MorningForest
|
ec17562f20
|
修改fdl 的prepare_*_dataloader,使其支持fastnlp的dataset
|
2022-06-04 15:08:50 +08:00 |
|
lxr-tech
|
3797f91434
|
update tutorial-3456 lxr 220603
|
2022-06-04 00:03:40 +08:00 |
|
yhcc
|
82b06767f5
|
修复n_batches导致的bug
|
2022-06-03 22:35:46 +08:00 |
|
MorningForest
|
745d254aa5
|
Merge branch 'dev0.8.0' of github.com:fastnlp/fastNLP into dev0.8.0
|
2022-06-03 21:01:46 +08:00 |
|
MorningForest
|
af6300e816
|
修改element,计算值转为非tensor形式
|
2022-06-03 21:01:39 +08:00 |
|
yhcc
|
21b7b58793
|
typo fix
|
2022-06-03 20:21:58 +08:00 |
|
yhcc
|
281ffc3b91
|
修复richcallback等在断点重训时显示进度不正确的问题
|
2022-06-03 20:09:11 +08:00 |
|
yhcc
|
fecd82aadf
|
修改element无法与element运算的问题
|
2022-06-03 19:51:54 +08:00 |
|
x54-729
|
f3a13ff885
|
Merge branch 'dev0.8.0' of github.com:fastnlp/fastNLP into dev0.8.0
|
2022-06-03 09:58:52 +00:00 |
|
x54-729
|
4d29524d81
|
删除初始化的种子设置函数
|
2022-06-03 09:54:33 +00:00 |
|
x54-729
|
2303079180
|
1. 删除 FASTNLP_SEED_WORKERS 2.修改各个框架的 seed_everything 函数,使得用户可以设置种子复现训练结果 3. 修改 sampler 的 seed 参数,在为 None 时可以根据时间戳生成随机数种子
|
2022-06-03 09:53:25 +00:00 |
|
yhcc
|
057806de7d
|
将total_batches改名为n_batches,以和n_epochs对应;增加n_batches作为Trainer的初始化参数
|
2022-06-02 22:52:14 +08:00 |
|
lxr-tech
|
fd91dc3734
|
update tutorial-045 lxr 220602
|
2022-06-02 22:43:48 +08:00 |
|
x54-729
|
a16a0e94f5
|
driver 在 dataloader 为空时进行警告
|
2022-06-01 16:19:00 +00:00 |
|
x54-729
|
d43159a576
|
Merge branch 'dev0.8.0' of github.com:fastnlp/fastNLP into dev0.8.0
|
2022-06-01 16:11:28 +00:00 |
|
lxr-tech
|
2d838b8927
|
update tutorial-45 lxr 220601
|
2022-06-01 23:15:21 +08:00 |
|
lxr-tech
|
e55ca8a1d1
|
update example-12 lxr 220531
|
2022-05-31 23:25:01 +08:00 |
|
yhcc
|
64c7ce5468
|
1.修复rich在jupyter的时候不打印的问题;2.修复若干其他bug
|
2022-05-31 22:44:32 +08:00 |
|
x54-729
|
c6cf4da25b
|
调整部分文档格式
|
2022-05-31 09:45:25 +00:00 |
|
x54-729
|
809b83b6c0
|
small
|
2022-05-31 04:15:54 +00:00 |
|
x54-729
|
f1cb6f6167
|
small
|
2022-05-31 03:56:39 +00:00 |
|
x54-729
|
79110d7c84
|
Merge branch 'dev0.8.0' of github.com:fastnlp/fastNLP into dev0.8.0
|
2022-05-31 03:47:42 +00:00 |
|
lxr-tech
|
c62f5bd2dc
|
update example-12 lxr 220530
|
2022-05-30 22:48:28 +08:00 |
|
yhcc
|
cef9dca103
|
修复bug
|
2022-05-30 21:02:29 +08:00 |
|
x54-729
|
edda3d3196
|
1.处理新测试例的import问题 2.将warnings.warn替换为logger.warn
|
2022-05-30 06:44:06 +00:00 |
|
yhcc
|
6d889c8624
|
Merge branch 'dev0.8.0' of github.com:fastnlp/fastNLP into dev0.8.0
|
2022-05-29 23:19:46 +08:00 |
|
yhcc
