search

neuralmagic / sparseml

Libraries for applying sparsification recipes to neural networks with a few lines of code, enabling faster and smaller models
Apache License 2.0
2.05k stars 144 forks source link

[Cherry-Pick] Make reloading compatible with safetensors #2291

Closed dbogunowicz closed 4 months ago

dbogunowicz commented 4 months ago

As confirmed by @shubhra, this commit needs to be cherry-picked into 1.7 to work with the quantized model, whenever the quantization type specified in the recipe is channel-wise.