mradermacher commited on
Commit
065a989
·
verified ·
1 Parent(s): a8cc209

auto-patch README.md

Browse files
Files changed (1) hide show
  1. README.md +18 -0
README.md CHANGED
@@ -39,12 +39,30 @@ more details, including on how to concatenate multi-part files.
39
 
40
  | Link | Type | Size/GB | Notes |
41
  |:-----|:-----|--------:|:------|
 
 
 
 
 
 
 
42
  | [GGUF](https://huggingface.co/mradermacher/lambda-qwen2.5-14b-dpo-test-i1-GGUF/resolve/main/lambda-qwen2.5-14b-dpo-test.i1-Q2_K.gguf) | i1-Q2_K | 5.9 | IQ3_XXS probably better |
43
  | [GGUF](https://huggingface.co/mradermacher/lambda-qwen2.5-14b-dpo-test-i1-GGUF/resolve/main/lambda-qwen2.5-14b-dpo-test.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 6.0 | lower quality |
 
 
 
44
  | [GGUF](https://huggingface.co/mradermacher/lambda-qwen2.5-14b-dpo-test-i1-GGUF/resolve/main/lambda-qwen2.5-14b-dpo-test.i1-IQ3_M.gguf) | i1-IQ3_M | 7.0 | |
45
  | [GGUF](https://huggingface.co/mradermacher/lambda-qwen2.5-14b-dpo-test-i1-GGUF/resolve/main/lambda-qwen2.5-14b-dpo-test.i1-Q3_K_M.gguf) | i1-Q3_K_M | 7.4 | IQ3_S probably better |
 
 
 
46
  | [GGUF](https://huggingface.co/mradermacher/lambda-qwen2.5-14b-dpo-test-i1-GGUF/resolve/main/lambda-qwen2.5-14b-dpo-test.i1-IQ4_NL.gguf) | i1-IQ4_NL | 8.6 | prefer IQ4_XS |
47
  | [GGUF](https://huggingface.co/mradermacher/lambda-qwen2.5-14b-dpo-test-i1-GGUF/resolve/main/lambda-qwen2.5-14b-dpo-test.i1-Q4_K_S.gguf) | i1-Q4_K_S | 8.7 | optimal size/speed/quality |
 
 
 
 
 
