mradermacher commited on
Commit
3a5f3fa
·
verified ·
1 Parent(s): 0006232

auto-patch README.md

Browse files
Files changed (1) hide show
  1. README.md +12 -0
README.md CHANGED
@@ -36,17 +36,29 @@ more details, including on how to concatenate multi-part files.
36
 
37
  | Link | Type | Size/GB | Notes |
38
  |:-----|:-----|--------:|:------|
 
39
  | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-IQ1_M.gguf) | i1-IQ1_M | 3.3 | mostly desperate |
40
  | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-IQ2_XXS.gguf) | i1-IQ2_XXS | 3.5 | |
 
 
41
  | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-IQ2_M.gguf) | i1-IQ2_M | 4.1 | |
42
  | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-Q2_K.gguf) | i1-Q2_K | 4.1 | IQ3_XXS probably better |
43
  | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 4.4 | lower quality |
 
 
44
  | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-Q3_K_S.gguf) | i1-Q3_K_S | 4.7 | IQ3_XS probably better |
45
  | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-IQ3_M.gguf) | i1-IQ3_M | 5.0 | |
46
  | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-Q3_K_M.gguf) | i1-Q3_K_M | 5.2 | IQ3_S probably better |
47
  | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-IQ4_XS.gguf) | i1-IQ4_XS | 5.4 | |
 
 
 
 
 
48
  | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-Q4_K_S.gguf) | i1-Q4_K_S | 5.9 | optimal size/speed/quality |
49
  | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-Q4_K_M.gguf) | i1-Q4_K_M | 6.4 | fast, recommended |
 
 
50
  | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-Q6_K.gguf) | i1-Q6_K | 8.4 | practically like static Q6_K |
51
 
52
  Here is a handy graph by ikawrakow comparing some lower-quality quant
 
36
 
37
  | Link | Type | Size/GB | Notes |
38
  |:-----|:-----|--------:|:------|
39
+ | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-IQ1_S.gguf) | i1-IQ1_S | 3.2 | for the desperate |
40
  | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-IQ1_M.gguf) | i1-IQ1_M | 3.3 | mostly desperate |
41
  | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-IQ2_XXS.gguf) | i1-IQ2_XXS | 3.5 | |
42
+ | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-IQ2_XS.gguf) | i1-IQ2_XS | 3.7 | |
43
+ | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-IQ2_S.gguf) | i1-IQ2_S | 3.9 | |
44
  | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-IQ2_M.gguf) | i1-IQ2_M | 4.1 | |
45
  | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-Q2_K.gguf) | i1-Q2_K | 4.1 | IQ3_XXS probably better |
46
  | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 4.4 | lower quality |
47
+ | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-IQ3_XS.gguf) | i1-IQ3_XS | 4.6 | |
48
+ | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-IQ3_S.gguf) | i1-IQ3_S | 4.7 | beats Q3_K* |
49
  | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-Q3_K_S.gguf) | i1-Q3_K_S | 4.7 | IQ3_XS probably better |
50
  | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-IQ3_M.gguf) | i1-IQ3_M | 5.0 | |
51
  | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-Q3_K_M.gguf) | i1-Q3_K_M | 5.2 | IQ3_S probably better |
52
  | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-IQ4_XS.gguf) | i1-IQ4_XS | 5.4 | |
53
+ | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-Q3_K_L.gguf) | i1-Q3_K_L | 5.4 | IQ3_M probably better |
54
+ | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-Q4_0_4_4.gguf) | i1-Q4_0_4_4 | 5.6 | fast on arm, low quality |
55
+ | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-Q4_0_4_8.gguf) | i1-Q4_0_4_8 | 5.6 | fast on arm+i8mm, low quality |
56
+ | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-Q4_0_8_8.gguf) | i1-Q4_0_8_8 | 5.6 | fast on arm+sve, low quality |
57
+ | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-Q4_0.gguf) | i1-Q4_0 | 5.6 | fast, low quality |
58
  | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-Q4_K_S.gguf) | i1-Q4_K_S | 5.9 | optimal size/speed/quality |
59
  | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-Q4_K_M.gguf) | i1-Q4_K_M | 6.4 | fast, recommended |
60
+ | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-Q5_K_S.gguf) | i1-Q5_K_S | 6.9 | |
61
+ | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-Q5_K_M.gguf) | i1-Q5_K_M | 7.3 | |
62
  | [GGUF](https://huggingface.co/mradermacher/glm-4-9b-chat-1m-i1-GGUF/resolve/main/glm-4-9b-chat-1m.i1-Q6_K.gguf) | i1-Q6_K | 8.4 | practically like static Q6_K |
63
 
64
  Here is a handy graph by ikawrakow comparing some lower-quality quant