v0.1.3

v0.1.3

Full Changelog: v0.1.2...v0.1.3

v0.1.2

v0.1.2

Full Changelog: v0.1.1...v0.1.2

v0.1.1

v0.1.1

FIX

setup.py

1. fix-code: add numpy requirment in setup.py

dl_torch

1. fix-bug: dl_torch.save_checkpoint err model_root
2. fix-bug: dl_torch.utils.init_model_parameter support None for weight and bias

ADD

base

1. add base.get_time as func to get str-time

dl_torch

1. add dl_torch.optim as sub-module
2. add dl_torch.optim.LrScheduler

v0.1.0

v0.1.0

ADD

stats

test

1. mannwhitneyu: func to do mannwhitneyu

FIX

stats

test

change test funcs to recive param in a uniform style as called mbapy-style

dl_torch

fix dl_torch sub-module code

bb

1. CnnCfg: a config class for simple cnn block
2. reshape: a nn.Module class for doing reshape in nn.Sequential
3. permute: a nn.Module class for doing permute in nn.Sequential
4. ScannCore: block for SCANN
5. SCANN: doing MHSA with cnn style in a 4-D tensor like a typical image input
6. PositionalEncoding: PositionalEncoding
7. PositionwiseFeedforwardLayer: PositionwiseFeedforwardLayer
8. MultiHeadAttentionLayer: MultiHeadAttentionLayer
9. FastMultiHeadAttentionLayer: warpper for FlashMHA in MultiHeadAttentionLayer style
10. OutMultiHeadAttentionLayer: MHSA for reduce dims, [b, l, c] -> [b, D]
11. EncoderLayer: transformer layer
12. OutEncoderLayer: EncoderLayer for OutMultiHeadAttentionLayer
13. Trans: Transformer
14. TransPE: transformer with positional encoding
15. OutEncoderLayerAvg: OutEncoderLayer like, but using avgpool
16. TransAvg: transformer for OutEncoderLayerAvg
17. SeparableConv2d: SeparableConv2d
18. ResBlock: ResBlock
19. ResBlockR: ResBlock like, Identity Mappings in Deep Residual Networks : exclusive gating
20. SABlock: Parallel convolution kernel with residual structure
21. SABlockR: SABlock like, exclusive gating
22. ScannBlock1d: edited from NonLocalBlock and scann_core
23. GenCnn1d: tool func, generate cnn block
24. SABlock1D: SABlock like, 3-D tensor
25. SABlock1DR: SABlock like, 3-D tensor, exclusive gating

m

1. TransCfg: config for a simple transformer
2. LayerCfg: config for a simple layer which contains cnn and transformer
3. calcu_q_len: tool func for caulcu aq_len for Out-style transformer
4. COneDLayer: 3-D tensor
5. MAlayer: 4-D tensor
6. MAvlayer: MAlayer like, with avgpool down-sampling
7. SCANlayer: MAlayer like, with SCANN
8. str2net: dict for get blocks and layers easily
9. MATTPBase: model using MA-style layer
10. COneD: COneDLayer
11. MATTP: MAlayer
12. MAvTTP: MAvlayer
13. MATTPE: MAlayer with pos_embedding
14. SCANNTTP: SCANlayer
15. MATTP_ViT: MAlayer with ViT performance

data

1. denarmalize_img: tool func for denarmalizing a image
2. SubDataSet: a high-level Dataset Loader

v0.0.14

ADD

file

1. update_excel : update a excel(xlsx) file with multi sheets

plot

1. get_palette: func, get a seq of hex colors

stats

1. pca : func, wrap of sklearn.decomposition.PCA

test

1. multicomp_turkeyHSD: do multicomp(turkeyHSD) using statsmodels(pairwise_tukeyhsd)
2. multicomp_dunnett: do multicomp(dunnett) using scipy.stats.dunnett
3. multicomp_bonferroni: do multicomp(bonferroni) using scikit_posthocs

dl_torch

paper

1. NonLocalBlock: Non-local Neural Networks (CVPR 2018)
2. FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness, just import flash_attn package
3. HydraAttention: Efficient Attention with Many Heads, arXiv:2209.07484v1

FIX

plot

1. get_df_data
   now this func support empty list for tags param

BTW

This package reached 3000 downloads according to https://pepy.tech/project/mbapy these days, I kown some mirror image servers make a lots effort in it. But, I just feel good. Self record HAHA.

v0.0.12

v0.0.12

ADD

stats.df sub-module

remove_simi(tag:str, df:pd.DataFrame, sh:float = 1., backend:str = 'numpy', tensor = None):

给定一组数，去除一些(最小数目)数，使任意两数差的绝对值大于或等于阈值
Given a set of numbers, remove some (minimum number) of numbers so that the absolute value of the difference between any two numbers is greater than or equal to the threshold

interp(long_one:pd.Series, short_one:pd.Series):

给定两个pd.Series，一长一短，用线性插值给短的插值，使其长度与长的pd.Series一样
Given two pd.Series, one long and one short, use linear interpolation to give the short one the same length as the long pd.Series

stats.reg sub-module

linear_reg as func

stats.test sub-module

some stats test func

plot

qqplot as func

save_show as func

v0.0.9-PreRelease

ADD

Add dl_torch as sub-module, which contains some class or func and is expected to be helpful with deeplearning coding (with torch). Some of them were written by FAIR or called MAIR (mostly from MoCo).
Add plot.pro_bar_data_R as a wrap

FIX

web.get_url_page uses header now

v0.0.6

ready to install with pypi.
actually it is the v0.0.4, but v0.0.4(and v.0.0.5) can not be installed correctly.

v0.0.4

Update:

plot

stats (subpackage)(in constructing)

v0.0.3-pipy_install

can install qith pip install mbapy==0.0.3
do not require a specific numpy, but still has problems with python 3.10