48
 
49
  Here is a handy graph by ikawrakow comparing some lower-quality quant
50
  types (lower is better):
 
39
 
40
  | Link | Type | Size/GB | Notes |
41
  |:-----|:-----|--------:|:------|
42
+ | [GGUF](https://huggingface.co/mradermacher/lambda-qwen2.5-14b-dpo-test-i1-GGUF/resolve/main/lambda-qwen2.5-14b-dpo-test.i1-IQ1_S.gguf) | i1-IQ1_S | 3.7 | for the desperate |
43
+ | [GGUF](https://huggingface.co/mradermacher/lambda-qwen2.5-14b-dpo-test-i1-GGUF/resolve/main/lambda-qwen2.5-14b-dpo-test.i1-IQ1_M.gguf) | i1-IQ1_M | 4.0 | mostly desperate |
44
+ | [GGUF](https://huggingface.co/mradermacher/lambda-qwen2.5-14b-dpo-test-i1-GGUF/resolve/main/lambda-qwen2.5-14b-dpo-test.i1-IQ2_XXS.gguf) | i1-IQ2_XXS | 4.4 | |
45
+ | [GGUF](https://huggingface.co/mradermacher/lambda-qwen2.5-14b-dpo-test-i1-GGUF/resolve/main/lambda-qwen2.5-14b-dpo-test.i1-IQ2_XS.gguf) | i1-IQ2_XS | 4.8 | |
46
+ | [GGUF](https://huggingface.co/mradermacher/lambda-qwen2.5-14b-dpo-test-i1-GGUF/resolve/main/lambda-qwen2.5-14b-dpo-test.i1-IQ2_S.gguf) | i1-IQ2_S | 5.1 | |
47
+ | [GGUF](https://huggingface.co/mradermacher/lambda-qwen2.5-14b-dpo-test-i1-GGUF/resolve/main/lambda-qwen2.5-14b-dpo-test.i1-IQ2_M.gguf) | i1-IQ2_M | 5.5 | |
48
+ | [GGUF](https://huggingface.co/mradermacher/lambda-qwen2.5-14b-dpo-test-i1-GGUF/resolve/main/lambda-qwen2.5-14b-dpo-test.i1-Q2_K_S.gguf) | i1-Q2_K_S | 5.5 | very low quality |
49
  | [GGUF](https://huggingface.co/mradermacher/lambda-qwen2.5-14b-dpo-test-i1-GGUF/resolve/main/lambda-qwen2.5-14b-dpo-test.i1-Q2_K.gguf) | i1-Q2_K | 5.9 | IQ3_XXS probably better |
50
  | [GGUF](https://huggingface.co/mradermacher/lambda-qwen2.5-14b-dpo-test-i1-GGUF/resolve/main/lambda-qwen2.5-14b-dpo-test.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 6.0 | lower quality |
51
+ | [GGUF](https://huggingface.co/mradermacher/lambda-qwen2.5-14b-dpo-test-i1-GGUF/resolve/main/lambda-qwen2.5-14b-dpo-test.i1-IQ3_XS.gguf) | i1-IQ3_XS | 6.5 | |
52
+ | [GGUF](https://huggingface.co/mradermacher/lambda-qwen2.5-14b-dpo-test-i1-GGUF/resolve/main/lambda-qwen2.5-14b-dpo-test.i1-Q3_K_S.gguf) | i1-Q3_K_S | 6.8 | IQ3_XS probably better |
53
+ | [GGUF](https://huggingface.co/mradermacher/lambda-qwen2.5-14b-dpo-test-i1-GGUF/resolve/main/lambda-qwen2.5-14b-dpo-test.i1-IQ3_S.gguf) | i1-IQ3_S | 6.8 | beats Q3_K* |
54
  | [GGUF](https://huggingface.co/mradermacher/lambda-qwen2.5-14b-dpo-test-i1-GGUF/resolve/main/lambda-qwen2.5-14b-dpo-test.i1-IQ3_M.gguf) | i1-IQ3_M | 7.0 | |
55
  | [GGUF](https://huggingface.co/mradermacher/lambda-qwen2.5-14b-dpo-test-i1-GGUF/resolve/main/lambda-qwen2.5-14b-dpo-test.i1-Q3_K_M.gguf) | i1-Q3_K_M | 7.4 | IQ3_S probably better |
56
+ | [GGUF](https://huggingface.co/mradermacher/lambda-qwen2.5-14b-dpo-test-i1-GGUF/resolve/main/lambda-qwen2.5-14b-dpo-test.i1-Q3_K_L.gguf) | i1-Q3_K_L | 8.0 | IQ3_M probably better |
57
+ | [GGUF](https://huggingface.co/mradermacher/lambda-qwen2.5-14b-dpo-test-i1-GGUF/resolve/main/lambda-qwen2.5-14b-dpo-test.i1-IQ4_XS.gguf) | i1-IQ4_XS | 8.2 | |
58
+ | [GGUF](https://huggingface.co/mradermacher/lambda-qwen2.5-14b-dpo-test-i1-GGUF/resolve/main/lambda-qwen2.5-14b-dpo-test.i1-Q4_0.gguf) | i1-Q4_0 | 8.6 | fast, low quality |
59
  | [GGUF](https://huggingface.co/mradermacher/lambda-qwen2.5-14b-dpo-test-i1-GGUF/resolve/main/lambda-qwen2.5-14b-dpo-test.i1-IQ4_NL.gguf) | i1-IQ4_NL | 8.6 | prefer IQ4_XS |
60
  | [GGUF](https://huggingface.co/mradermacher/lambda-qwen2.5-14b-dpo-test-i1-GGUF/resolve/main/lambda-qwen2.5-14b-dpo-test.i1-Q4_K_S.gguf) | i1-Q4_K_S | 8.7 | optimal size/speed/quality |
61
+ | [GGUF](https://huggingface.co/mradermacher/lambda-qwen2.5-14b-dpo-test-i1-GGUF/resolve/main/lambda-qwen2.5-14b-dpo-test.i1-Q4_K_M.gguf) | i1-Q4_K_M | 9.1 | fast, recommended |
62
+ | [GGUF](https://huggingface.co/mradermacher/lambda-qwen2.5-14b-dpo-test-i1-GGUF/resolve/main/lambda-qwen2.5-14b-dpo-test.i1-Q4_1.gguf) | i1-Q4_1 | 9.5 | |
63
+ | [GGUF](https://huggingface.co/mradermacher/lambda-qwen2.5-14b-dpo-test-i1-GGUF/resolve/main/lambda-qwen2.5-14b-dpo-test.i1-Q5_K_S.gguf) | i1-Q5_K_S | 10.4 | |
64
+ | [GGUF](https://huggingface.co/mradermacher/lambda-qwen2.5-14b-dpo-test-i1-GGUF/resolve/main/lambda-qwen2.5-14b-dpo-test.i1-Q5_K_M.gguf) | i1-Q5_K_M | 10.6 | |
65
+ | [GGUF](https://huggingface.co/mradermacher/lambda-qwen2.5-14b-dpo-test-i1-GGUF/resolve/main/lambda-qwen2.5-14b-dpo-test.i1-Q6_K.gguf) | i1-Q6_K | 12.2 | practically like static Q6_K |
66
 
67
  Here is a handy graph by ikawrakow comparing some lower-quality quant
68
  types (lower is better):