Update README.md
Browse files
README.md
CHANGED
|
@@ -37,11 +37,7 @@ print(output["generated_text"])
|
|
| 37 |
## Evals
|
| 38 |
|
| 39 |
|
| 40 |
-
|
| 41 |
-
|
| 42 |
-
## Training procedure
|
| 43 |
-
|
| 44 |
-
SFT dataset mixture:
|
| 45 |
|
| 46 |
|dataset|weight (examples)| weight (tokens)|
|
| 47 |
|:------|:----:|:----:|
|
|
@@ -49,6 +45,9 @@ SFT dataset mixture:
|
|
| 49 |
|DebateLabKIT/deep-argmap-conversations|25%|18%|
|
| 50 |
|allenai/tulu-3-sft-mixture|50%|33%|
|
| 51 |
|
|
|
|
|
|
|
|
|
|
| 52 |
Trained with SFT on **1M examples** and for 1 epoch with
|
| 53 |
|
| 54 |
* context length 8196
|
|
|
|
| 37 |
## Evals
|
| 38 |
|
| 39 |
|
| 40 |
+
## SFT dataset mixture
|
|
|
|
|
|
|
|
|
|
|
|
|
| 41 |
|
| 42 |
|dataset|weight (examples)| weight (tokens)|
|
| 43 |
|:------|:----:|:----:|
|
|
|
|
| 45 |
|DebateLabKIT/deep-argmap-conversations|25%|18%|
|
| 46 |
|allenai/tulu-3-sft-mixture|50%|33%|
|
| 47 |
|
| 48 |
+
|
| 49 |
+
## Training procedure
|
| 50 |
+
|
| 51 |
Trained with SFT on **1M examples** and for 1 epoch with
|
| 52 |
|
| 53 |
* context length 8196
|