PyTorch fast

August 26, 2019 by admin

[D] Why is PyTorch as fast as (and sometimes faster than) TensorFlow? from MachineLearning

ShujianFollow

Retweet on Twitter Shujian Retweeted

CNBC@CNBC·

16 Jun

Pfizer CEO sees a return to 'normal' globally by the end of 2022 https://cnb.cx/3xq36Bt

Reply on Twitter 1405191025279262729 Retweet on Twitter 140519102527926272958 Like on Twitter 1405191025279262729143 Twitter 1405191025279262729

Shujian @Shujian_Liu·

16 Jun

Rules! My favorite: Controlling Neural Networks with Rule Representations https://arxiv.org/abs/2106.07804

Reply on Twitter 1405170194058776578 Retweet on Twitter 1405170194058776578 Like on Twitter 1405170194058776578 Twitter 1405170194058776578

Retweet on Twitter Shujian Retweeted

Tanishq Gautam @Closedcv·

15 Jun

Finally back on @kaggle!
Worked on Swin Transformers with CutMix augmentation. I also show you how to use @huggingface's Accelerate to train @PyTorch models with GPU and AMP, along with @weights_biases for visualising and tracking my evaluation metrics.
https://www.kaggle.com/ligtfeather/swinformer-cutmix-amp-accelerate-w-b/output?scriptVersionId=65790719

Reply on Twitter 1404716120376967169 Retweet on Twitter 140471612037696716921 Like on Twitter 140471612037696716999 Twitter 1404716120376967169

Retweet on Twitter Shujian Retweeted

Jo Kristian Bergum @jobergum·

15 Jun

The multiphase retrieval and ranking pipeline described in the paper "Pre-trained Language Model for Web-scale Retrieval in Baidu Search" https://arxiv.org/abs/2106.03373

Could easily be represented with http://Vespa.ai
- Hybrid text and embedding retrieval
- GBDT re-ranking

Reply on Twitter 1404701905176125442 Retweet on Twitter 14047019051761254423 Like on Twitter 140470190517612544218 Twitter 1404701905176125442

Retweet on Twitter Shujian Retweeted

Vladimir Haltakov @haltakov·

9 Jun

Can you detect COVID-19 using Machine Learning? 🤔

You have an X-ray or CT scan and the task is to detect if the patient has COVID-19 or not. Sounds doable, right?

None of the 415 ML papers published on the subject in 2020 was usable. Not a single one!

Let's see why 👇

Reply on Twitter 1402717293801508867 Retweet on Twitter 14027172938015088671176 Like on Twitter 14027172938015088673539 Twitter 1402717293801508867