1pikachu1111 commited on
Commit
1dc7724
1 Parent(s): 4843e13

update int8 onnx model and readme

Browse files

Signed-off-by: dujun <jun.du@intel.com>

Files changed (2) hide show
  1. README.md +4 -2
  2. model.onnx +2 -2
README.md CHANGED
@@ -61,12 +61,14 @@ This is an INT8 ONNX model quantized with [Intel® Neural Compressor](https://gi
61
 
62
  The original fp32 model comes from the fine-tuned model [roberta-base-mrpc](https://huggingface.co/Intel/roberta-base-mrpc).
63
 
 
 
64
  #### Test result
65
 
66
  | |INT8|FP32|
67
  |---|:---:|:---:|
68
- | **Accuracy (eval-f1)** |0.9073|0.9138|
69
- | **Model size (MB)** |243|476|
70
 
71
 
72
  #### Load ONNX model:
61
 
62
  The original fp32 model comes from the fine-tuned model [roberta-base-mrpc](https://huggingface.co/Intel/roberta-base-mrpc).
63
 
64
+ The calibration dataloader is the eval dataloader. The calibration sampling size is 100.
65
+
66
  #### Test result
67
 
68
  | |INT8|FP32|
69
  |---|:---:|:---:|
70
+ | **Accuracy (eval-f1)** |0.9100|0.9138|
71
+ | **Model size (MB)** |294|476|
72
 
73
 
74
  #### Load ONNX model:
model.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:51de3a4577a50af94e72dd8a75d77b9145726d3f83de336a2573fd12180c6075
3
- size 254669454
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:315f3dfad2e4344cfc4688634a4909c0505467bb3cec620509ad204af3662cea
3
+ size 307815333