Omdano
/

INT8-H16P

Model card Files Files and versions

848 MB

Ctrl+K

Ctrl+K

1 contributor

History: 8 commits

Omdano's picture

Re-quantize with float16 for T4 GPU compatibility

c68d969 verified 6 months ago

.gitattributes

1.52 kB
initial commit 6 months ago
README.md

1.21 kB
Upload README.md with huggingface_hub 6 months ago
config.json

744 Bytes
Re-quantize with float16 for T4 GPU compatibility 6 months ago
preprocessor_config.json

585 Bytes
Upload folder using huggingface_hub 6 months ago
pytorch_model.bin
Detected Pickle imports (15)
- "torch.int8",
- "torchao.dtypes.uintx.plain_layout.PlainAQTTensorImpl",
- "torch._tensor._rebuild_from_type_v2",
- "torch.float16",
- "collections.OrderedDict",
- "torch.device",
- "torch.LongStorage",
- "torch._utils._rebuild_wrapper_subclass",
- "torchao.quantization.quant_primitives.ZeroPointDomain",
- "torchao.dtypes.utils.PlainLayout",
- "torch.serialization._get_layout",
- "torch.HalfStorage",
- "torch._utils._rebuild_tensor_v2",
- "torchao.dtypes.affine_quantized_tensor.AffineQuantizedTensor",
- "torch.CharStorage"
How to fix it?
848 MB
xet

Re-quantize with float16 for T4 GPU compatibility 6 months ago