mradermacher commited on
Commit
b5a5417
·
verified ·
1 Parent(s): 20b0ad6

auto-patch README.md

Browse files
Files changed (1) hide show
  1. README.md +7 -0
README.md CHANGED
@@ -222,6 +222,7 @@ more details, including on how to concatenate multi-part files.
222
  | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.imatrix.gguf) | imatrix | 0.1 | imatrix file (for creating your own qwuants) |
223
  | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-IQ1_S.gguf) | i1-IQ1_S | 2.1 | for the desperate |
224
  | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-IQ1_M.gguf) | i1-IQ1_M | 2.3 | mostly desperate |
 
225
  | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-IQ2_XS.gguf) | i1-IQ2_XS | 2.7 | |
226
  | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-IQ2_S.gguf) | i1-IQ2_S | 2.9 | |
227
  | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-Q2_K_S.gguf) | i1-Q2_K_S | 3.1 | very low quality |
@@ -229,11 +230,17 @@ more details, including on how to concatenate multi-part files.
229
  | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 3.4 | lower quality |
230
  | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-IQ3_XS.gguf) | i1-IQ3_XS | 3.6 | |
231
  | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-Q3_K_S.gguf) | i1-Q3_K_S | 3.8 | IQ3_XS probably better |
 
232
  | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-IQ3_M.gguf) | i1-IQ3_M | 3.9 | |
233
  | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-Q3_K_M.gguf) | i1-Q3_K_M | 4.1 | IQ3_S probably better |
 
 
 
234
  | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-Q4_K_S.gguf) | i1-Q4_K_S | 4.8 | optimal size/speed/quality |
 
235
  | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-Q4_1.gguf) | i1-Q4_1 | 5.2 | |
236
  | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-Q5_K_S.gguf) | i1-Q5_K_S | 5.7 | |
 
237
  | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-Q6_K.gguf) | i1-Q6_K | 6.7 | practically like static Q6_K |
238
 
239
  Here is a handy graph by ikawrakow comparing some lower-quality quant
 
222
  | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.imatrix.gguf) | imatrix | 0.1 | imatrix file (for creating your own qwuants) |
223
  | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-IQ1_S.gguf) | i1-IQ1_S | 2.1 | for the desperate |
224
  | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-IQ1_M.gguf) | i1-IQ1_M | 2.3 | mostly desperate |
225
+ | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-IQ2_XXS.gguf) | i1-IQ2_XXS | 2.5 | |
226
  | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-IQ2_XS.gguf) | i1-IQ2_XS | 2.7 | |
227
  | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-IQ2_S.gguf) | i1-IQ2_S | 2.9 | |
228
  | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-Q2_K_S.gguf) | i1-Q2_K_S | 3.1 | very low quality |
 
230
  | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 3.4 | lower quality |
231
  | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-IQ3_XS.gguf) | i1-IQ3_XS | 3.6 | |
232
  | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-Q3_K_S.gguf) | i1-Q3_K_S | 3.8 | IQ3_XS probably better |
233
+ | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-IQ3_S.gguf) | i1-IQ3_S | 3.8 | beats Q3_K* |
234
  | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-IQ3_M.gguf) | i1-IQ3_M | 3.9 | |
235
  | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-Q3_K_M.gguf) | i1-Q3_K_M | 4.1 | IQ3_S probably better |
236
+ | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-Q3_K_L.gguf) | i1-Q3_K_L | 4.4 | IQ3_M probably better |
237
+ | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-IQ4_XS.gguf) | i1-IQ4_XS | 4.5 | |
238
+ | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-IQ4_NL.gguf) | i1-IQ4_NL | 4.8 | prefer IQ4_XS |
239
  | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-Q4_K_S.gguf) | i1-Q4_K_S | 4.8 | optimal size/speed/quality |
240
+ | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-Q4_K_M.gguf) | i1-Q4_K_M | 5.0 | fast, recommended |
241
  | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-Q4_1.gguf) | i1-Q4_1 | 5.2 | |
242
  | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-Q5_K_S.gguf) | i1-Q5_K_S | 5.7 | |
243
+ | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-Q5_K_M.gguf) | i1-Q5_K_M | 5.8 | |
244
  | [GGUF](https://huggingface.co/mradermacher/Llama-3.1-8B-Instruct-heretic-i1-GGUF/resolve/main/Llama-3.1-8B-Instruct-heretic.i1-Q6_K.gguf) | i1-Q6_K | 6.7 | practically like static Q6_K |
245
 
246
  Here is a handy graph by ikawrakow comparing some lower-quality quant