Hugging Face's logo

mradermacher
/
MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF

Transformers
GGUF
Ukrainian
English
gemma3
instruct
mamaylm
insait
imatrix
conversational
Model card Files
xet
Community
mradermacher commited on
Commit
382abce
·
verified ·
1 Parent(s): 4f8ce13

auto-patch README.md

Browse files
Files changed (1) hide show
  1. README.md +17 -0
README.md CHANGED
@@ -53,13 +53,30 @@ more details, including on how to concatenate multi-part files.
53
  | Link | Type | Size/GB | Notes |
54
  |:-----|:-----|--------:|:------|
55
  | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.imatrix.gguf) | imatrix | 0.1 | imatrix file (for creating your own qwuants) |
 
 
 
 
 
 
 
56
  | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.i1-Q2_K.gguf) | i1-Q2_K | 4.9 | IQ3_XXS probably better |
57
  | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 4.9 | lower quality |
 
 
 
58
  | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.i1-IQ3_M.gguf) | i1-IQ3_M | 5.8 | |
59
  | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.i1-Q3_K_M.gguf) | i1-Q3_K_M | 6.1 | IQ3_S probably better |
 
 
60
  | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.i1-IQ4_NL.gguf) | i1-IQ4_NL | 7.0 | prefer IQ4_XS |
 
61
  | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.i1-Q4_K_S.gguf) | i1-Q4_K_S | 7.0 | optimal size/speed/quality |
62
  | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.i1-Q4_K_M.gguf) | i1-Q4_K_M | 7.4 | fast, recommended |
 
 
 
 
63
 
64
  Here is a handy graph by ikawrakow comparing some lower-quality quant
65
  types (lower is better):
 
53
  | Link | Type | Size/GB | Notes |
54
  |:-----|:-----|--------:|:------|
55
  | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.imatrix.gguf) | imatrix | 0.1 | imatrix file (for creating your own qwuants) |
56
+ | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.i1-IQ1_S.gguf) | i1-IQ1_S | 3.0 | for the desperate |
57
+ | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.i1-IQ1_M.gguf) | i1-IQ1_M | 3.3 | mostly desperate |
58
+ | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.i1-IQ2_XXS.gguf) | i1-IQ2_XXS | 3.6 | |
59
+ | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.i1-IQ2_XS.gguf) | i1-IQ2_XS | 3.9 | |
60
+ | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.i1-IQ2_S.gguf) | i1-IQ2_S | 4.1 | |
61
+ | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.i1-IQ2_M.gguf) | i1-IQ2_M | 4.4 | |
62
+ | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.i1-Q2_K_S.gguf) | i1-Q2_K_S | 4.5 | very low quality |
63
  | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.i1-Q2_K.gguf) | i1-Q2_K | 4.9 | IQ3_XXS probably better |
64
  | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 4.9 | lower quality |
65
+ | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.i1-IQ3_XS.gguf) | i1-IQ3_XS | 5.3 | |
66
+ | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.i1-IQ3_S.gguf) | i1-IQ3_S | 5.6 | beats Q3_K* |
67
+ | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.i1-Q3_K_S.gguf) | i1-Q3_K_S | 5.6 | IQ3_XS probably better |
68
  | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.i1-IQ3_M.gguf) | i1-IQ3_M | 5.8 | |
69
  | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.i1-Q3_K_M.gguf) | i1-Q3_K_M | 6.1 | IQ3_S probably better |
70
+ | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.i1-Q3_K_L.gguf) | i1-Q3_K_L | 6.6 | IQ3_M probably better |
71
+ | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.i1-IQ4_XS.gguf) | i1-IQ4_XS | 6.7 | |
72
  | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.i1-IQ4_NL.gguf) | i1-IQ4_NL | 7.0 | prefer IQ4_XS |
73
+ | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.i1-Q4_0.gguf) | i1-Q4_0 | 7.0 | fast, low quality |
74
  | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.i1-Q4_K_S.gguf) | i1-Q4_K_S | 7.0 | optimal size/speed/quality |
75
  | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.i1-Q4_K_M.gguf) | i1-Q4_K_M | 7.4 | fast, recommended |
76
+ | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.i1-Q4_1.gguf) | i1-Q4_1 | 7.7 | |
77
+ | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.i1-Q5_K_S.gguf) | i1-Q5_K_S | 8.3 | |
78
+ | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.i1-Q5_K_M.gguf) | i1-Q5_K_M | 8.5 | |
79
+ | [GGUF](https://huggingface.co/mradermacher/MamayLM-Gemma-3-12B-IT-v1.0-i1-GGUF/resolve/main/MamayLM-Gemma-3-12B-IT-v1.0.i1-Q6_K.gguf) | i1-Q6_K | 9.8 | practically like static Q6_K |
80
 
81
  Here is a handy graph by ikawrakow comparing some lower-quality quant
82
  types (lower is better):