feat: support INT8 ConvRot safetensors (#1857)

This commit is contained in:
leejet
2026-08-11 23:45:02 +08:00
committed by GitHub
parent 487de75c96
commit bcc7e29568
9 changed files with 384 additions and 25 deletions
+3
View File
@@ -1508,6 +1508,9 @@ bool ModelLoader::load_tensors(std::map<std::string, ggml_tensor*>& tensors,
bool ModelLoader::tensor_should_be_converted(const TensorStorage& tensor_storage, ggml_type type) {
const std::string& name = tensor_storage.name;
if (tensor_storage.is_int8_tensorwise) {
return false;
}
if (type != GGML_TYPE_COUNT) {
if (ggml_is_quantized(type) && tensor_storage.ne[0] % ggml_blck_size(type) != 0) {
// Pass, do not convert