Update README.md
Browse files
README.md
CHANGED
|
@@ -6,16 +6,13 @@ datasets:
|
|
| 6 |
- skvarre/sv-instruct-v1
|
| 7 |
---
|
| 8 |
|
| 9 |
-
# Model Card for Model ID
|
| 10 |
-
|
| 11 |
-
<!-- Provide a quick summary of what the model is/does. -->
|
| 12 |
-
|
| 13 |
-
|
| 14 |
-
|
| 15 |
## Model Details
|
| 16 |
|
| 17 |
Finetune of [gpt-sw3-6.7b-v2](https://huggingface.co/AI-Sweden-Models/gpt-sw3-6.7b-v2) using **LoRA** with 4-bit quantization. Adapters merged with base model, however with ```bfloat16``` tensors.
|
| 18 |
|
|
|
|
|
|
|
|
|
|
| 19 |
|
| 20 |
## Evaluation
|
| 21 |
ScandEval benchmarks:
|
|
@@ -29,3 +26,5 @@ ScandEval benchmarks:
|
|
| 29 |
| mmlu-sv | 5.45 ± 0.91 / 28.14 ± 0.82 |
|
| 30 |
| hellaswag-sv | 27.95 ± 0.73 / 4.19 ± 0.94 |
|
| 31 |
| speed | 5322.20 ± 1132.75 / 1280.06 ± 408.08 |
|
|
|
|
|
|
|
|
|
| 6 |
- skvarre/sv-instruct-v1
|
| 7 |
---
|
| 8 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 9 |
## Model Details
|
| 10 |
|
| 11 |
Finetune of [gpt-sw3-6.7b-v2](https://huggingface.co/AI-Sweden-Models/gpt-sw3-6.7b-v2) using **LoRA** with 4-bit quantization. Adapters merged with base model, however with ```bfloat16``` tensors.
|
| 12 |
|
| 13 |
+
## Usage:
|
| 14 |
+
|
| 15 |
+
This is a finetune experiment. How to use will be provided later.
|
| 16 |
|
| 17 |
## Evaluation
|
| 18 |
ScandEval benchmarks:
|
|
|
|
| 26 |
| mmlu-sv | 5.45 ± 0.91 / 28.14 ± 0.82 |
|
| 27 |
| hellaswag-sv | 27.95 ± 0.73 / 4.19 ± 0.94 |
|
| 28 |
| speed | 5322.20 ± 1132.75 / 1280.06 ± 408.08 |
|
| 29 |
+
|
| 30 |
+
|