|
004c344e4c
|
1.增加torch modules;torch models
|
2022-05-29 23:19:40 +08:00 |
|
x54-729
|
9b0a30c8fb
|
small bug
|
2022-05-29 13:51:10 +00:00 |
|
x54-729
|
aeeb1bc415
|
Merge branch 'dev0.8.0' of github.com:fastnlp/fastNLP into dev0.8.0
|
2022-05-29 12:23:40 +00:00 |
|
yhcc
|
639f57ad8b
|
update topk_saver.py
|
2022-05-29 19:25:07 +08:00 |
|
x54-729
|
badd07f8f1
|
修改 ReproducibleSampler ReproducibleBatchSampler 的 num_samples 获取方式
|
2022-05-29 07:04:24 +00:00 |
|
yh
|
1d665c9480
|
unrepeatedSampler支持chunk_split
|
2022-05-29 14:41:59 +08:00 |
|
x54-729
|
f7b3fe6a06
|
small
|
2022-05-29 02:04:05 +00:00 |
|
x54-729
|
e402893017
|
Merge branch 'dev0.8.0' of github.com:fastnlp/fastNLP into dev0.8.0
|
2022-05-29 01:32:31 +00:00 |
|
x54-729
|
75a3278d69
|
1. 完成JittorSingleDriver的功能,并添加测试用例 2.在Sampler中添加属性num_samplers 用于动态获取dataset的长度 3.添加便于测试断点重训的数据集 4.修改jittor其它测试的一些bug,统一ArgMaxDataset 的命名
|
2022-05-29 01:32:16 +00:00 |
|
lxr-tech
|
ec76ba8887
|
update example-2 lxr 220528
|
2022-05-28 17:06:57 +08:00 |
|
lxr-tech
|
19a48c7101
|
update example-12 lxr 220527
|
2022-05-27 22:47:11 +08:00 |
|
yh
|
49e8ae2daa
|
Merge branch 'dev0.8.0' of github.com:fastnlp/fastNLP into dev0.8.0
|
2022-05-27 12:04:43 +08:00 |
|
yh
|
6903253e59
|
1.修改了发
f_rich_progress在没有bar的时候暂时关闭live;2.修改sampler获取dataset长度的方式以适配jittor
|
2022-05-27 12:04:37 +08:00 |
|
lxr-tech
|
8f211ef8ab
|
add example-12 lxr 220526
|
2022-05-26 21:57:44 +08:00 |
|
yh
|
1ec3337a49
|
Merge branch 'dev0.8.0' of github.com:fastnlp/fastNLP into dev0.8.0
|
2022-05-26 16:04:02 +08:00 |
|
yh
|
0be2357f46
|
优化了monitor相关callback的信息打印
|
2022-05-26 16:03:56 +08:00 |
|
x54-729
|
a6cfc4086f
|
修改seq_len_to_mask的jittor实现及测试例中的一处传参错误
|
2022-05-25 17:34:32 +00:00 |
|
x54-729
|
b29b002bcc
|
修改 paddle_to 中关于 device 类型的文档
|
2022-05-25 17:17:58 +00:00 |
|
x54-729
|
999520e4b0
|
Merge branch 'dev0.8.0' of github.com:fastnlp/fastNLP into dev0.8.0
|
2022-05-25 17:14:50 +00:00 |
|
x54-729
|
28eb1a5836
|
添加了对tensor_to_numeric reduce 参数的测试
|
2022-05-25 17:14:01 +00:00 |
|
x54-729
|
efc2675741
|
1.修改Driver.py的小部分文档;2.完善 JittorDriver JittorSingleDriver 的部分基础函数
|
2022-05-25 17:03:12 +00:00 |
|
yh
|
7a7d75d100
|
Merge branch 'dev0.8.0' of github.com:fastnlp/fastNLP into dev0.8.0
|
2022-05-26 00:05:42 +08:00 |
|
yh
|
dcdd484eb2
|
增加seq_len_to_mask
|
2022-05-26 00:05:36 +08:00 |
|
x54-729
|
abeb22c700
|
Merge branch 'dev0.8.0' of github.com:fastnlp/fastNLP into dev0.8.0
|
2022-05-25 15:30:35 +00:00 |
|