Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,113 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
base_model: Qwen/Qwen3-VL-30B-A3B-Instruct
|
| 4 |
+
base_model_relation: quantized
|
| 5 |
+
quantized_by: turboderp
|
| 6 |
+
tags:
|
| 7 |
+
- exl3
|
| 8 |
+
---
|
| 9 |
+
|
| 10 |
+
EXL3 quants of [Qwen3-VL-30B-A3B-Instruct](https://huggingface.co/Qwen/Qwen3-VL-30B-A3B-Instruct)
|
| 11 |
+
|
| 12 |
+
⚠️ Requires ExLlamaV3 v0.0.13 (or v0.0.12 `dev` branch)
|
| 13 |
+
|
| 14 |
+
Base bitrates:
|
| 15 |
+
|
| 16 |
+
[2.00 bits per weight](https://huggingface.co/turboderp/Qwen3-VL-30B-A3B-Instruct-exl3/2.00bpw)
|
| 17 |
+
[3.00 bits per weight](https://huggingface.co/turboderp/Qwen3-VL-30B-A3B-Instruct-exl3/3.00bpw)
|
| 18 |
+
[4.00 bits per weight](https://huggingface.co/turboderp/Qwen3-VL-30B-A3B-Instruct-exl3/4.00bpw)
|
| 19 |
+
[5.00 bits per weight](https://huggingface.co/turboderp/Qwen3-VL-30B-A3B-Instruct-exl3/5.00bpw)
|
| 20 |
+
[6.00 bits per weight](https://huggingface.co/turboderp/Qwen3-VL-30B-A3B-Instruct-exl3/6.00bpw)
|
| 21 |
+
|
| 22 |
+
Optimized:
|
| 23 |
+
|
| 24 |
+
[2.09 bits per weight](https://huggingface.co/turboderp/Qwen3-VL-30B-A3B-Instruct-exl3/2.09bpw)
|
| 25 |
+
[2.28 bits per weight](https://huggingface.co/turboderp/Qwen3-VL-30B-A3B-Instruct-exl3/2.28bpw)
|
| 26 |
+
[2.54 bits per weight](https://huggingface.co/turboderp/Qwen3-VL-30B-A3B-Instruct-exl3/2.54bpw)
|
| 27 |
+
[2.84 bits per weight](https://huggingface.co/turboderp/Qwen3-VL-30B-A3B-Instruct-exl3/2.84bpw)
|
| 28 |
+
[3.09 bits per weight](https://huggingface.co/turboderp/Qwen3-VL-30B-A3B-Instruct-exl3/3.09bpw)
|
| 29 |
+
[3.54 bits per weight](https://huggingface.co/turboderp/Qwen3-VL-30B-A3B-Instruct-exl3/3.54bpw)
|
| 30 |
+
[4.06 bits per weight](https://huggingface.co/turboderp/Qwen3-VL-30B-A3B-Instruct-exl3/4.06bpw)
|
| 31 |
+
|
| 32 |
+
# SVG Catbench
|
| 33 |
+
|
| 34 |
+
<table>
|
| 35 |
+
<tr>
|
| 36 |
+
<td align="center">
|
| 37 |
+
<a href="https://huggingface.co/turboderp/Qwen3-VL-8B-Instruct-exl3/blob/main/2.00bpw.svg">
|
| 38 |
+
<img src="2.00bpw.svg" alt="2.00 bpw" width="160">
|
| 39 |
+
</a>
|
| 40 |
+
<div>2.00 bpw</div>
|
| 41 |
+
</td>
|
| 42 |
+
<td align="center">
|
| 43 |
+
<a href="https://huggingface.co/turboderp/Qwen3-VL-8B-Instruct-exl3/blob/main/2.28bpw.svg">
|
| 44 |
+
<img src="2.28bpw.svg" alt="2.28 bpw" width="160">
|
| 45 |
+
</a>
|
| 46 |
+
<div>2.28 bpw</div>
|
| 47 |
+
</td>
|
| 48 |
+
<td align="center">
|
| 49 |
+
<a href="https://huggingface.co/turboderp/Qwen3-VL-8B-Instruct-exl3/blob/main/2.54bpw.svg">
