Update README.md
Browse files
README.md
CHANGED
|
@@ -4,7 +4,9 @@ license: apache-2.0
|
|
| 4 |
|
| 5 |
Converted using ggerganov/ggml's "stablelm" conversion script and quantization code as of [commit 05f3079](https://github.com/ggerganov/ggml/tree/05f307971862b83df12fada0c42ee027ba5a82b5/examples/stablelm) (2023-04-20). Should be compatible with KoboldCpp.
|
| 6 |
|
| 7 |
-
|
|
|
|
|
|
|
| 8 |
|
| 9 |
# RAM USAGE (koboldcpp):
|
| 10 |
Model | Loaded RAM usage
|
|
|
|
| 4 |
|
| 5 |
Converted using ggerganov/ggml's "stablelm" conversion script and quantization code as of [commit 05f3079](https://github.com/ggerganov/ggml/tree/05f307971862b83df12fada0c42ee027ba5a82b5/examples/stablelm) (2023-04-20). Should be compatible with KoboldCpp.
|
| 6 |
|
| 7 |
+
These conversions are based off of [the updated Pythia Deduped checkpoints](https://huggingface.co/models?sort=downloads&search=pythia+deduped), not [the original v0 trainings](https://huggingface.co/models?search=pythia%20deduped%20v0).
|
| 8 |
+
|
| 9 |
+
**NOTE:** pythia-70m-deduped gave garbage output in my testing.
|
| 10 |
|
| 11 |
# RAM USAGE (koboldcpp):
|
| 12 |
Model | Loaded RAM usage
|