ASHISH TENDULKAR: So how many of you are doing research?
Yay.
A large amount of you.
So you know the importance of this, right?
If you can get your experiment complete
in minutes or hours, great.
You can iterate quite fast.
One to four days is tolerable.
Right?
You can try to run many experiments in parallel
and try to reduce your time.
One to four weeks.
OK.
Only if it's high value experiment, progress stalls.
More than one m...
| repo | techs | stars | weekly | forks | weekly |
|---|---|---|---|---|---|
| roboflow/supervision | Python | 5.8k | +369 | 398 | +31 |
| CorentinJ/Real-Time-Voice-Cloning | Python | 48k | +278 | 8.2k | +24 |
| microsoft/onnxruntime | C++CPython | 10.6k | +143 | 2.3k | +16 |
| ray-project/ray | PythonC++Java | 28.1k | +113 | 4.8k | +13 |
| d2l-ai/d2l-en | PythonTeXHTML | 19.3k | +85 | 3.9k | +7 |
| onnx/onnx | PythonC++CMake | 15.7k | +48 | 3.6k | +4 |
| google-research/bert | PythonJupyter Notebook | 35.4k | +46 | 9.4k | +8 |
| mozilla/DeepSpeech | C++PythonC | 23.1k | +36 | 3.9k | +1 |
| datawhalechina/fun-rec | Jupyter NotebookPythonVue | 2.7k | +30 | 638 | +4 |
| satellite-image-deep-learning/techniques | 6.9k | +30 | 1.2k | +5 |