Transformers
GGUF
imatrix
mradermacher commited on
Commit
8677442
·
verified ·
1 Parent(s): c15bb7c

auto-patch README.md

Browse files
Files changed (1) hide show
  1. README.md +3 -0
README.md CHANGED
@@ -75,8 +75,11 @@ more details, including on how to concatenate multi-part files.
75
  | Link | Type | Size/GB | Notes |
76
  |:-----|:-----|--------:|:------|
77
  | [PART 1](https://huggingface.co/mradermacher/bloom-i1-GGUF/resolve/main/bloom.i1-Q2_K.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/bloom-i1-GGUF/resolve/main/bloom.i1-Q2_K.gguf.part2of2) | i1-Q2_K | 68.2 | IQ3_XXS probably better |
 
78
  | [PART 1](https://huggingface.co/mradermacher/bloom-i1-GGUF/resolve/main/bloom.i1-IQ3_M.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/bloom-i1-GGUF/resolve/main/bloom.i1-IQ3_M.gguf.part2of2) | i1-IQ3_M | 87.4 | |
 
79
  | [PART 1](https://huggingface.co/mradermacher/bloom-i1-GGUF/resolve/main/bloom.i1-Q4_K_S.gguf.part1of3) [PART 2](https://huggingface.co/mradermacher/bloom-i1-GGUF/resolve/main/bloom.i1-Q4_K_S.gguf.part2of3) [PART 3](https://huggingface.co/mradermacher/bloom-i1-GGUF/resolve/main/bloom.i1-Q4_K_S.gguf.part3of3) | i1-Q4_K_S | 103.1 | optimal size/speed/quality |
 
80
 
81
  Here is a handy graph by ikawrakow comparing some lower-quality quant
82
  types (lower is better):
 
75
  | Link | Type | Size/GB | Notes |
76
  |:-----|:-----|--------:|:------|
77
  | [PART 1](https://huggingface.co/mradermacher/bloom-i1-GGUF/resolve/main/bloom.i1-Q2_K.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/bloom-i1-GGUF/resolve/main/bloom.i1-Q2_K.gguf.part2of2) | i1-Q2_K | 68.2 | IQ3_XXS probably better |
78
+ | [PART 1](https://huggingface.co/mradermacher/bloom-i1-GGUF/resolve/main/bloom.i1-IQ3_XXS.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/bloom-i1-GGUF/resolve/main/bloom.i1-IQ3_XXS.gguf.part2of2) | i1-IQ3_XXS | 70.9 | lower quality |
79
  | [PART 1](https://huggingface.co/mradermacher/bloom-i1-GGUF/resolve/main/bloom.i1-IQ3_M.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/bloom-i1-GGUF/resolve/main/bloom.i1-IQ3_M.gguf.part2of2) | i1-IQ3_M | 87.4 | |
80
+ | [PART 1](https://huggingface.co/mradermacher/bloom-i1-GGUF/resolve/main/bloom.i1-Q3_K_M.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/bloom-i1-GGUF/resolve/main/bloom.i1-Q3_K_M.gguf.part2of2) | i1-Q3_K_M | 94.5 | IQ3_S probably better |
81
  | [PART 1](https://huggingface.co/mradermacher/bloom-i1-GGUF/resolve/main/bloom.i1-Q4_K_S.gguf.part1of3) [PART 2](https://huggingface.co/mradermacher/bloom-i1-GGUF/resolve/main/bloom.i1-Q4_K_S.gguf.part2of3) [PART 3](https://huggingface.co/mradermacher/bloom-i1-GGUF/resolve/main/bloom.i1-Q4_K_S.gguf.part3of3) | i1-Q4_K_S | 103.1 | optimal size/speed/quality |
82
+ | [PART 1](https://huggingface.co/mradermacher/bloom-i1-GGUF/resolve/main/bloom.i1-Q4_K_M.gguf.part1of3) [PART 2](https://huggingface.co/mradermacher/bloom-i1-GGUF/resolve/main/bloom.i1-Q4_K_M.gguf.part2of3) [PART 3](https://huggingface.co/mradermacher/bloom-i1-GGUF/resolve/main/bloom.i1-Q4_K_M.gguf.part3of3) | i1-Q4_K_M | 114.8 | fast, recommended |
83
 
84
  Here is a handy graph by ikawrakow comparing some lower-quality quant
85
  types (lower is better):