Yoni Kremer
Yoni Kremer
When I'm training a model using keras, I always calculate the runtime by hand (Using the estimated time per epoch) It would be great if when I use mlnotify the...
The error: Collecting flopco-keras Downloading flopco-keras-0.1.2.tar.gz (3.4 kB) Preparing metadata (setup.py) ... error error: subprocess-exited-with-error × python setup.py egg_info did not run successfully. │ exit code: 1 ╰─> [10 lines...
PyCharm gave me a lot of warnings about the current main branch so I fixed many of those (not all). The vast majority of fixes are auto-fixes. The inference works...
The correct answer to PCA/week 2/General inner products: lengths and distances/question 1 is sqrt(26) and nor sqrt(29)
(shapes (2, 1), (3,) mismatch) x: array([[0.666667], [1.333333]]) y: array([0.5, 1.5, 1. ]) def mean(X): D, N = X.shape mean = np.zeros((D,1)) mean = np.mean(X, axis=1, keepdims=True) return mean
I am working on the growing datasets lab. When I am trying to load to embedding for the train set, I am loading 32% of the data very fast (33...
### Description In numpy, you can use `==` and `!=` to check which item in an `ndarray` is None. In cupy, comparing an array to None raises `NotImplemented`. Numpy example:...
### Description When running the tests on my 4GB GPU (NVIDIA GeForce RTX 3050) I got a cuda out of memory. I want to add a pull request to lower...
Preformence: change TopKMaskLogitsKernel to oveewrite the tensor instead of allocating a new tensor
Hi, I am tying to scale up to large batch sized and noticed that TopKMaskLogitsKernel is allocating a new tensor instead of modified the tensor in-place. In the vast majority...
Hey, I am using a 16 bit floating point model that outputs a 16 bit floating point logits. I noticed that when in `flashinfer.sampling.top_p_sampling_from_probs` the logits are converted into a...