Update README.md
Browse files
README.md
CHANGED
|
@@ -17,12 +17,13 @@ metrics:
|
|
| 17 |
- accuracy
|
| 18 |
---
|
| 19 |
|
| 20 |
-
|
| 21 |
This repo contains the model and tokenizer checkpoints for:
|
| 22 |
- model family [<b>mistralai/Mistral-7B-Instruct-v0.2</b>](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2)
|
| 23 |
- optimized with the loss [<b>KTO</b>](https://twitter.com/winniethexu/status/1732839295365554643)
|
| 24 |
- aligned using the [snorkelai/Snorkel-Mistral-PairRM-DPO-Dataset](https://huggingface.co/datasets/snorkelai/Snorkel-Mistral-PairRM-DPO-Dataset)
|
| 25 |
-
- via 3 iterations of KTO on one epoch of each training partition.
|
|
|
|
|
|
|
| 26 |
|
| 27 |
To prompt this model, ensure that the format is consistent with that of TuluV2.
|
| 28 |
For example, a prompt should be formatted as follows, where `<|user|>` corresponds to the human's role and `<|assistant|>` corresponds to the LLM's role.
|
|
|
|
| 17 |
- accuracy
|
| 18 |
---
|
| 19 |
|
|
|
|
| 20 |
This repo contains the model and tokenizer checkpoints for:
|
| 21 |
- model family [<b>mistralai/Mistral-7B-Instruct-v0.2</b>](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2)
|
| 22 |
- optimized with the loss [<b>KTO</b>](https://twitter.com/winniethexu/status/1732839295365554643)
|
| 23 |
- aligned using the [snorkelai/Snorkel-Mistral-PairRM-DPO-Dataset](https://huggingface.co/datasets/snorkelai/Snorkel-Mistral-PairRM-DPO-Dataset)
|
| 24 |
+
- via 3 iterations of KTO on one epoch of each training partition, each previous iteration's model serving as the reference for the subsequeent.
|
| 25 |
+
|
| 26 |
+
**[03/06/2024]**: We are #2 on the (verified) [Alpaca Eval 2.0 Leaderboard](https://tatsu-lab.github.io/alpaca_eval/) scoring **33.23**!
|
| 27 |
|
| 28 |
To prompt this model, ensure that the format is consistent with that of TuluV2.
|
| 29 |
For example, a prompt should be formatted as follows, where `<|user|>` corresponds to the human's role and `<|assistant|>` corresponds to the LLM's role.
|