Triangle104 commited on
Commit
190690d
1 Parent(s): d10fc09

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +52 -0
README.md CHANGED
@@ -107,6 +107,58 @@ model-index:
107
  This model was converted to GGUF format from [`lemon07r/Gemma-2-Ataraxy-v2-9B`](https://huggingface.co/lemon07r/Gemma-2-Ataraxy-v2-9B) using llama.cpp via the ggml.ai's [GGUF-my-repo](https://huggingface.co/spaces/ggml-org/gguf-my-repo) space.
108
  Refer to the [original model card](https://huggingface.co/lemon07r/Gemma-2-Ataraxy-v2-9B) for more details on the model.
109
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
110
  ## Use with llama.cpp
111
  Install llama.cpp through brew (works on Mac and Linux)
112
 
 
107
  This model was converted to GGUF format from [`lemon07r/Gemma-2-Ataraxy-v2-9B`](https://huggingface.co/lemon07r/Gemma-2-Ataraxy-v2-9B) using llama.cpp via the ggml.ai's [GGUF-my-repo](https://huggingface.co/spaces/ggml-org/gguf-my-repo) space.
108
  Refer to the [original model card](https://huggingface.co/lemon07r/Gemma-2-Ataraxy-v2-9B) for more details on the model.
109
 
110
+ ---
111
+ Model details:
112
+ -
113
+
114
+ Gemma 2 Ataraxy v2 9B
115
+
116
+ Finally, after much testing, a sucessor to the first Gemma 2 Ataraxy 9B. Same kind of recipe, using the same principles, same concept as the last Ataraxy. It's not quite a better overall model, v1 is more well rounded, v2 is a little better at writing but has a little more slop and some other issues. consider this a sidegrade.
117
+
118
+ Ataraxy GGUF / EXL2 Quants
119
+
120
+ Bartowski quants (imatrix): https://huggingface.co/bartowski/Gemma-2-Ataraxy-v2-9B-GGUF
121
+
122
+ Mradermacher quants (static): https://huggingface.co/mradermacher/Gemma-2-Ataraxy-v2-9B-GGUF
123
+
124
+ Mradermacher quants (imatrix): https://huggingface.co/mradermacher/Gemma-2-Ataraxy-v2-9B-i1-GGUF
125
+
126
+ Bartowski and mradermacher use different calibration data for their imatrix quants I believe, and the static quant of course uses none. Pick your poison.
127
+
128
+ More coming soon. Format
129
+
130
+ Use Gemma 2 format. Merge Details Merge Method
131
+
132
+ This model was merged using the SLERP merge method. Models Merged
133
+
134
+ This is a merge of pre-trained language models created using mergekit.
135
+
136
+ The following models were included in the merge:
137
+
138
+ ifable/gemma-2-Ifable-9B
139
+ jsgreenawalt/gemma-2-9B-it-advanced-v2.1
140
+
141
+ Configuration
142
+
143
+ The following YAML configuration was used to produce this model:
144
+
145
+ base_model: ifable/gemma-2-Ifable-9B dtype: bfloat16 merge_method: slerp parameters: t:
146
+
147
+ filter: self_attn value: [0.0, 0.5, 0.3, 0.7, 1.0]
148
+ filter: mlp value: [1.0, 0.5, 0.7, 0.3, 0.0]
149
+ value: 0.5 slices:
150
+ sources:
151
+ layer_range: [0, 42] model: jsgreenawalt/gemma-2-9B-it-advanced-v2.1
152
+ layer_range: [0, 42] model: ifable/gemma-2-Ifable-9B
153
+
154
+ Open LLM Leaderboard Evaluation Results
155
+
156
+ Detailed results can be found here Metric Value Avg. 19.16 IFEval (0-Shot) 21.36 BBH (3-Shot) 39.80 MATH Lvl 5 (4-Shot) 0.83 GPQA (0-shot) 12.30 MuSR (0-shot) 4.88 MMLU-PRO (5-shot) 35.79
157
+
158
+ Second highest ranked open weight model in EQ-Bench.
159
+
160
+ ---
161
+
162
  ## Use with llama.cpp
163
  Install llama.cpp through brew (works on Mac and Linux)
164