Update README.md
Browse files
README.md
CHANGED
|
@@ -36,6 +36,16 @@ developed by Università di Bari. For the implementation we combined different a
|
|
| 36 |
- **Developed by:** [giux78](https://alessandroercolani.webflow.io/)
|
| 37 |
- **Funded by:** [Business Operating System](https://www.businessos.xyz)
|
| 38 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 39 |
## Evaluations:
|
| 40 |
|
| 41 |
| Model | Arc-c | HellaS | MMUL | AVG |
|
|
|
|
| 36 |
- **Developed by:** [giux78](https://alessandroercolani.webflow.io/)
|
| 37 |
- **Funded by:** [Business Operating System](https://www.businessos.xyz)
|
| 38 |
|
| 39 |
+
## Code
|
| 40 |
+
|
| 41 |
+
I followed the [alingnment handbook](https://github.com/huggingface/alignment-handbook/blob/main/recipes/zephyr-7b-beta/sft/config_qlora.yaml) from HuggingfaceH4 team
|
| 42 |
+
, I just changed the base model and some parameters
|
| 43 |
+
|
| 44 |
+
## Computation
|
| 45 |
+
It has been trained on two A100 from [seeweb.it](https://www.seeweb.it/) who sponsered the training. I strongly
|
| 46 |
+
suggest is one of the cheapest and solid GPUs provider.
|
| 47 |
+
|
| 48 |
+
|
| 49 |
## Evaluations:
|
| 50 |
|
| 51 |
| Model | Arc-c | HellaS | MMUL | AVG |
|