apps/FastDeploy
1
0
Fork 0
mirror of https://github.com/PaddlePaddle/FastDeploy.git synced 2026-04-23 17:11:21 +08:00
Code Issues Actions 10 Packages Projects Releases Wiki Activity

[Other] Adjust GPUModelRunner to enhance compatibility (#6851)

Browse Source
This commit is contained in:
ming1753
2026-03-16 14:49:19 +08:00
committed by GitHub
parent 04fde3b227
commit bb925c605f
25 changed files with 224 additions and 248 deletions
Download Patch File Download Diff File Expand all files Collapse all files
fastdeploy/model_executor/models/paddleocr_vl/paddleocr_vl.py
+4 -3
View File
@@ -221,7 +221,7 @@ class PaddleOCRVLForConditionalGeneration(ModelForCasualLM):
def name(self):
return "PaddleOCRVLForConditionalGeneration"
def compute_logits(self, hidden_states: paddle.Tensor):
def compute_logits(self, hidden_states: paddle.Tensor, forward_meta: ForwardMeta = None):
logits = self.lm_head(hidden_states)
logits = paddle.cast(logits, paddle.float32)
logits[:, self.vocab_size :] = -float("inf")
@@ -246,10 +246,11 @@ class PaddleOCRVLForConditionalGeneration(ModelForCasualLM):
def forward(
self,
ids_remove_padding: paddle.Tensor,
image_features: Optional[paddle.Tensor],
inputs: Dict,
forward_meta: ForwardMeta,
):
ids_remove_padding = inputs["ids_remove_padding"]
image_features = inputs["image_features"]
input_embeddings = self.get_input_embeddings(
ids_remove_padding=ids_remove_padding, image_features=image_features, forward_meta=forward_meta
)