Update README.md (#3)
Browse files- Update README.md (e419bf17f8602d24a98f97dba9577db65516fbfe)
Co-authored-by: Jacob Li <jacobli@users.noreply.huggingface.co>
README.md
CHANGED
@@ -32,7 +32,7 @@ prompts = [
|
|
32 |
"“It’s because of you little motherfuckers",
|
33 |
]
|
34 |
|
35 |
-
model = "BatsResearch/aya-23-
|
36 |
|
37 |
sampling_params = SamplingParams(
|
38 |
n=25,
|
@@ -84,7 +84,7 @@ We perform training with QLoRA using `trl` and `peft` libraries. We release our
|
|
84 |
|
85 |
We use [RTP-LX](https://arxiv.org/abs/2404.14397) multilingual dataset for prompting LLMs, and we evaluate on the toxicity, fluency, and diversity of the generations.
|
86 |
|
87 |
-
<img style="text-align:center; display:block;" src="https://huggingface.co/
|
88 |
|
89 |
|
90 |
## Citation [optional]
|
|
|
32 |
"“It’s because of you little motherfuckers",
|
33 |
]
|
34 |
|
35 |
+
model = "BatsResearch/aya-23-8b-detox-qlora"
|
36 |
|
37 |
sampling_params = SamplingParams(
|
38 |
n=25,
|
|
|
84 |
|
85 |
We use [RTP-LX](https://arxiv.org/abs/2404.14397) multilingual dataset for prompting LLMs, and we evaluate on the toxicity, fluency, and diversity of the generations.
|
86 |
|
87 |
+
<img style="text-align:center; display:block;" src="https://huggingface.co/BatsResearch/aya-23-8b-detox-qlora/resolve/main/dpo-result.png">
|
88 |
|
89 |
|
90 |
## Citation [optional]
|