Pinned Loading

1. visual-token-pruningvisual-token-pruningPublic

   evaluating signals for pruning visual tokens in VLMs

   Python

2. ScalingIntelligence/KernelBenchScalingIntelligence/KernelBenchPublic

   KernelBench: Can LLMs Write GPU Kernels? - Benchmark + Toolkit with Torch -> CUDA (+ more DSLs)

   Jupyter Notebook 1.2k 192

3. vlm-normsvlm-normsPublic

   investigating norm discrepancies in VLMs

   Python 1

4. vrlm-trajectoriesvrlm-trajectoriesPublic

   dataset for creating RLM trajectories on video understanding tasks

   Python

5. forgeforgePublic

   single gpu machine job queuing cli for background work (i am so sick of clearml)

   Python