Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -117,8 +117,7 @@ The model was evaluated using hellaswag dataset. TBD results.
|
|
| 117 |
|
| 118 |
The `nano-gpt train` command has built in support for evaluating against
|
| 119 |
the val dataset as well as HellaSwag in between training steps. Every 500 steps
|
| 120 |
-
the model was evaluated against the val dataset
|
| 121 |
-
geneated annecdotal samples from each worker.
|
| 122 |
|
| 123 |
## Environmental Impact
|
| 124 |
|
|
|
|
| 117 |
|
| 118 |
The `nano-gpt train` command has built in support for evaluating against
|
| 119 |
the val dataset as well as HellaSwag in between training steps. Every 500 steps
|
| 120 |
+
the model was evaluated against the val dataset and HellaSwag.
|
|
|
|
| 121 |
|
| 122 |
## Environmental Impact
|
| 123 |
|