HF-Quantization

/

Llama-3.2-1B-TORCHAO-W8A8

like 0

Follow

HF Quantization 2

• "torch.int8",
• "torch.bfloat16",
• "torch.CharStorage",
• "collections.OrderedDict",
• "torch.BFloat16Storage",
• "torchao.quantization.quant_primitives.ZeroPointDomain",
• "torch.device",
• "torch._utils._rebuild_tensor_v2",
• "torch.LongStorage",
• "torchao.quantization.quant_api._int8_symm_per_token_reduced_range_quant",
• "torchao.dtypes.uintx.plain_layout.PlainAQTTensorImpl",
• "torch._tensor._rebuild_from_type_v2",
• "torchao.quantization.linear_activation_quantized_tensor.LinearActivationQuantizedTensor",
• "torch.serialization._get_layout",
• "torchao.dtypes.utils.PlainLayout",
• "torchao.dtypes.affine_quantized_tensor.AffineQuantizedTensor",
• "torch._utils._rebuild_wrapper_subclass"

How to fix it?