Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,93 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: apache-2.0
|
| 3 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
---
|
| 4 |
+
|
| 5 |
+
gobean: the q4_k_m in the main model repo was hitting some tokenizer issues - after about five prompt exchanges it goes wild with repeats and unprintable tokens. I made this for comparison, it's not as bad as the q4_k_m but it isn't perfect. Both models fail the knowledge benchmark "describe the difference between a bear credit spread and a poor man's covered call," but they come really really close to getting it (just describes a standard covered call). Overall not bad. Inference is fast with q4_0 on 24gb vram - and the bug could very likely be with llama.cpp, so I may start looking into other frontends.
|
| 6 |
+
|
| 7 |
+
# ~ Original Model Card ~
|
| 8 |
+
|
| 9 |
+
<div align="center">
|
| 10 |
+
|
| 11 |
+
<picture>
|
| 12 |
+
<img src="https://raw.githubusercontent.com/01-ai/Yi/main/assets/img/Yi_logo_icon_light.svg" width="150px">
|
| 13 |
+
</picture>
|
| 14 |
+
|
| 15 |
+
</div>
|
| 16 |
+
|
| 17 |
+
<p align="center">
|
| 18 |
+
<a href="https://github.com/01-ai">🐙 GitHub</a> •
|
| 19 |
+
<a href="https://discord.gg/hYUwWddeAu">👾 Discord</a> •
|
| 20 |
+
<a href="https://twitter.com/01ai_yi">🐤 Twitter</a> •
|
| 21 |
+
<a href="https://github.com/01-ai/Yi-1.5/issues/2">💬 WeChat</a>
|
| 22 |
+
<br/>
|
| 23 |
+
<a href="https://arxiv.org/abs/2403.04652">📝 Paper</a> •
|
| 24 |
+
<a href="https://github.com/01-ai/Yi/tree/main?tab=readme-ov-file#faq">🙌 FAQ</a> •
|
| 25 |
+
<a href="https://github.com/01-ai/Yi/tree/main?tab=readme-ov-file#learning-hub">📗 Learning Hub</a>
|
| 26 |
+
</p>
|
| 27 |
+
|
| 28 |
+
# Intro
|
| 29 |
+
|
| 30 |
+
Yi-1.5 is an upgraded version of Yi. It is continuously pre-trained on Yi with a high-quality corpus of 500B tokens and fine-tuned on 3M diverse fine-tuning samples.
|
| 31 |
+
|
| 32 |
+
Compared with Yi, Yi-1.5 delivers stronger performance in coding, math, reasoning, and instruction-following capability, while still maintaining excellent capabilities in language understanding, commonsense reasoning, and reading comprehension.
|
| 33 |
+
|
| 34 |
+
<div align="center">
|
| 35 |
+
|
| 36 |
+
Model | Context Length | Pre-trained Tokens
|
| 37 |
+
| :------------: | :------------: | :------------: |
|
| 38 |
+
| Yi-1.5 | 4K | 3.6T
|
| 39 |
+
|
| 40 |
+
</div>
|
| 41 |
+
|
| 42 |
+
# Models
|
| 43 |
+
|
| 44 |
+
- Chat models
|
| 45 |
+
|
| 46 |
+
<div align="center">
|
| 47 |
+
|
| 48 |
+
| Name | Download |
|
| 49 |
+
| --------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
|
| 50 |
+
| Yi-1.5-34B-Chat | • [🤗 Hugging Face](https://huggingface.co/collections/01-ai/yi-15-2024-05-663f3ecab5f815a3eaca7ca8) • [🤖 ModelScope](https://www.modelscope.cn/organization/01ai) |
|
| 51 |
+
| Yi-1.5-9B-Chat | • [🤗 Hugging Face](https://huggingface.co/collections/01-ai/yi-15-2024-05-663f3ecab5f815a3eaca7ca8) • [🤖 ModelScope](https://www.modelscope.cn/organization/01ai) |
|
| 52 |
+
| Yi-1.5-6B-Chat | • [🤗 Hugging Face](https://huggingface.co/collections/01-ai/yi-15-2024-05-663f3ecab5f815a3eaca7ca8) • [🤖 ModelScope](https://www.modelscope.cn/organization/01ai) |
|
| 53 |
+
|
| 54 |
+
</div>
|
| 55 |
+
|
| 56 |
+
- Base models
|
| 57 |
+
|
| 58 |
+
<div align="center">
|
| 59 |
+
|
| 60 |
+
| Name | Download |
|
| 61 |
+
| ---------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
|
| 62 |
+
| Yi-1.5-34B | • [🤗 Hugging Face](https://huggingface.co/collections/01-ai/yi-15-2024-05-663f3ecab5f815a3eaca7ca8) • [🤖 ModelScope](https://www.modelscope.cn/organization/01ai) |
|
| 63 |
+
| Yi-1.5-9B | • [🤗 Hugging Face](https://huggingface.co/collections/01-ai/yi-15-2024-05-663f3ecab5f815a3eaca7ca8) • [🤖 ModelScope](https://www.modelscope.cn/organization/01ai) |
|
| 64 |
+
| Yi-1.5-6B | • [🤗 Hugging Face](https://huggingface.co/collections/01-ai/yi-15-2024-05-663f3ecab5f815a3eaca7ca8) • [🤖 ModelScope](https://www.modelscope.cn/organization/01ai) |
|
| 65 |
+
|
| 66 |
+
</div>
|
| 67 |
+
|
| 68 |
+
# Benchmarks
|
| 69 |
+
|
| 70 |
+
- Chat models
|
| 71 |
+
|
| 72 |
+
Yi-1.5-34B-Chat is on par with or excels beyond larger models in most benchmarks.
|
| 73 |
+
|
| 74 |
+
|
| 75 |
+
|
| 76 |
+
Yi-1.5-9B-Chat is the top performer among similarly sized open-source models.
|
| 77 |
+
|
| 78 |
+
|
| 79 |
+
|
| 80 |
+
- Base models
|
| 81 |
+
|
| 82 |
+
Yi-1.5-34B is on par with or excels beyond larger models in some benchmarks.
|
| 83 |
+
|
| 84 |
+
|
| 85 |
+
|
| 86 |
+
Yi-1.5-9B is the top performer among similarly sized open-source models.
|
| 87 |
+
|
| 88 |
+
|
| 89 |
+
|
| 90 |
+
# Quick Start
|
| 91 |
+
|
| 92 |
+
For getting up and running with Yi-1.5 models quickly, see [README](https://github.com/01-ai/Yi-1.5).
|
| 93 |
+
|