Trappu commited on
Commit
b208f48
1 Parent(s): 5ddd5da

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +19 -14
README.md CHANGED
@@ -3,11 +3,11 @@ tags:
3
  - merge
4
  - mergekit
5
  - lazymergekit
6
- - Trappu/Nemo-Picaro-fixed
7
  - anthracite-org/magnum-v2-12b
 
8
  base_model:
9
- - Trappu/Nemo-Picaro-fixed
10
  - anthracite-org/magnum-v2-12b
 
11
  model-index:
12
  - name: Magnum-Picaro-0.7-v2-12b
13
  results:
@@ -24,7 +24,8 @@ model-index:
24
  value: 30.03
25
  name: strict accuracy
26
  source:
27
- url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=Trappu/Magnum-Picaro-0.7-v2-12b
 
28
  name: Open LLM Leaderboard
29
  - task:
30
  type: text-generation
@@ -39,7 +40,8 @@ model-index:
39
  value: 35.75
40
  name: normalized accuracy
41
  source:
42
- url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=Trappu/Magnum-Picaro-0.7-v2-12b
 
43
  name: Open LLM Leaderboard
44
  - task:
45
  type: text-generation
@@ -54,7 +56,8 @@ model-index:
54
  value: 4.76
55
  name: exact match
56
  source:
57
- url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=Trappu/Magnum-Picaro-0.7-v2-12b
 
58
  name: Open LLM Leaderboard
59
  - task:
60
  type: text-generation
@@ -69,7 +72,8 @@ model-index:
69
  value: 9.73
70
  name: acc_norm
71
  source:
72
- url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=Trappu/Magnum-Picaro-0.7-v2-12b
 
73
  name: Open LLM Leaderboard
74
  - task:
75
  type: text-generation
@@ -84,7 +88,8 @@ model-index:
84
  value: 19.56
85
  name: acc_norm
86
  source:
87
- url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=Trappu/Magnum-Picaro-0.7-v2-12b
 
88
  name: Open LLM Leaderboard
89
  - task:
90
  type: text-generation
@@ -101,7 +106,8 @@ model-index:
101
  value: 28.67
102
  name: accuracy
103
  source:
104
- url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=Trappu/Magnum-Picaro-0.7-v2-12b
 
105
  name: Open LLM Leaderboard
106
  license: apache-2.0
107
  pipeline_tag: text-generation
@@ -110,7 +116,7 @@ library_name: transformers
110
 
111
  # Details
112
 
113
- This model is a merge between [Trappu/Nemo-Picaro-fixed](https://huggingface.co/Trappu/Nemo-Picaro-fixed), a model trained on my own little dataset free of synthetic data, which focuses solely on storywriting and scenrio prompting (Example: `[ Scenario: bla bla bla; Tags: bla bla bla ]`), and [anthracite-org/magnum-v2-12b](https://huggingface.co/anthracite-org/magnum-v2-12b).
114
 
115
  The reason why I decided to merge it with Magnum (and don't recommend Picaro alone) is because that model, aside from its obvious flaws (rampant impersonation, stupid, etc...), is a one-trick pony and will be really rough for the average LLM user to handle. The idea was to have Magnum work as some sort of stabilizer to fix the issues that emerge from the lack of multiturn/smart data in Picaro's dataset. It worked, I think. I enjoy the outputs and it's smart enough to work with.
116
 
@@ -159,14 +165,14 @@ Static: https://huggingface.co/mradermacher/Magnum-Picaro-0.7-v2-12b-GGUF
159
  # Magnum-Picaro-0.7-v2-12b
160
 
161
  Magnum-Picaro-0.7-v2-12b is a merge of the following models using [LazyMergekit](https://colab.research.google.com/drive/1obulZ1ROXHjYLn6PPZJwRR6GzgQogxxb?usp=sharing):
162
- * [Trappu/Nemo-Picaro-fixed](https://huggingface.co/Trappu/Nemo-Picaro-fixed)
163
  * [anthracite-org/magnum-v2-12b](https://huggingface.co/anthracite-org/magnum-v2-12b)
164
 
