Hardik Shah

activity

20222026

most citedVoLTA: Vision-Language Transformer with Weakly-Supervised Local-Feature Alignment

11 citations · 14 across the 8 of their papers we have counts for

collaborators

Showing cs.LGShow all

1 paper · 1 filter

cs.LG2023

End-to-End Neural Network Compression via $\frac{ℓ _{1}}{ℓ _{2}}$ Regularized Latency Surrogates

Anshul Nasery, Hardik Shah, Arun Sai Suggala +1

Neural network (NN) compression via techniques such as pruning, quantization requires setting compression hyperparameters (e.g., number of channels to be pruned, bitwidths for quan…

End-to-End Neural Network Compression via ℓ2​ℓ1​​ Regularized Latency Surrogates

End-to-End Neural Network Compression via $\frac{ℓ _{1}}{ℓ _{2}}$ Regularized Latency Surrogates