Commit
•
780c487
1
Parent(s):
a6ff6c2
Update README.md
Browse files
README.md
CHANGED
|
@@ -8,7 +8,7 @@ inference: false
|
|
| 8 |
|
| 9 |
|
| 10 |
|
| 11 |
-
# Gecko-7B-v0.1
|
| 12 |
Designed to generate instructive and narrative text, with a specific focus on mathematics & numeracy.
|
| 13 |
|
| 14 |
Full-parameter fine-tune (FFT) of Mistral-7B-Instruct-v0.2, with apache-2.0 license.
|
|
@@ -39,16 +39,6 @@ The user is responsible for their output and must use this model responsibly.
|
|
| 39 |
|
| 40 |
### Hardware and Training
|
| 41 |
|
| 42 |
-
|
| 43 |
-
|
| 44 |
-
```
|
| 45 |
-
|
| 46 |
-
n_epochs = 3,
|
| 47 |
-
n_checkpoints = 3,
|
| 48 |
-
batch_size = 12,
|
| 49 |
-
learning_rate = 1e-5,
|
| 50 |
-
|
| 51 |
-
|
| 52 |
-
```
|
| 53 |
|
| 54 |
*Sincere appreciation to Techmind for their generous sponsorship.*
|
|
|
|
| 8 |
|
| 9 |
|
| 10 |
|
| 11 |
+
# Gecko-7B-v0.1-DPO
|
| 12 |
Designed to generate instructive and narrative text, with a specific focus on mathematics & numeracy.
|
| 13 |
|
| 14 |
Full-parameter fine-tune (FFT) of Mistral-7B-Instruct-v0.2, with apache-2.0 license.
|
|
|
|
| 39 |
|
| 40 |
### Hardware and Training
|
| 41 |
|
| 42 |
+
Trained on A100 for 2 hours trained using Axolotl
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 43 |
|
| 44 |
*Sincere appreciation to Techmind for their generous sponsorship.*
|