Hugging Face's logo

mradermacher
/
commandr-35b-cpt-sft-i1-GGUF

Text Generation
Transformers
GGUF
English
causal-lm
two-stage-training
continual-pretraining
supervised-fine-tuning
synthetic-qa
lora
axolotl
deepspeed
commandr
cohere
eu-hpc
imatrix
conversational
Model card Files
xet
Community
mradermacher commited on
Commit
bcbe658
·
verified ·
1 Parent(s): cabf80e

auto-patch README.md

Browse files
Files changed (1) hide show
  1. README.md +10 -0
README.md CHANGED
@@ -60,13 +60,23 @@ more details, including on how to concatenate multi-part files.
60
  | Link | Type | Size/GB | Notes |
61
  |:-----|:-----|--------:|:------|
62
  | [GGUF](https://huggingface.co/mradermacher/commandr-35b-cpt-sft-i1-GGUF/resolve/main/commandr-35b-cpt-sft.imatrix.gguf) | imatrix | 0.1 | imatrix file (for creating your own qwuants) |
 
 
 
 
 
63
  | [GGUF](https://huggingface.co/mradermacher/commandr-35b-cpt-sft-i1-GGUF/resolve/main/commandr-35b-cpt-sft.i1-IQ2_M.gguf) | i1-IQ2_M | 12.8 | |
 
64
  | [GGUF](https://huggingface.co/mradermacher/commandr-35b-cpt-sft-i1-GGUF/resolve/main/commandr-35b-cpt-sft.i1-Q2_K.gguf) | i1-Q2_K | 13.9 | IQ3_XXS probably better |
65
  | [GGUF](https://huggingface.co/mradermacher/commandr-35b-cpt-sft-i1-GGUF/resolve/main/commandr-35b-cpt-sft.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 13.9 | lower quality |
 
66
  | [GGUF](https://huggingface.co/mradermacher/commandr-35b-cpt-sft-i1-GGUF/resolve/main/commandr-35b-cpt-sft.i1-IQ3_M.gguf) | i1-IQ3_M | 16.8 | |
67
  | [GGUF](https://huggingface.co/mradermacher/commandr-35b-cpt-sft-i1-GGUF/resolve/main/commandr-35b-cpt-sft.i1-Q3_K_M.gguf) | i1-Q3_K_M | 17.7 | IQ3_S probably better |
 
 
68
  | [GGUF](https://huggingface.co/mradermacher/commandr-35b-cpt-sft-i1-GGUF/resolve/main/commandr-35b-cpt-sft.i1-Q4_K_S.gguf) | i1-Q4_K_S | 20.5 | optimal size/speed/quality |
69
  | [GGUF](https://huggingface.co/mradermacher/commandr-35b-cpt-sft-i1-GGUF/resolve/main/commandr-35b-cpt-sft.i1-Q4_K_M.gguf) | i1-Q4_K_M | 21.6 | fast, recommended |
 
70
  | [GGUF](https://huggingface.co/mradermacher/commandr-35b-cpt-sft-i1-GGUF/resolve/main/commandr-35b-cpt-sft.i1-Q6_K.gguf) | i1-Q6_K | 28.8 | practically like static Q6_K |
71
 
