Commit Graph

143 Commits (7840e6f2704191e3645a696990e852e5c05cbe3b)

Author SHA1 Message Date
gongjy 7840e6f270 update model 2024-10-20 15:13:58 +08:00
gongjy 17f6021d32 update num_workers 2024-10-19 06:31:15 +08:00
gongjy 4e7bf23414 update sft 2024-10-16 22:58:06 +08:00
gongjy 4556cd5386 update rlhf 2024-10-15 15:20:31 +08:00
gongjy 829689cb55 update rlhf 2024-10-15 15:04:38 +08:00
gongjy 25e505ccff update rlhf 2024-10-15 10:30:32 +08:00
gongjy 3e424076fb update trl 2024-10-13 22:44:28 +08:00
gongjy 42c36f4916 update inference 2024-10-13 13:11:41 +08:00
gongjy f54130aa8e update inference 2024-10-13 13:09:52 +08:00
gongjy 7d720f5ea2 update data_process 2024-10-12 19:46:08 +08:00
gongjy 0fbeba5e5d update data_process 2024-10-12 18:47:08 +08:00
gongjy fed0ccc8c2 update lora-sft 2024-10-11 17:43:52 +08:00
gongjy c19d111cea update readme 2024-10-10 22:13:10 +08:00
gongjy a775e8ddd8 update readme 2024-10-09 09:12:55 +08:00
gongjy 89f7b18035 update readme 2024-10-09 00:03:07 +08:00
gongjy e232ed82bb update readme 2024-10-08 23:40:29 +08:00
gongjy 93853be90a update readme 2024-10-08 23:16:29 +08:00
gongjy 0fe2351882 update readme 2024-10-07 18:42:53 +08:00
gongjy 36e70d30e2 update readme 2024-10-05 22:59:00 +08:00
gongjy acea505af0 update readme 2024-10-05 00:35:54 +08:00
gongjy 571fec6fdb update requirements.txt 2024-10-03 14:45:58 +08:00
gongjy 5e67759d1a update model (fix loss bug) 2024-09-29 16:58:48 +08:00
gongjy 1119025ac5 fix deepspeed local_rank bug 2024-09-27 23:25:56 +08:00
gongjy 43a7652134 Update data preprocessing methods 2024-09-27 22:34:30 +08:00
gongjy 682003cec6 Update data preprocessing methods 2024-09-27 17:19:03 +08:00
gongjy a80e8e5f52 update readme info 2024-09-27 16:38:18 +08:00
gongjy f351a61c8a Update data preprocessing methods 2024-09-27 16:19:30 +08:00
gongjy 89b91ceaa0 update batchsize 2024-09-25 12:35:29 +08:00
gongjy fc1400909a fix dtype bug 2024-09-25 10:07:30 +08:00
jingyaogong a0081c1931 Merge pull request #44 from iomgaa-ycz/wandb
修复wandb bug & 添加了argparse
2024-09-24 14:10:05 +08:00
jingyaogong 53379b9912 Merge branch 'master' into wandb 2024-09-24 14:09:54 +08:00
Yu Chengzhang b43de06a70 更新了ReadMe 2024-09-24 12:45:21 +08:00
Yu Chengzhang adbfdad1c1 添加了argparse,方便命令行输入参数 2024-09-24 12:41:58 +08:00
Yu Chengzhang 9172ec2f15 修复了wandb的bug,避免了多次产生项目 2024-09-24 11:43:30 +08:00
gongjy f24f8bdc63 update wandb monitor 2024-09-23 22:16:21 +08:00
gongjy dc7d1066e8 update wandb monitor 2024-09-23 22:14:52 +08:00
jingyaogong e63ef387ad Merge pull request #43 from iomgaa-ycz/wandb
use wandb to monitor training process
2024-09-23 21:54:48 +08:00
Yu Chengzhang 4895d77540 添加了wandb 2024-09-23 20:11:45 +08:00
gongjy 4b8df4a7e0 fix data_process bug 2024-09-23 20:11:19 +08:00
jingyaogong 130293a11f Merge pull request #42 from iomgaa-ycz/wandb
修复data_process.py文件追加的bug
2024-09-23 20:08:13 +08:00
Yu Chengzhang 9291c7c8e9 更新了忽视列表 2024-09-23 18:36:04 +08:00
Yu Chengzhang 19a2428b60 修复了process_seq_monkey写入错误的bug 2024-09-23 18:34:58 +08:00
gongjy 44813d5f85 update data_process 2024-09-22 21:17:05 +08:00
gongjy 77a0528bbf update readme 2024-09-22 14:49:05 +08:00
gongjy 58999b591e update readme 2024-09-22 14:45:37 +08:00
gongjy 589a22f451 update readme 2024-09-21 23:11:02 +08:00
gongjy 8f0fd7affc update data_process Chunk_size Read 2024-09-21 22:57:22 +08:00
gongjy ef45699989 update model mask 2024-09-21 20:00:25 +08:00
gongjy e892df833b Efficient implementation of Inference KV cache 2024-09-21 00:01:05 +08:00
gongjy 2d4163a98a update readme 2024-09-20 21:21:25 +08:00