Improved the description.
Browse files
README.md
CHANGED
@@ -183,19 +183,19 @@ print(output[0].outputs[0].text)
|
|
183 |
|
184 |
### Instruction Tuning
|
185 |
|
186 |
-
The following
|
|
|
187 |
- Japanese
|
188 |
- `lmsys-chat-1m-synth-ja-wo-pii-and-template-instructions`
|
189 |
-
- Single-turn Japanese
|
190 |
-
|
191 |
-
- As implied by the dataset name, conversations that contain personally identifiable information (PII) or template-based user instructions have been removed. Duplicate instuctions have also been removed.
|
192 |
- `filtered-magpie-ultra-ja`
|
193 |
-
- A Japanese variant of the `filtered-magpie-ultra-en` dataset,
|
194 |
- `gemma-magpie`
|
195 |
-
- A Japanese synthetic Q&A dataset from scratch, generated
|
196 |
- English
|
197 |
- `lmsys-chat-1m-synth-en-wo-pii-and-template-instructions`
|
198 |
-
-
|
199 |
- `filtered-magpie-ultra-en`
|
200 |
- A subset of the [magpie-ultra](https://huggingface.co/datasets/argilla/magpie-ultra-v0.1) dataset, developed following the MAGPIE recipe [\[Xu+, arXiv24\]](https://arxiv.org/abs/2406.08464) using [Llama-3.1-405B-Instruct](https://huggingface.co/meta-llama/Llama-3.1-405B-Instruct). This subset includes only samples rated as 'average,' 'good,' or 'excellent.'
|
201 |
|
|
|
183 |
|
184 |
### Instruction Tuning
|
185 |
|
186 |
+
The following datasets were used for the instruction tuning.
|
187 |
+
|
188 |
- Japanese
|
189 |
- `lmsys-chat-1m-synth-ja-wo-pii-and-template-instructions`
|
190 |
+
- Single-turn Japanese instruction dataset synthesized and derived from [lmsys-chat-1m](https://huggingface.co/datasets/lmsys/lmsys-chat-1m) [\[Zhang+, ICLR24\]](https://openreview.net/forum?id=BOfDKxfwt0)). First-turn user instructions were translated into Japanese via DeepL (machine translation), and assistant responses were generated using [Llama-3.1-405B-Instruct](https://huggingface.co/meta-llama/Llama-3.1-405B-Instruct). [Llama-3.1-70B-Instruct](https://huggingface.co/meta-llama/Llama-3.1-70B-Instruct) served as a judge for rejection sampling (n=6).
|
191 |
+
Conversations containing personally identifiable information (PII) and template-based user instructions were removed. Duplicate instructions were removed.
|
|
|
192 |
- `filtered-magpie-ultra-ja`
|
193 |
+
- A Japanese variant of the `filtered-magpie-ultra-en` dataset, translated into Japanese using [gemma-2-27b-it](https://huggingface.co/google/gemma-2-27b-it).
|
194 |
- `gemma-magpie`
|
195 |
+
- A Japanese synthetic Q&A dataset from scratch, generated by [gemma-2-27b-it](https://huggingface.co/google/gemma-2-27b-it). User instructions were created with prompts specific to each topic, and assistant responses were generated for these instructions. The conversations were then heuristically filtered for quality and length.
|
196 |
- English
|
197 |
- `lmsys-chat-1m-synth-en-wo-pii-and-template-instructions`
|
198 |
+
- The creation process is similar to `lmsys-chat-1m-synth-ja-wo-pii-and-template-instructions`, but this version uses the original English user instructions. The assistant responses were generated in English as well. Rejection sampling was not applied in this version.
|
199 |
- `filtered-magpie-ultra-en`
|
200 |
- A subset of the [magpie-ultra](https://huggingface.co/datasets/argilla/magpie-ultra-v0.1) dataset, developed following the MAGPIE recipe [\[Xu+, arXiv24\]](https://arxiv.org/abs/2406.08464) using [Llama-3.1-405B-Instruct](https://huggingface.co/meta-llama/Llama-3.1-405B-Instruct). This subset includes only samples rated as 'average,' 'good,' or 'excellent.'
|
201 |
|