manojpreveen commited on
Commit
da08437
·
1 Parent(s): cf5b212

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +5 -0
README.md CHANGED
@@ -10,6 +10,11 @@ First Version of Instruction Tuned Bloomz-7B1 model on ChatGPT dataset (85k data
10
  **Training Details :**
11
  * Epochs: 5
12
  * Batch Size : 5 instantaneous per device x 2 gradient accumulation steps x 8 gpus = 80
 
 
 
 
 
13
  * Machine : 8xA100 80GB
14
 
15
  **Dataset Details :**
 
10
  **Training Details :**
11
  * Epochs: 5
12
  * Batch Size : 5 instantaneous per device x 2 gradient accumulation steps x 8 gpus = 80
13
+ * Max Length : 512
14
+ * Weight Decay : 0
15
+ * Learning Rate : 5e-5
16
+ * Learning Rate Scheduler Type : Linear
17
+ * Number of warmup steps : 0
18
  * Machine : 8xA100 80GB
19
 
20
  **Dataset Details :**