Mohamed Mekkouri

Hello @lucidrains Sorry if i'm mistaken, but in the line 380 of the file denoising_diffusion_pytorch.py, where the final ResNetBlock is defined for the UNet : `self.final_res_block = block_klass(dim * 2,...

RuntimeError: CUDA error: an illegal instruction was encountered when runing test.py

2

Hello, When running `python test.py` I get the error : ===================================== ERROR: test_groups (__main__.Test) ---------------------------------------------------------------------- Traceback (most recent call last): File "/fsx/mohamed/dev/marlin/test.py", line 155, in test_groups self.run_problem(m, n, k, *thread_shape,...

Hello everyone ! I am trying to understand how the marlin kernel works in depth to adapt it for int2 quantization, do you have any pointers please ? I appreciate...

Changing __repr__ in torchao to show quantized Linear

1

# What does this PR do? When a model is quantized using TorchAO and then loaded, the representation of its Linear layers is expected to be different compared to the...

As the title implies, I added the Blackwell architecture to the list of supported architectures. Since it was already mentioned at the beginning, I simply included it in the list.

Add Finegrained FP8

3

# What does this PR do? Adds finegrained FP8

Hey Team, In our tests in transformers we were expecting the layer type to be `tritonv2` for T4 gpus, but after the latest release it's `torch`. Any ideas why ?...

[core] fix fp-quant

4

# What does this PR do? Fix fp-quant