jerryzh168

/

Qwen3-4B-FP8

like 0

• "torch.BFloat16Storage",
• "torchao.quantization.quantize_.workflows.float8.float8_tensor.QuantizeTensorToFloat8Kwargs",
• "torch.FloatStorage",
• "torch._utils._rebuild_wrapper_subclass",
• "torch.device",
• "torch._utils._rebuild_tensor_v2",
• "torch.bfloat16",
• "torch._utils._rebuild_tensor_v3",
• "collections.OrderedDict",
• "torchao.quantization.granularity.PerRow",
• "torch.float8_e4m3fn",
• "torchao.quantization.Float8Tensor",
• "torch.storage.UntypedStorage",
• "torchao.quantization.quantize_.common.kernel_preference.KernelPreference",
• "torch._tensor._rebuild_from_type_v2",
• "torchao.float8.inference.Float8MMConfig",
• "torch.serialization._get_layout"

How to fix it?