Update README.md
Browse files
README.md
CHANGED
|
@@ -75,6 +75,16 @@ Total 153,013 samples.
|
|
| 75 |
| Average | |
|
| 76 |
|
| 77 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 78 |
## Training Arguments
|
| 79 |
| | |
|
| 80 |
|------ | ------ |
|
|
@@ -97,7 +107,6 @@ Total 153,013 samples.
|
|
| 97 |
| grandient_accumulation_steps | 16 |
|
| 98 |
| bf16 | True |
|
| 99 |
|
| 100 |
-
A40-48G x 2
|
| 101 |
|
| 102 |
| | |
|
| 103 |
|------ | ------ |
|
|
|
|
| 75 |
| Average | |
|
| 76 |
|
| 77 |
|
| 78 |
+
H800-80G x 2
|
| 79 |
+
|
| 80 |
+
transformers=4.33.0
|
| 81 |
+
|
| 82 |
+
flash-attn=2.1.0
|
| 83 |
+
|
| 84 |
+
bitsandbytes=0.41.1
|
| 85 |
+
|
| 86 |
+
peft=0.5.0
|
| 87 |
+
|
| 88 |
## Training Arguments
|
| 89 |
| | |
|
| 90 |
|------ | ------ |
|
|
|
|
| 107 |
| grandient_accumulation_steps | 16 |
|
| 108 |
| bf16 | True |
|
| 109 |
|
|
|
|
| 110 |
|
| 111 |
| | |
|
| 112 |
|------ | ------ |
|