• Joined on 2025-04-24
iomgaa created branch old/lzn in iomgaa/Minimind 2025-05-14 00:05:54 +08:00
iomgaa deleted branch lzn from iomgaa/Minimind 2025-05-14 00:05:54 +08:00
iomgaa created branch old/SLM in iomgaa/Minimind 2025-05-14 00:05:48 +08:00
iomgaa deleted branch SLM from iomgaa/Minimind 2025-05-14 00:05:48 +08:00
iomgaa created branch old/HPC in iomgaa/Minimind 2025-05-14 00:05:41 +08:00
iomgaa deleted branch HPC from iomgaa/Minimind 2025-05-14 00:05:41 +08:00
iomgaa created branch old/master in iomgaa/Minimind 2025-05-14 00:05:31 +08:00
iomgaa deleted branch master-new from iomgaa/Minimind 2025-05-14 00:05:31 +08:00
iomgaa pushed to master at iomgaa/Minimind 2025-05-14 00:03:13 +08:00
iomgaa created branch master in iomgaa/Minimind 2025-05-14 00:03:13 +08:00
iomgaa pushed to master-new at iomgaa/Minimind 2025-05-14 00:02:10 +08:00
089afd6728 DynamicKV-LLM Pretrain v1.1.0
iomgaa pushed to master-new at iomgaa/Minimind 2025-05-13 23:36:20 +08:00
1a8c86360d update
iomgaa pushed to master-new at iomgaa/Minimind 2025-05-13 16:28:05 +08:00
fc688ddde4 删除了与kv cache有关的代码
f31e17030c update
7ba51b8571 位置编码从复数变为两次实数计算
7cf4228401 update
caa9c23bc5 使用accelerate和deepseed替代torchrun
Compare 8 commits »
iomgaa created branch master-new in iomgaa/Minimind 2025-05-13 16:28:05 +08:00
iomgaa pushed to HPC at iomgaa/Minimind 2025-05-12 12:24:06 +08:00
decec67b78 update
iomgaa pushed to SLM at iomgaa/Minimind 2025-05-12 12:16:41 +08:00
77cf24c925 update
iomgaa pushed to SLM at iomgaa/Minimind 2025-05-12 12:11:34 +08:00
6eaca41018 update
iomgaa pushed to HPC at iomgaa/Minimind 2025-05-12 11:53:22 +08:00
d93889194d update
iomgaa pushed to HPC at iomgaa/Minimind 2025-05-12 00:21:12 +08:00
a3ea93597c DynamicKV-LLM Pretrain v1.1.0
iomgaa pushed to HPC at iomgaa/Minimind 2025-05-12 00:15:25 +08:00
da5ac6a5c0 Merge branch 'SLM' into HPC
8dd7cfaf72 修正了loss为nan的错误
cb286d26d1 wandb包含config信息
0c8c6e5d1a 添加了忽视数据库模式
b6bd97aaaa 抽取self.downsample_v与self.downsample_q的共同部分,并使用可分离卷积降低参数量
Compare 5 commits »