Route forward

models/quantization.py:None–None · view source on GitHub ↗

custom_fwd(cast_inputs=torch.float16)

Source from the content-addressed store, hash-verified

1	import numpy as np
2	import torch
3	import torch.nn as nn
4	from torch.cuda.amp import custom_bwd, custom_fwd

nothing calls this directly

no outgoing calls

no test coverage detected