|
| 50 |
+
<img src="2.54bpw.svg" alt="2.54 bpw" width="160">
|
| 51 |
+
</a>
|
| 52 |
+
<div>2.54 bpw</div>
|
| 53 |
+
</td>
|
| 54 |
+
<td align="center">
|
| 55 |
+
<a href="https://huggingface.co/turboderp/Qwen3-VL-8B-Instruct-exl3/blob/main/2.84bpw.svg">
|
| 56 |
+
<img src="2.84bpw.svg" alt="2.84 bpw" width="160">
|
| 57 |
+
</a>
|
| 58 |
+
<div>2.84 bpw</div>
|
| 59 |
+
</td>
|
| 60 |
+
</tr>
|
| 61 |
+
<tr>
|
| 62 |
+
<td align="center">
|
| 63 |
+
<a href="https://huggingface.co/turboderp/Qwen3-VL-8B-Instruct-exl3/blob/main/3.00bpw.svg">
|
| 64 |
+
<img src="3.00bpw.svg" alt="3.00 bpw" width="160">
|
| 65 |
+
</a>
|
| 66 |
+
<div>3.00 bpw</div>
|
| 67 |
+
</td>
|
| 68 |
+
<td align="center">
|
| 69 |
+
<a href="https://huggingface.co/turboderp/Qwen3-VL-8B-Instruct-exl3/blob/main/3.09bpw.svg">
|
| 70 |
+
<img src="3.09bpw.svg" alt="3.09 bpw" width="160">
|
| 71 |
+
</a>
|
| 72 |
+
<div>3.09 bpw</div>
|
| 73 |
+
</td>
|
| 74 |
+
<td align="center">
|
| 75 |
+
<a href="https://huggingface.co/turboderp/Qwen3-VL-8B-Instruct-exl3/blob/main/3.54bpw.svg">
|
| 76 |
+
<img src="3.54bpw.svg" alt="3.54 bpw" width="160">
|
| 77 |
+
</a>
|
| 78 |
+
<div>3.54 bpw</div>
|
| 79 |
+
</td>
|
| 80 |
+
<td align="center">
|
| 81 |
+
<a href="https://huggingface.co/turboderp/Qwen3-VL-8B-Instruct-exl3/blob/main/4.00bpw.svg">
|
| 82 |
+
<img src="4.00bpw.svg" alt="4.00 bpw" width="160">
|
| 83 |
+
</a>
|
| 84 |
+
<div>4.00 bpw</div>
|
| 85 |
+
</td>
|
| 86 |
+
</tr>
|
| 87 |
+
<tr>
|
| 88 |
+
<td align="center">
|
| 89 |
+
<a href="https://huggingface.co/turboderp/Qwen3-VL-8B-Instruct-exl3/blob/main/4.09bpw.svg">
|
| 90 |
+
<img src="4.09bpw.svg" alt="4.09 bpw" width="160">
|
| 91 |
+
</a>
|
| 92 |
+
<div>4.09 bpw</div>
|
| 93 |
+
</td>
|
| 94 |
+
<td align="center">
|
| 95 |
+
<a href="https://huggingface.co/turboderp/Qwen3-VL-8B-Instruct-exl3/blob/main/5.00bpw.svg">
|
| 96 |
+
<img src="5.0bpw.svg" alt="5.00 bpw" width="160">
|
| 97 |
+
</a>
|
| 98 |
+
<div>5.00 bpw</div>
|
| 99 |
+
</td>
|
| 100 |
+
<td align="center">
|
| 101 |
+
<a href="https://huggingface.co/turboderp/Qwen3-VL-8B-Instruct-exl3/blob/main/6.00bpw.svg">
|
| 102 |
+
<img src="6.0bpw.svg" alt="6.00 bpw" width="160">
|
| 103 |
+
</a>
|
| 104 |
+
<div>6.00 bpw</div>
|
| 105 |
+
</td>
|
| 106 |
+
<td align="center">
|
| 107 |
+
<a href="https://huggingface.co/turboderp/Qwen3-VL-8B-Instruct-exl3/blob/main/api.svg">
|
| 108 |
+
<img src="api.svg" alt="API" width="160">
|
| 109 |
+
</a>
|
| 110 |
+
<div>API</div>
|
| 111 |
+
</td>
|
| 112 |
+
</tr>
|
| 113 |
+
</table>
|