HF-Quantization

/

Llama-3.2-1B-TORCHAO-W4

like 0

Follow

HF Quantization 1

• "torch.serialization._get_layout",
• "torchao.dtypes.affine_quantized_tensor.AffineQuantizedTensor",
• "torch.bfloat16",
• "torch.int32",
• "torchao.dtypes.uintx.tensor_core_tiled_layout.TensorCoreTiledLayout",
• "torch.BFloat16Storage",
• "torch.IntStorage",
• "torch._utils._rebuild_wrapper_subclass",
• "collections.OrderedDict",
• "torchao.dtypes.uintx.tensor_core_tiled_layout.TensorCoreTiledAQTTensorImpl",
• "torchao.quantization.quant_primitives.ZeroPointDomain",
• "torch.device",
• "torch._utils._rebuild_tensor_v2",
• "torch._tensor._rebuild_from_type_v2"

How to fix it?