Commit
•
273ff7d
1
Parent(s):
22089c5
Update README.md
Browse files
README.md
CHANGED
|
@@ -17,4 +17,17 @@ Mistral-7B-v0.1 is a transformer model, with the following architecture choices:
|
|
| 17 |
Byte-fallback BPE tokenizer
|
| 18 |
|
| 19 |
|
|
|
|
| 20 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 17 |
Byte-fallback BPE tokenizer
|
| 18 |
|
| 19 |
|
| 20 |
+
Troubleshooting
|
| 21 |
|
| 22 |
+
If you see the following error:
|
| 23 |
+
|
| 24 |
+
KeyError: 'mistral'
|
| 25 |
+
|
| 26 |
+
Or:
|
| 27 |
+
|
| 28 |
+
NotImplementedError: Cannot copy out of meta tensor; no data!
|
| 29 |
+
|
| 30 |
+
Ensure you are utilizing a stable version of Transformers, 4.34.0 or newer.
|
| 31 |
+
Notice
|
| 32 |
+
|
| 33 |
+
Mistral 7B is a pretrained base model and therefore does not have any moderation mechanisms.
|