diff options
author | Keiichi Tabata <keiichi.tabata@outlook.com> | 2023-08-06 15:34:05 +0900 |
---|---|---|
committer | GitHub <noreply@github.com> | 2023-08-06 09:34:05 +0300 |
commit | 2e8265ae1764d6288aab0e2df641909072e2d58e (patch) | |
tree | 339cb1a562357c7682a49e3b3841807bd76b5fc6 | |
parent | f514d1b306e1114c2884fcb25dd9bd48ae64ba32 (diff) |
convert.py : add missing abstract methods for quantized data (#2491)
-rw-r--r-- | convert.py | 7 |
1 files changed, 7 insertions, 0 deletions
@@ -465,6 +465,13 @@ class GGMLQuantizedTensor(Tensor): def permute(self, n_head: int, n_kv_head: Optional[int] = None) -> 'GGMLQuantizedTensor': return GGMLQuantizedTensor(permute(self.ndarray, n_head, n_kv_head), self.shape, self.data_type) + def permute_part(self, n_part: int, n_head: int) -> 'UnquantizedTensor': + r = self.ndarray.shape[0] // 3 + return UnquantizedTensor(permute(self.ndarray[r * n_part : r * n_part + r, ...], n_head)) + + def part(self, n_part: int) -> 'UnquantizedTensor': + r = self.ndarray.shape[0] // 3 + return UnquantizedTensor(self.ndarray[r * n_part : r * n_part + r, ...]) GGMLCompatibleTensor = Union[UnquantizedTensor, GGMLQuantizedTensor] |