Upload README.md with huggingface_hub
Browse files
README.md
ADDED
|
@@ -0,0 +1,68 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
|
| 2 |
+
---
|
| 3 |
+
|
| 4 |
+
license: apache-2.0
|
| 5 |
+
datasets:
|
| 6 |
+
- AuriAetherwiing/Allura
|
| 7 |
+
- kalomaze/Opus_Instruct_25k
|
| 8 |
+
base_model:
|
| 9 |
+
- AuriAetherwiing/Yi-1.5-9B-32K-tokfix
|
| 10 |
+
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
[](https://hf.co/QuantFactory)
|
| 14 |
+
|
| 15 |
+
|
| 16 |
+
# QuantFactory/EVA-Yi-1.5-9B-32K-V1-GGUF
|
| 17 |
+
This is quantized version of [EVA-UNIT-01/EVA-Yi-1.5-9B-32K-V1](https://huggingface.co/EVA-UNIT-01/EVA-Yi-1.5-9B-32K-V1) created using llama.cpp
|
| 18 |
+
|
| 19 |
+
# Original Model Card
|
| 20 |
+
|
| 21 |
+
|
| 22 |
+
**EVA Yi 1.5 9B v1**
|
| 23 |
+
|
| 24 |
+
<p>
|
| 25 |
+
A RP/storywriting focused model, full-parameter finetune of Yi-1.5-9B-32K on mixture of synthetic and natural data.<br>
|
| 26 |
+
A continuation of nothingiisreal's Celeste 1.x series, made to improve stability and versatility, without losing unique, diverse writing style of Celeste.
|
| 27 |
+
</p>
|
| 28 |
+
|
| 29 |
+
<p>
|
| 30 |
+
<h3>Quants: (GGUF is not recommended, lcpp breaks tokenizer fix)</h3>
|
| 31 |
+
<ul>
|
| 32 |
+
<li><a href=https://huggingface.co/bartowski/EVA-Yi-1.5-9B-32K-V1-GGUF>IMatrix GGUF by bartowski</a></li>
|
| 33 |
+
<li><a href=https://huggingface.co/mradermacher/EVA-Yi-1.5-9B-32K-V1-GGUF>Static GGUF by Mradermacher</a></li>
|
| 34 |
+
<li><a href=https://huggingface.co/bartowski/EVA-Yi-1.5-9B-32K-V1-exl2>EXL2 by bartowski</a></li>
|
| 35 |
+
</ul>
|
| 36 |
+
We recommend using original BFloat16 weights, quantization seems to affect Yi significantly more than other model architectures.
|
| 37 |
+
</p>
|
| 38 |
+
<p>
|
| 39 |
+
Prompt format is ChatML.<br>
|
| 40 |
+
<h3>Recommended sampler values:</h3>
|
| 41 |
+
|
| 42 |
+
- Temperature: 1
|
| 43 |
+
- Min-P: 0.05
|
| 44 |
+
|
| 45 |
+
<h3>Recommended SillyTavern presets (via CalamitousFelicitousness):</h3>
|
| 46 |
+
|
| 47 |
+
- [Context](https://huggingface.co/EVA-UNIT-01/EVA-Yi-1.5-9B-32K-V1/blob/main/%5BChatML%5D%20Roleplay-v1.9%20Context.json)
|
| 48 |
+
- [Instruct and System Prompt](https://huggingface.co/EVA-UNIT-01/EVA-Yi-1.5-9B-32K-V1/blob/main/%5BChatML%5D%20Roleplay-v1.9%20Instruct.json)
|
| 49 |
+
</p>
|
| 50 |
+
|
| 51 |
+
<p>
|
| 52 |
+
<br>
|
| 53 |
+
<h3>
|
| 54 |
+
Training data:
|
| 55 |
+
</h3>
|
| 56 |
+
<ul>
|
| 57 |
+
<li>Celeste 70B 0.1 data mixture minus Opus Instruct subset. See that model's <a href=https://huggingface.co/nothingiisreal/L3.1-70B-Celeste-V0.1-BF16>card</a> for details.</li>
|
| 58 |
+
<li>Kalomaze's Opus_Instruct_25k dataset, filtered for refusals.</li></ul>
|
| 59 |
+
<h3>
|
| 60 |
+
Hardware used:
|
| 61 |
+
</h3>
|
| 62 |
+
<ul><li>4x3090Ti for 5 days.</li></ul><br>
|
| 63 |
+
</p>
|
| 64 |
+
Model was trained by Kearm and Auri.
|
| 65 |
+
<h4>Special thanks:</h4><ul>
|
| 66 |
+
<li>to Lemmy, Gryphe, Kalomaze and Nopm for the data</li>
|
| 67 |
+
<li>to ALK, Fizz and CalamitousFelicitousness for Yi tokenizer fix</li>
|
| 68 |
+
<li>and to InfermaticAI's community for their continued support for our endeavors</li></ul>
|