#HDFC

Upward sloping parallel channel- characterized by HH-HL's.

Resistance zone: 2350-2372
Support: 2240

Important paper from Google on large batch optimization. They do impressively careful experiments measuring # iterations needed to achieve target validation error at various batch sizes. The main "surprise" is the lack of surprises. [thread]

https://t.co/7QIx5CFdfJ

The paper is a good example of lots of elements of good experimental design. They validate their metric by showing lots of variants give consistent results. They tune hyperparamters separately for each condition, check that optimum isn't at the endpoints, and measure sensitivity.

They have separate experiments where the hold fixed # iterations and # epochs, which (as they explain) measure very different things. They avoid confounds, such as batch norm's artificial dependence between batch size and regularization strength.

When the experiments are done carefully enough, the results are remarkably consistent between different datasets and architectures. Qualitatively, MNIST behaves just like ImageNet.

Importantly, they don't find any evidence for a "sharp/flat optima" effect whereby better optimization leads to worse final results. They have a good discussion of experimental artifacts/confounds in past papers where such effects were reported.

Ever wonder how traders get in stocks before they explode? 🧐

Or find options that move 1,000% in a single day/week?🤯

I made a thread going over exactly how I find these stocks BEFORE they take off 🚀
👇👇👇

So, let's start with the software I use…

Finviz!! Finviz is a free website that I use to scan for stocks.

They have a lot of other cool features as well, but I only use their scanner.
👇

Head on over to finviz and go to the screener

Click on the “All” tab on the top and put in the criteria in the next tweet
👇

This is the first scan that I use for bullish swing setups
✅Quarter +10%
✅Week Down
✅Vol over 400k
✅Optionable
✅Price > $20
👇

The next scan I use is the opposite of the first one.

Simply change:
✅Quarter +10% to -10%
✅Week Down to Week up

This is my bearish scan for swing setups. 👇

Important paper from Google on large batch optimization. They do impressively careful experiments measuring # iterations needed to achieve target validation error at various batch sizes. The main "surprise" is the lack of surprises. [thread]

https://t.co/7QIx5CFdfJ

The paper is a good example of lots of elements of good experimental design. They validate their metric by showing lots of variants give consistent results. They tune hyperparamters separately for each condition, check that optimum isn't at the endpoints, and measure sensitivity.

They have separate experiments where the hold fixed # iterations and # epochs, which (as they explain) measure very different things. They avoid confounds, such as batch norm's artificial dependence between batch size and regularization strength.

When the experiments are done carefully enough, the results are remarkably consistent between different datasets and architectures. Qualitatively, MNIST behaves just like ImageNet.

Importantly, they don't find any evidence for a "sharp/flat optima" effect whereby better optimization leads to worse final results. They have a good discussion of experimental artifacts/confounds in past papers where such effects were reported.