virilo

@shakthimaan @imawhale did you solve it? the carousel would be a great enhancement

I'm interested too. My email is [email protected] Thanks in advance!

I tried to force this behavior by introducing a `\n` or a `` but neither of them worked.

Test code: (you can see it in [Google Colab](https://colab.research.google.com/drive/11DEJ0De30n0ZrdRMOzxPkzdFULWvuee9#scrollTo=Zw75gPQ3aKXk)) # -*- coding: utf-8 -*- !pip uninstall lucid --yes !pip install git+https://github.com/virilo/lucid.git import lucid lucid.__version__ !git clone https://github.com/virilo/lucid-issue !find "lucid-issue/keras-2.2.4/" -type...

I had the same issue. It is a ZIP64. You can uncompress it with java 64 bits version: jar xvf world-cleanup-app-images_20190306.zip > uncompress.log

I solved this issue by using model._blocks[-1]._project_conv as layer parameter, instead of None. But [this solution](https://github.com/sidml/EfficientNet-GradCam-Visualization/issues/2#issuecomment-608279845) is better. Although both gives me the same results. BTW @HelloAlone @shravan2x **@ all**...

Thanks a lot @kermitt2 @lfoppiano ! I've seen coords attributes, but they are only present in `` tags (for my grobid extractions). Is there any way to get them also...

Thanks a lot @kermitt2 @lfoppiano! Just sharing some early experiments using GROBID to extract chunks for a RAG pipeline, splitting the document by sections and subsections. Overall, the results are...

I'm sharing a patch to overcome the tie_weights issue. This patch allows you to instantiate the model object: ``` import os from vllm import LLM, SamplingParams AVAILABLE_GPUS = [1,2] #...

@pluiez I tested it by adding an assertion: ``` _original_init = DeepseekV2MoE.__init__ def patched_init(self, config, quant_config=None, prefix=""): assert False, "patched DeepseekV2MoE(...) was called!" ... ``` It was executed and stopped...