165
  ## 🧩 Configuration
166
 
167
  ```yaml
168
  models:
169
- - model: Trappu/Nemo-Picaro-fixed
170
  parameters:
171
  density: 0.7
172
  weight: 0.5
@@ -176,7 +182,7 @@ models:
176
  weight: 0.5
177
 
178
  merge_method: ties
179
- base_model: Trappu/Nemo-Picaro-fixed
180
  parameters:
181
  normalize: true
182
  int8_mask: true
@@ -219,5 +225,4 @@ Detailed results can be found [here](https://huggingface.co/datasets/open-llm-le
219
  |MATH Lvl 5 (4-Shot)| 4.76|
220
  |GPQA (0-shot) | 9.73|
221
  |MuSR (0-shot) |19.56|
222
- |MMLU-PRO (5-shot) |28.67|
223
-
 
3
  - merge
4
  - mergekit
5
  - lazymergekit
 
6
  - anthracite-org/magnum-v2-12b
7
+ - Trappu/Nemo-Picaro-12B
8
  base_model:
 
9
  - anthracite-org/magnum-v2-12b
10
+ - Trappu/Nemo-Picaro-12B
11
  model-index:
12
  - name: Magnum-Picaro-0.7-v2-12b
13
  results:
 
24
  value: 30.03
25
  name: strict accuracy
26
  source:
27
+ url: >-
28
+ https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=Trappu/Magnum-Picaro-0.7-v2-12b
29
  name: Open LLM Leaderboard
30
  - task:
31
  type: text-generation
 
40
  value: 35.75
41
  name: normalized accuracy
42
  source:
43
+ url: >-
44
+ https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=Trappu/Magnum-Picaro-0.7-v2-12b
45
  name: Open LLM Leaderboard
46
  - task:
47
  type: text-generation
 
56
  value: 4.76
57
  name: exact match
58
  source:
59
+ url: >-
60
+ https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=Trappu/Magnum-Picaro-0.7-v2-12b
61
  name: Open LLM Leaderboard
62
  - task:
63
  type: text-generation
 
72
  value: 9.73
73
  name: acc_norm
74
  source:
75
+ url: >-
76
+ https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=Trappu/Magnum-Picaro-0.7-v2-12b
77
  name: Open LLM Leaderboard
78
  - task:
79
  type: text-generation
 
88
  value: 19.56
89
  name: acc_norm
90
  source:
91
+ url: >-
92
+ https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=Trappu/Magnum-Picaro-0.7-v2-12b
93
  name: Open LLM Leaderboard
94
  - task:
95
  type: text-generation
 
106
  value: 28.67
107
  name: accuracy
108
  source:
109
+ url: >-
110
+ https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=Trappu/Magnum-Picaro-0.7-v2-12b
111
  name: Open LLM Leaderboard
112
  license: apache-2.0
113
  pipeline_tag: text-generation
 
116
 
117
  # Details
118
 
119
+ This model is a merge between [Trappu/Nemo-Picaro-12B](https://huggingface.co/Trappu/Nemo-Picaro-12B), a model trained on my own little dataset free of synthetic data, which focuses solely on storywriting and scenrio prompting (Example: `[ Scenario: bla bla bla; Tags: bla bla bla ]`), and [anthracite-org/magnum-v2-12b](https://huggingface.co/anthracite-org/magnum-v2-12b).
120
 
121
  The reason why I decided to merge it with Magnum (and don't recommend Picaro alone) is because that model, aside from its obvious flaws (rampant impersonation, stupid, etc...), is a one-trick pony and will be really rough for the average LLM user to handle. The idea was to have Magnum work as some sort of stabilizer to fix the issues that emerge from the lack of multiturn/smart data in Picaro's dataset. It worked, I think. I enjoy the outputs and it's smart enough to work with.
122
 
 
165
  # Magnum-Picaro-0.7-v2-12b
166
 
167
  Magnum-Picaro-0.7-v2-12b is a merge of the following models using [LazyMergekit](https://colab.research.google.com/drive/1obulZ1ROXHjYLn6PPZJwRR6GzgQogxxb?usp=sharing):
168
+ * [Trappu/Nemo-Picaro-12B](https://huggingface.co/Trappu/Nemo-Picaro-12B)
169
  * [anthracite-org/magnum-v2-12b](https://huggingface.co/anthracite-org/magnum-v2-12b)
170
 
171
  ## 🧩 Configuration
172
 
173
  ```yaml
174
  models:
175
+ - model: Trappu/Nemo-Picaro-12B
176
  parameters:
177
  density: 0.7
178
  weight: 0.5
 
182
  weight: 0.5
183
 
184
  merge_method: ties
185
+ base_model: Trappu/Nemo-Picaro-12B
186
  parameters:
187
  normalize: true
188
  int8_mask: true
 
225
  |MATH Lvl 5 (4-Shot)| 4.76|
226
  |GPQA (0-shot) | 9.73|
227
  |MuSR (0-shot) |19.56|
228
+ |MMLU-PRO (5-shot) |28.67|