CefSharp VS pytorch-lightning

The Fullscreen Chromium (cefsharp) based Web-Browser ignores optionally user input like Mouse clicks, and is optionally always in the foreground (Force-Top-Most), so it can be used for any situation.

Just as a background: I'm working on a façade for the CefSharp utility. This utility requires the programmer to initialize and terminate it in the same thread, once-per-process. My current implementation uses a lazy initialization algorithm requiring inter-thread communication. But the possibility of AppDomains arising or passing away complicates the algorithm by requiring threads to potentially signal one another across AppDomain boundaries. It would be easier to just let the threads communicate without worrying about which AppDomain they belonged to.

I dont want to state the obvious but of course there is CefSharp .. But of course thats pretty heavy duty as you end shipping a headless version of Chrome in your app, but it can do anything a browser can do (of course it can, its Chrome), including exectuting a JavaScript string on the fly and getting a response. Because its Chrome you evenn have access the the dev tool and everything like that. But the other suggestions might be better suited as they might be more lightweight.

Two options: Either use an embedded browser you can control programmatically like CefSharp (https://cefsharp.github.io/) or spawn a real browser and use sendkeys() to control the browser by emulating user input.

Chrome Runtime discussion in CefSharp repository.

https://github.com/cefsharp/CefSharp/ if i spiked your intrigue

CefSharp

I always use https://cefsharp.github.io/

def on_validation_end(self, trainer: Trainer, pl_module: LightningModule) -> None: # Trainer calls `on_validation_end` for sanity check. Therefore, it is necessary to avoid # calling `trial.report` multiple times at epoch 0. For more details, see # https://github.com/PyTorchLightning/pytorch-lightning/issues/1391. if trainer.sanity_checking: return

Under the hood, the ModelWrapper object will create a ML model based on the config (so far, an XGBoost model and a PyTorch Lightning model). Each of those will have a wrapper that conducts training and evaluation (since from my understanding of Lightning, Trainers are required to be outside of the class). In lack of a better name, I call these wrappers Fitters. For uniformity, I thought about adding a common interface IFitter, which is inherited by all model wrappers as outlined below.

Join their Slack to ask the community questions and check out the GitHub here.

Pytorch lightning benchmarks against pytorch on every PR (benchmarks to make sure that it is mot slower.

There is already a ton of momentum around automating ML workflows. I would suggest you contribute to a preexisting project like, for instance, PyTorch Lightning or fast.ai.

Is the problem the sheer number of options, or the fact that they are all together in one place? Would it be better if they were organized into the different trainer entrypoints (fit, validate, ...)? If that is the case, there was an RFC proposing this which you might find interesting, feel free to drop by and comment on the issue: https://github.com/PyTorchLightning/pytorch-lightning/issues/10444

I wanted to make a quick performance comparison between the GPU (Tesla K80) and TPU (v2-8) available in Google Colab with PyTorch. To do so quickly, I used an MNIST example from pytorch-lightning that trains a simple CNN.

We've noticed GPU 0 on our 3 GPU system is sometimes idle (which would explain performance differences). However its unclear to us why that may be. Similar to this issue

If you want a good example for reference, take a look at Pytorch Lightning's readme (https://github.com/PyTorchLightning/pytorch-lightning) It answers the 3 questions of "what is this", "why should I care", and "how do i use it" almost instantly