update int8 onnx model and readme

Signed-off-by: dujun <[email protected]>

README.md

@@ -13,7 +13,7 @@ metrics:
 - f1
 ---
 
-# INT8 roberta base finetuned MRPC
+# INT8 xlm-roberta base finetuned MRPC
 
 ## Post-training dynamic quantization
 
@@ -27,8 +27,8 @@ The original fp32 model comes from the fine-tuned model [Intel/xlm-roberta-base-
 
 |   |INT8|FP32|
 |---|:---:|:---:|
-| **Accuracy (eval-f1)** |0.8945|0.9010|
-| **Model size (MB)**  |275|1061|
+| **Accuracy (eval-f1)** |0.8966|0.9010|
+| **Model size (MB)**  |354|1061|
 
 
 #### Load ONNX model:

model.onnx

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f6d813dd10c5fe77b93e9f93b50fe31b73f84c18cbb223b3ee126d728dadb4ea
-size 287513036
+oid sha256:b64eb841911f82bf28d8f832429f0dad30731cb47eccda96944af48d84f288d9
+size 370569392