hoho2b

icon location Beijing, China icon location Fun Hunter | Biotech PhD | Data Painter | Deep Learner

Results 19 comments of hoho2b

Mixup or CutMix augmentations

Ref PR #1147

SSL certificate is expired

Solution: `import ssl` `ssl._create_default_https_context = ssl._create_unverified_context` _Originally posted in https://github.com/Cadene/pretrained-models.pytorch/issues/193#issuecomment-635730515_

limit number of loadings in pl.pca_loadings

- Fixed by #2075

Syllabus parsing won't use proxy: Can't parse algorithms-part1 syllabus

I think #675 or #682 may solve this issue.

Instantiating densenet fails

Hi, @mahehu @manideep2510 I checked this issue and found out that every model in [models_factory.py](classification_models/models_factory.py) that defined by `keras_applications` has the same issue: https://github.com/qubvel/classification_models/blob/d445959e742385cd70d5f7c27f224984f2105629/classification_models/models_factory.py#L30-L32 https://github.com/qubvel/classification_models/blob/d445959e742385cd70d5f7c27f224984f2105629/classification_models/models_factory.py#L39-L59 I think this problem caused...

Model output shape does not make sense.

More details: https://github.com/nanoporetech/bonito/issues/101#issuecomment-754611097

[BUG] Float coordinate cause error (ValueError: invalid literal for int() with base 10)

- This code seems to be **NOT** actively maintained - Fixed by #11 **or** #13

Understanding the output shape and scores.

This comment might help you: https://github.com/nanoporetech/bonito/issues/101#issuecomment-754611097

Understanding the output shape and scores.

@mamdouhJ That's because the v3 model has a Conv1D layer with stride=5 to "downsample/compression" the input features: https://github.com/nanoporetech/bonito/blob/651b9c14bbac7fb8a1b01db1610dde97eb4d2d76/bonito/models/configs/dna_r9.4.1%40v3.toml#L16-L18 https://github.com/nanoporetech/bonito/blob/d0cf56f5053494d722b3349319928b140cf3ae33/bonito/crf/model.py#L149-L155

scgen Literal deprecation error

- related to #76