| # Copyright 2024-2025 ModelCloud.ai | |
| # Copyright 2024-2025 qubitium@modelcloud.ai | |
| # Contact: qubitium@modelcloud.ai, x.com/qubitium | |
| # | |
| # Licensed under the Apache License, Version 2.0 (the "License"); | |
| # you may not use this file except in compliance with the License. | |
| # You may obtain a copy of the License at | |
| # | |
| # http://www.apache.org/licenses/LICENSE-2.0 | |
| # | |
| # Unless required by applicable law or agreed to in writing, software | |
| # distributed under the License is distributed on an "AS IS" BASIS, | |
| # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. | |
| # See the License for the specific language governing permissions and | |
| # limitations under the License. | |
| import math | |
| import numpy as np | |
| import torch | |
| # helper method to get accurate parameter count for a gptq model | |
| def tensor_parameters( | |
| tensor_name: str, # name of tensor weight in model | |
| tensor_shape: torch.Size, # shape of tensor | |
| bits: int, # gptq bits | |
| ) -> int: | |
| # only .qweight is relevant for `parameters` in gptq model | |
| if tensor_name.endswith(".qweight"): | |
| real_infeatures = math.ceil(tensor_shape[0] / bits * 32) | |
| real_tensor_shape = (real_infeatures,) + tensor_shape[1:] | |
| return np.prod(real_tensor_shape) | |
| # .scales and .qzeros are not model parameters but aux data for .qweight | |
| elif tensor_name.endswith((".scales", ".qzeros", ".g_idx")): | |
| return 0 | |
| else: | |
| return np.prod(tensor_shape) | |