yiyison
b446a39fe7 增加load_checkpoint_utils.py以及run_check.py的测试用例 2025-12-03 10:53:03 +08:00
i-robot
0d0565acac !7774 【master】增加adamw.py测试用例
Merge pull request !7774 from AAA碧根果批发赵少/test_adamw
2025-12-03 02:09:06 +00:00
i-robot
2d5147f4a9 !7790 【master】【bug-fix】解决q_lora_rank为None时,跑推理任务权重加载不上的问题
Merge pull request !7790 from zhouxq/weight_q_lora_rank_master
2025-12-03 01:31:00 +00:00
zxq
2ef148e418 【master】【bug-fix】解决q_lora_rank为None时,跑推理任务权重加载不上的问题 2025-12-02 20:59:36 +08:00
i-robot
fbe98f2e18 !7792 【bugfix】【master】test_checkpoint测试用例bugfix
Merge pull request !7792 from AAA碧根果批发赵少/bugfix
2025-12-02 12:55:37 +00:00
yiyison
b45dd45901 test_checkpoint测试用例bugfix 2025-12-02 17:01:44 +08:00
i-robot
549a2ae2c7 !7723 【master】权重2.0日志打印优化
Merge pull request !7723 from AAA碧根果批发赵少/weight2
2025-12-02 06:40:00 +00:00
i-robot
584dc5cd5e !7788 【master】【bug-fix】修改文档中的拼写错误
Merge pull request !7788 from zhouxq/code_docs_bug_master
2025-12-02 06:27:28 +00:00
i-robot
2a4488ff2e !7772 新增 blended_megatron_dataset_builder 测试用例
Merge pull request !7772 from zzzkeke/new/add_builder_test
2025-12-02 03:46:38 +00:00
zxq
e2ee4478fb 【master】【bug-fix】修改文档中的拼写错误 2025-12-02 09:59:46 +08:00
i-robot
3874984d5c !7766 增加 gpt_dataset 测试用例
Merge pull request !7766 from zzzkeke/new/add_test
2025-12-02 01:54:47 +00:00
yiyison
da696d857c 增加adamw.py测试用例 2025-12-02 09:29:03 +08:00
i-robot
bc4ed9d124 !7758 【master】增加checkpoint.py测试用例
Merge pull request !7758 from AAA碧根果批发赵少/test_ckpt_py
2025-12-01 13:00:46 +00:00
i-robot
3eff54dbde !7760 【master】增加model_mixin.py测试用例
Merge pull request !7760 from AAA碧根果批发赵少/test_model_mixin_py
2025-12-01 13:00:34 +00:00
i-robot
3d6fdb117c !7784 【master】【bugfix】【文档】DeepSeek-V3离线脚本文档修改
Merge pull request !7784 from SaiYao/code_docs_improve_reading_fluency_20251201
2025-12-01 12:54:39 +00:00
yiyison
1c97f6d4ae 日志打印优化 2025-12-01 20:51:01 +08:00
i-robot
f1d287320e !7762 【master】【用例】为LayerSetting增加swap用例
Merge pull request !7762 from kongziyi/swap_ut
2025-12-01 12:24:22 +00:00
SaiYao
6e8243346a 【master】【bugfix】【文档】DeepSeek-V3离线脚本文档修改 2025-12-01 20:15:32 +08:00
i-robot
a9ae855ca3 !7780 【master】【bugfix】【文档】文档通顺度修复
Merge pull request !7780 from SaiYao/code_docs_improve_reading_fluency_20251201
2025-12-01 11:50:53 +00:00
SaiYao
217fbfd67f 【master】【bugfix】【文档】文档通顺度修复 2025-12-01 17:02:11 +08:00
yiyison
3783cd7f3e 增加model_mixin.py测试用例 2025-12-01 15:58:21 +08:00
i-robot
fcf881cdf3 !7756 【master】【覆盖率】增加转权重脚本用例
Merge pull request !7756 from zyw_hw/add_test_convert_weight_case
2025-12-01 07:34:49 +00:00
i-robot
6a5ecd811d !7770 【master】【覆盖率】增加tokenizer用例
Merge pull request !7770 from zyw_hw/add_tokenizer_cases
2025-12-01 07:34:31 +00:00
JavaZero
4dc65d8960 fix redistribution op in flash_attn 2025-12-01 15:27:03 +08:00
kongziyi
177ecf1cec 【master】【用例】为LayerSetting增加swap用例 2025-12-01 15:24:33 +08:00
zzzkeke
a5c1669074 新增 blended_megatron_dataset_builder 测试用例 2025-12-01 15:04:19 +08:00
i-robot
76249cf34d !7722 【master】【bugfix】fix profiler step question
Merge pull request !7722 from zyw_hw/fix_profiler_step_ques
2025-12-01 07:02:39 +00:00
i-robot
1d007c3205 !7753 【master】【覆盖率】修复函数级用例执行时用例报错问题
Merge pull request !7753 from zyw_hw/fix_tokenizer_case_bug
2025-12-01 04:33:03 +00:00
i-robot
d71ace34dc !7767 新增callback测试用例
Merge pull request !7767 from lan/callback_test
2025-12-01 02:11:22 +00:00
zyw_hw
55e6743837 add convert weight test cases 2025-12-01 10:03:43 +08:00
zyw_hw
04872b4531 add tokenizer cases 2025-12-01 10:02:28 +08:00
lanxiang
36983773ef 新增callback测试用例 2025-11-29 18:15:55 +08:00
i-robot
2d6206ed09 !7764 【master】【bugfix】trainingstatemonitor文档注释规范修改
Merge pull request !7764 from 魏琢艺/code_docs_ts
2025-11-29 09:38:09 +00:00
i-robot
b046aff038 !7754 【master】【bugfix】添加TokenDispatcher用例
Merge pull request !7754 from 魏琢艺/token_testcase
2025-11-29 09:34:36 +00:00
i-robot
3cb502b25f !7759 【master】【bugfix】【UT】添加sharded_tensor的UT用例
Merge pull request !7759 from SaiYao/add_sharded_tensor_ut
2025-11-29 09:33:31 +00:00
魏琢艺
daba22f898 trainingstatemonitor doc fix 2025-11-29 17:15:29 +08:00
yiyison
f62c549a84 增加checkpoint.py测试用例 2025-11-29 17:05:48 +08:00
zzzkeke
1018297846 Add gpt dataset test UT 2025-11-29 16:00:55 +08:00
i-robot
f478b4bfec !7752 【master】【bugfix】增加muon优化器开启时dp>=op和swap=False的校验
Merge pull request !7752 from kongziyi/fix_muon
2025-11-29 07:56:54 +00:00
i-robot
63c3d85881 !7738 【test】【master】add testcase for trainer
Merge pull request !7738 from hsshuai/test/master/trainer
2025-11-29 07:05:00 +00:00
SaiYao
b0e4c234ec 【UT】添加sharded_tensor的UT用例 2025-11-29 15:01:21 +08:00
i-robot
7c9c9de11c !7642 【bugfix】【master】绑核配置非法device id拦截
Merge pull request !7642 from AAA碧根果批发赵少/affinity
2025-11-29 06:40:15 +00:00
kongziyi
a6a71000d5 【master】【bugfix】增加muon优化器开启时dp>=op和swap=False的校验 2025-11-29 14:37:21 +08:00
魏琢艺
ecf9b0e0da add TokenDispatcher testcase 2025-11-29 12:56:27 +08:00
i-robot
268a799b38 !7724 【master】【feature】升级transformers版本4.51.3->4.57.1
Merge pull request !7724 from Yule100/tf_version
2025-11-29 03:05:36 +00:00