Update README.md
Browse files
README.md
CHANGED
|
@@ -29,6 +29,38 @@ All quants were made using the default [calibration files](https://github.com/tu
|
|
| 29 |
|
| 30 |
<hr>
|
| 31 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 32 |
# Perplexity table (the lower the better)
|
| 33 |
|
| 34 |
<hr>
|
|
|
|
| 29 |
|
| 30 |
<hr>
|
| 31 |
|
| 32 |
+
# Branches
|
| 33 |
+
|
| 34 |
+
| Precision | Size |
|
| 35 |
+
| ------------------------------------------------------------------------------------------- | ------- |
|
| 36 |
+
| [2.5bpw](https://huggingface.co/ThomasBaruzier/Qwen2.5-Coder-32B-Instruct-EXL2/tree/2.5bpw) | 11374MB |
|
| 37 |
+
| [2.6bpw](https://huggingface.co/ThomasBaruzier/Qwen2.5-Coder-32B-Instruct-EXL2/tree/2.6bpw) | 11743MB |
|
| 38 |
+
| [2.7bpw](https://huggingface.co/ThomasBaruzier/Qwen2.5-Coder-32B-Instruct-EXL2/tree/2.7bpw) | 12118MB |
|
| 39 |
+
| [2.8bpw](https://huggingface.co/ThomasBaruzier/Qwen2.5-Coder-32B-Instruct-EXL2/tree/2.8bpw) | 12490MB |
|
| 40 |
+
| [2.9bpw](https://huggingface.co/ThomasBaruzier/Qwen2.5-Coder-32B-Instruct-EXL2/tree/2.9bpw) | 12863MB |
|
| 41 |
+
| [3.0bpw](https://huggingface.co/ThomasBaruzier/Qwen2.5-Coder-32B-Instruct-EXL2/tree/3.0bpw) | 13234MB |
|
| 42 |
+
| [3.1bpw](https://huggingface.co/ThomasBaruzier/Qwen2.5-Coder-32B-Instruct-EXL2/tree/3.1bpw) | 13601MB |
|
| 43 |
+
| [3.2bpw](https://huggingface.co/ThomasBaruzier/Qwen2.5-Coder-32B-Instruct-EXL2/tree/3.2bpw) | 13978MB |
|
| 44 |
+
| [3.3bpw](https://huggingface.co/ThomasBaruzier/Qwen2.5-Coder-32B-Instruct-EXL2/tree/3.3bpw) | 14348MB |
|
| 45 |
+
| [3.4bpw](https://huggingface.co/ThomasBaruzier/Qwen2.5-Coder-32B-Instruct-EXL2/tree/3.4bpw) | 14722MB |
|
| 46 |
+
| [3.5bpw](https://huggingface.co/ThomasBaruzier/Qwen2.5-Coder-32B-Instruct-EXL2/tree/3.5bpw) | 15088MB |
|
| 47 |
+
| [3.6bpw](https://huggingface.co/ThomasBaruzier/Qwen2.5-Coder-32B-Instruct-EXL2/tree/3.6bpw) | 15466MB |
|
| 48 |
+
| [3.7bpw](https://huggingface.co/ThomasBaruzier/Qwen2.5-Coder-32B-Instruct-EXL2/tree/3.7bpw) | 15836MB |
|
| 49 |
+
| [3.8bpw](https://huggingface.co/ThomasBaruzier/Qwen2.5-Coder-32B-Instruct-EXL2/tree/3.8bpw) | 16195MB |
|
| 50 |
+
| [3.9bpw](https://huggingface.co/ThomasBaruzier/Qwen2.5-Coder-32B-Instruct-EXL2/tree/3.9bpw) | 16570MB |
|
| 51 |
+
| [4.0bpw](https://huggingface.co/ThomasBaruzier/Qwen2.5-Coder-32B-Instruct-EXL2/tree/4.0bpw) | 16954MB |
|
| 52 |
+
| [4.1bpw](https://huggingface.co/ThomasBaruzier/Qwen2.5-Coder-32B-Instruct-EXL2/tree/4.1bpw) | 17323MB |
|
| 53 |
+
| [4.2bpw](https://huggingface.co/ThomasBaruzier/Qwen2.5-Coder-32B-Instruct-EXL2/tree/4.2bpw) | 17694MB |
|
| 54 |
+
| [4.3bpw](https://huggingface.co/ThomasBaruzier/Qwen2.5-Coder-32B-Instruct-EXL2/tree/4.3bpw) | 18068MB |
|
| 55 |
+
| [4.4bpw](https://huggingface.co/ThomasBaruzier/Qwen2.5-Coder-32B-Instruct-EXL2/tree/4.4bpw) | 18440MB |
|
| 56 |
+
| [4.5bpw](https://huggingface.co/ThomasBaruzier/Qwen2.5-Coder-32B-Instruct-EXL2/tree/4.5bpw) | 18811MB |
|
| 57 |
+
| [5.0bpw](https://huggingface.co/ThomasBaruzier/Qwen2.5-Coder-32B-Instruct-EXL2/tree/5.0bpw) | 20657MB |
|
| 58 |
+
| [5.5bpw](https://huggingface.co/ThomasBaruzier/Qwen2.5-Coder-32B-Instruct-EXL2/tree/5.5bpw) | 22527MB |
|
| 59 |
+
| [6.0bpw](https://huggingface.co/ThomasBaruzier/Qwen2.5-Coder-32B-Instruct-EXL2/tree/6.0bpw) | 24389MB |
|
| 60 |
+
| [8.0bpw](https://huggingface.co/ThomasBaruzier/Qwen2.5-Coder-32B-Instruct-EXL2/tree/8.0bpw) | 30158MB |
|
| 61 |
+
|
| 62 |
+
<hr>
|
| 63 |
+
|
| 64 |
# Perplexity table (the lower the better)
|
| 65 |
|
| 66 |
<hr>
|