Update README.md
Browse files
README.md
CHANGED
@@ -86,8 +86,8 @@ The model was trained using LoRA applied to Q and V matrices.
|
|
86 |
| Hyperparameter | Value |
|
87 |
|-------------------------|---------------------------|
|
88 |
| Max Sequence Length | 600 |
|
89 |
-
| Batch Size |
|
90 |
-
| Number of Epochs |
|
91 |
| Learning Rate | 5e-4 |
|
92 |
| Weight Decay | 0.01 |
|
93 |
| Logging Steps | 100 |
|
@@ -107,7 +107,7 @@ The model was trained using LoRA applied to Q and V matrices.
|
|
107 |
|
108 |
## Evaluation
|
109 |
|
110 |
-
ThreatDetect-C-Cpp reaches an accruacy of
|
111 |
|
112 |
|
113 |
|
|
|
86 |
| Hyperparameter | Value |
|
87 |
|-------------------------|---------------------------|
|
88 |
| Max Sequence Length | 600 |
|
89 |
+
| Batch Size | 32 |
|
90 |
+
| Number of Epochs | 9 |
|
91 |
| Learning Rate | 5e-4 |
|
92 |
| Weight Decay | 0.01 |
|
93 |
| Logging Steps | 100 |
|
|
|
107 |
|
108 |
## Evaluation
|
109 |
|
110 |
+
ThreatDetect-C-Cpp reaches an accruacy of 86% on the eval set.
|
111 |
|
112 |
|
113 |
|