Commit
•
a8c03f7
1
Parent(s):
7029f7b
Update README.md
Browse files
README.md
CHANGED
|
@@ -19,21 +19,20 @@ model-index:
|
|
| 19 |
num_few_shot: 25
|
| 20 |
metrics:
|
| 21 |
- type: acc
|
| 22 |
-
value:
|
| 23 |
name: accuracy
|
| 24 |
- task:
|
| 25 |
type: text-generation
|
| 26 |
name: Text Generation
|
| 27 |
dataset:
|
| 28 |
-
name:
|
| 29 |
-
type:
|
| 30 |
-
|
| 31 |
-
split: test
|
| 32 |
args:
|
| 33 |
-
num_few_shot:
|
| 34 |
metrics:
|
| 35 |
- type: acc
|
| 36 |
-
value:
|
| 37 |
name: accuracy
|
| 38 |
- task:
|
| 39 |
type: text-generation
|
|
@@ -48,7 +47,7 @@ model-index:
|
|
| 48 |
metrics:
|
| 49 |
- type: acc
|
| 50 |
name: accuracy
|
| 51 |
-
value:
|
| 52 |
- task:
|
| 53 |
type: text-generation
|
| 54 |
name: Text Generation
|
|
@@ -61,7 +60,7 @@ model-index:
|
|
| 61 |
num_few_shot: 5
|
| 62 |
metrics:
|
| 63 |
- type: acc
|
| 64 |
-
value:
|
| 65 |
name: accuracy
|
| 66 |
- task:
|
| 67 |
type: text-generation
|
|
@@ -75,7 +74,7 @@ model-index:
|
|
| 75 |
num_few_shot: 5
|
| 76 |
metrics:
|
| 77 |
- type: acc
|
| 78 |
-
value:
|
| 79 |
name: accuracy
|
| 80 |
pipeline_tag: text-generation
|
| 81 |
---
|
|
@@ -89,17 +88,15 @@ pipeline_tag: text-generation
|
|
| 89 |
<aside>by <a href="https://curiosity.tech">Curiosity Technology</a></aside>
|
| 90 |
</div>
|
| 91 |
|
| 92 |
-
MARS is the
|
| 93 |
|
| 94 |
-
We
|
| 95 |
-
translations.
|
| 96 |
-
It is our intention to release Turkish translations in near future for community to have their go on them.
|
| 97 |
|
| 98 |
MARS have been trained for 3 days on 4xA100.
|
| 99 |
|
| 100 |
## Model Details
|
| 101 |
|
| 102 |
-
- **Base Model**: Meta Llama 3 8B Instruct
|
| 103 |
- **Training Dataset**: In-house & Translated Open Source Turkish Datasets
|
| 104 |
- **Training Method**: LoRA Fine Tuning
|
| 105 |
|
|
|
|
| 19 |
num_few_shot: 25
|
| 20 |
metrics:
|
| 21 |
- type: acc
|
| 22 |
+
value: 43.85
|
| 23 |
name: accuracy
|
| 24 |
- task:
|
| 25 |
type: text-generation
|
| 26 |
name: Text Generation
|
| 27 |
dataset:
|
| 28 |
+
name: HellaSwag TR
|
| 29 |
+
type: hellaswag
|
| 30 |
+
split: validation
|
|
|
|
| 31 |
args:
|
| 32 |
+
num_few_shot: 10
|
| 33 |
metrics:
|
| 34 |
- type: acc
|
| 35 |
+
value: 46.64
|
| 36 |
name: accuracy
|
| 37 |
- task:
|
| 38 |
type: text-generation
|
|
|
|
| 47 |
metrics:
|
| 48 |
- type: acc
|
| 49 |
name: accuracy
|
| 50 |
+
value: 48.66
|
| 51 |
- task:
|
| 52 |
type: text-generation
|
| 53 |
name: Text Generation
|
|
|
|
| 60 |
num_few_shot: 5
|
| 61 |
metrics:
|
| 62 |
- type: acc
|
| 63 |
+
value: 52.84
|
| 64 |
name: accuracy
|
| 65 |
- task:
|
| 66 |
type: text-generation
|
|
|
|
| 74 |
num_few_shot: 5
|
| 75 |
metrics:
|
| 76 |
- type: acc
|
| 77 |
+
value: 59.30
|
| 78 |
name: accuracy
|
| 79 |
pipeline_tag: text-generation
|
| 80 |
---
|
|
|
|
| 88 |
<aside>by <a href="https://curiosity.tech">Curiosity Technology</a></aside>
|
| 89 |
</div>
|
| 90 |
|
| 91 |
+
MARS-v0.2 is the second iteration of Curiosity Technology models, built on the foundation of Llama 3.1 8B. This version expands upon the initial MARS model by fine-tuning it with a more comprehensive dataset, with an increased emphasis on mathematical data to enhance its reasoning and problem-solving capabilities.
|
| 92 |
|
| 93 |
+
We've continued our commitment to Turkish language processing, utilizing both in-house Turkish datasets and a broader selection of translated open-source datasets. We believe this version will serve the community with even more versatility and depth.
|
|
|
|
|
|
|
| 94 |
|
| 95 |
MARS have been trained for 3 days on 4xA100.
|
| 96 |
|
| 97 |
## Model Details
|
| 98 |
|
| 99 |
+
- **Base Model**: Meta Llama 3.1 8B Instruct
|
| 100 |
- **Training Dataset**: In-house & Translated Open Source Turkish Datasets
|
| 101 |
- **Training Method**: LoRA Fine Tuning
|
| 102 |
|