Nicolas Iglesias
commited on
Commit
•
d2d5712
1
Parent(s):
ffc0e1d
Update README.md
Browse files
README.md
CHANGED
@@ -16,6 +16,11 @@ Zenos GPT-J 6B Alpaca Evol 4-bit is a Spanish Instruction capable model based on
|
|
16 |
|
17 |
### Requirements
|
18 |
|
|
|
|
|
|
|
|
|
|
|
19 |
Since this is a compressed version (4 bits), it can fit into ~7GB of VRAM.
|
20 |
|
21 |
## Usage
|
@@ -68,7 +73,7 @@ print(answer)
|
|
68 |
|
69 |
# Inference
|
70 |
|
71 |
-
Currently, the HuggingFace's Inference Tool UI doesn't properly load the model. However, you can use it with regular Python code as shown above.
|
72 |
|
73 |
# Acknowledgments
|
74 |
|
|
|
16 |
|
17 |
### Requirements
|
18 |
|
19 |
+
The following **specific** versions are required in order to load and/or manipulate the present model. At least, until the existing PRs are approved,
|
20 |
+
|
21 |
+
- [bitsandbytes](https://github.com/poedator/bitsandbytes/tree/save4)
|
22 |
+
- [transformers](https://github.com/poedator/transformers/tree/save4)
|
23 |
+
|
24 |
Since this is a compressed version (4 bits), it can fit into ~7GB of VRAM.
|
25 |
|
26 |
## Usage
|
|
|
73 |
|
74 |
# Inference
|
75 |
|
76 |
+
Currently, the HuggingFace's Inference Tool UI doesn't properly load the model. However, you can use it with regular Python code as shown above once you meet the [requirements](#requirements).
|
77 |
|
78 |
# Acknowledgments
|
79 |
|