72
  Here is a handy graph by ikawrakow comparing some lower-quality quant
 
60
  | Link | Type | Size/GB | Notes |
61
  |:-----|:-----|--------:|:------|
62
  | [GGUF](https://huggingface.co/mradermacher/commandr-35b-cpt-sft-i1-GGUF/resolve/main/commandr-35b-cpt-sft.imatrix.gguf) | imatrix | 0.1 | imatrix file (for creating your own qwuants) |
63
+ | [GGUF](https://huggingface.co/mradermacher/commandr-35b-cpt-sft-i1-GGUF/resolve/main/commandr-35b-cpt-sft.i1-IQ1_S.gguf) | i1-IQ1_S | 8.6 | for the desperate |
64
+ | [GGUF](https://huggingface.co/mradermacher/commandr-35b-cpt-sft-i1-GGUF/resolve/main/commandr-35b-cpt-sft.i1-IQ1_M.gguf) | i1-IQ1_M | 9.2 | mostly desperate |
65
+ | [GGUF](https://huggingface.co/mradermacher/commandr-35b-cpt-sft-i1-GGUF/resolve/main/commandr-35b-cpt-sft.i1-IQ2_XXS.gguf) | i1-IQ2_XXS | 10.3 | |
66
+ | [GGUF](https://huggingface.co/mradermacher/commandr-35b-cpt-sft-i1-GGUF/resolve/main/commandr-35b-cpt-sft.i1-IQ2_XS.gguf) | i1-IQ2_XS | 11.2 | |
67
+ | [GGUF](https://huggingface.co/mradermacher/commandr-35b-cpt-sft-i1-GGUF/resolve/main/commandr-35b-cpt-sft.i1-IQ2_S.gguf) | i1-IQ2_S | 11.9 | |
68
  | [GGUF](https://huggingface.co/mradermacher/commandr-35b-cpt-sft-i1-GGUF/resolve/main/commandr-35b-cpt-sft.i1-IQ2_M.gguf) | i1-IQ2_M | 12.8 | |
69
+ | [GGUF](https://huggingface.co/mradermacher/commandr-35b-cpt-sft-i1-GGUF/resolve/main/commandr-35b-cpt-sft.i1-Q2_K_S.gguf) | i1-Q2_K_S | 12.8 | very low quality |
70
  | [GGUF](https://huggingface.co/mradermacher/commandr-35b-cpt-sft-i1-GGUF/resolve/main/commandr-35b-cpt-sft.i1-Q2_K.gguf) | i1-Q2_K | 13.9 | IQ3_XXS probably better |
71
  | [GGUF](https://huggingface.co/mradermacher/commandr-35b-cpt-sft-i1-GGUF/resolve/main/commandr-35b-cpt-sft.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 13.9 | lower quality |
72
+ | [GGUF](https://huggingface.co/mradermacher/commandr-35b-cpt-sft-i1-GGUF/resolve/main/commandr-35b-cpt-sft.i1-Q3_K_S.gguf) | i1-Q3_K_S | 16.0 | IQ3_XS probably better |
73
  | [GGUF](https://huggingface.co/mradermacher/commandr-35b-cpt-sft-i1-GGUF/resolve/main/commandr-35b-cpt-sft.i1-IQ3_M.gguf) | i1-IQ3_M | 16.8 | |
74
  | [GGUF](https://huggingface.co/mradermacher/commandr-35b-cpt-sft-i1-GGUF/resolve/main/commandr-35b-cpt-sft.i1-Q3_K_M.gguf) | i1-Q3_K_M | 17.7 | IQ3_S probably better |
75
+ | [GGUF](https://huggingface.co/mradermacher/commandr-35b-cpt-sft-i1-GGUF/resolve/main/commandr-35b-cpt-sft.i1-Q3_K_L.gguf) | i1-Q3_K_L | 19.2 | IQ3_M probably better |
76
+ | [GGUF](https://huggingface.co/mradermacher/commandr-35b-cpt-sft-i1-GGUF/resolve/main/commandr-35b-cpt-sft.i1-IQ4_XS.gguf) | i1-IQ4_XS | 19.3 | |
77
  | [GGUF](https://huggingface.co/mradermacher/commandr-35b-cpt-sft-i1-GGUF/resolve/main/commandr-35b-cpt-sft.i1-Q4_K_S.gguf) | i1-Q4_K_S | 20.5 | optimal size/speed/quality |
78
  | [GGUF](https://huggingface.co/mradermacher/commandr-35b-cpt-sft-i1-GGUF/resolve/main/commandr-35b-cpt-sft.i1-Q4_K_M.gguf) | i1-Q4_K_M | 21.6 | fast, recommended |
79
+ | [GGUF](https://huggingface.co/mradermacher/commandr-35b-cpt-sft-i1-GGUF/resolve/main/commandr-35b-cpt-sft.i1-Q5_K_S.gguf) | i1-Q5_K_S | 24.4 | |
80
  | [GGUF](https://huggingface.co/mradermacher/commandr-35b-cpt-sft-i1-GGUF/resolve/main/commandr-35b-cpt-sft.i1-Q6_K.gguf) | i1-Q6_K | 28.8 | practically like static Q6_K |
81
 
82
  Here is a handy graph by ikawrakow comparing some lower-quality quant