Hugging Face's logo

RichardErkhov
/
nbeerbower_-_Llama3.1-Gutenberg-Doppel-70B-gguf

GGUF
conversational
Model card Files
xet
Community
RichardErkhov commited on
Commit
acc0e70
·
verified ·
1 Parent(s): 0a79b6d

uploaded readme

Browse files
Files changed (1) hide show
  1. README.md +173 -0
README.md ADDED
@@ -0,0 +1,173 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ Quantization made by Richard Erkhov.
2
+
3
+ [Github](https://github.com/RichardErkhov)
4
+
5
+ [Discord](https://discord.gg/pvy7H8DZMG)
6
+
7
+ [Request more models](https://github.com/RichardErkhov/quant_request)
8
+
9
+
10
+ Llama3.1-Gutenberg-Doppel-70B - GGUF
11
+ - Model creator: https://huggingface.co/nbeerbower/
12
+ - Original model: https://huggingface.co/nbeerbower/Llama3.1-Gutenberg-Doppel-70B/
13
+
14
+
15
+ | Name | Quant method | Size |
16
+ | ---- | ---- | ---- |
17
+ | [Llama3.1-Gutenberg-Doppel-70B.Q2_K.gguf](https://huggingface.co/RichardErkhov/nbeerbower_-_Llama3.1-Gutenberg-Doppel-70B-gguf/blob/main/Llama3.1-Gutenberg-Doppel-70B.Q2_K.gguf) | Q2_K | 24.56GB |
18
+ | [Llama3.1-Gutenberg-Doppel-70B.IQ3_XS.gguf](https://huggingface.co/RichardErkhov/nbeerbower_-_Llama3.1-Gutenberg-Doppel-70B-gguf/blob/main/Llama3.1-Gutenberg-Doppel-70B.IQ3_XS.gguf) | IQ3_XS | 27.29GB |
19
+ | [Llama3.1-Gutenberg-Doppel-70B.IQ3_S.gguf](https://huggingface.co/RichardErkhov/nbeerbower_-_Llama3.1-Gutenberg-Doppel-70B-gguf/blob/main/Llama3.1-Gutenberg-Doppel-70B.IQ3_S.gguf) | IQ3_S | 28.79GB |
20
+ | [Llama3.1-Gutenberg-Doppel-70B.Q3_K_S.gguf](https://huggingface.co/RichardErkhov/nbeerbower_-_Llama3.1-Gutenberg-Doppel-70B-gguf/blob/main/Llama3.1-Gutenberg-Doppel-70B.Q3_K_S.gguf) | Q3_K_S | 28.79GB |
21
+ | [Llama3.1-Gutenberg-Doppel-70B.IQ3_M.gguf](https://huggingface.co/RichardErkhov/nbeerbower_-_Llama3.1-Gutenberg-Doppel-70B-gguf/blob/main/Llama3.1-Gutenberg-Doppel-70B.IQ3_M.gguf) | IQ3_M | 29.74GB |
22
+ | [Llama3.1-Gutenberg-Doppel-70B.Q3_K.gguf](https://huggingface.co/RichardErkhov/nbeerbower_-_Llama3.1-Gutenberg-Doppel-70B-gguf/blob/main/Llama3.1-Gutenberg-Doppel-70B.Q3_K.gguf) | Q3_K | 31.91GB |
23
+ | [Llama3.1-Gutenberg-Doppel-70B.Q3_K_M.gguf](https://huggingface.co/RichardErkhov/nbeerbower_-_Llama3.1-Gutenberg-Doppel-70B-gguf/blob/main/Llama3.1-Gutenberg-Doppel-70B.Q3_K_M.gguf) | Q3_K_M | 31.91GB |
24
+ | [Llama3.1-Gutenberg-Doppel-70B.Q3_K_L.gguf](https://huggingface.co/RichardErkhov/nbeerbower_-_Llama3.1-Gutenberg-Doppel-70B-gguf/blob/main/Llama3.1-Gutenberg-Doppel-70B.Q3_K_L.gguf) | Q3_K_L | 34.59GB |
25
+ | [Llama3.1-Gutenberg-Doppel-70B.IQ4_XS.gguf](https://huggingface.co/RichardErkhov/nbeerbower_-_Llama3.1-Gutenberg-Doppel-70B-gguf/blob/main/Llama3.1-Gutenberg-Doppel-70B.IQ4_XS.gguf) | IQ4_XS | 35.64GB |
26
+ | [Llama3.1-Gutenberg-Doppel-70B.Q4_0.gguf](https://huggingface.co/RichardErkhov/nbeerbower_-_Llama3.1-Gutenberg-Doppel-70B-gguf/blob/main/Llama3.1-Gutenberg-Doppel-70B.Q4_0.gguf) | Q4_0 | 37.22GB |
27
+ | [Llama3.1-Gutenberg-Doppel-70B.IQ4_NL.gguf](https://huggingface.co/RichardErkhov/nbeerbower_-_Llama3.1-Gutenberg-Doppel-70B-gguf/tree/main/) | IQ4_NL | 37.58GB |
28
+ | [Llama3.1-Gutenberg-Doppel-70B.Q4_K_S.gguf](https://huggingface.co/RichardErkhov/nbeerbower_-_Llama3.1-Gutenberg-Doppel-70B-gguf/tree/main/) | Q4_K_S | 37.58GB |
29
+ | [Llama3.1-Gutenberg-Doppel-70B.Q4_K.gguf](https://huggingface.co/RichardErkhov/nbeerbower_-_Llama3.1-Gutenberg-Doppel-70B-gguf/tree/main/) | Q4_K | 39.6GB |
30
+ | [Llama3.1-Gutenberg-Doppel-70B.Q4_K_M.gguf](https://huggingface.co/RichardErkhov/nbeerbower_-_Llama3.1-Gutenberg-Doppel-70B-gguf/tree/main/) | Q4_K_M | 39.6GB |
31
+ | [Llama3.1-Gutenberg-Doppel-70B.Q4_1.gguf](https://huggingface.co/RichardErkhov/nbeerbower_-_Llama3.1-Gutenberg-Doppel-70B-gguf/tree/main/) | Q4_1 | 41.27GB |
32
+ | [Llama3.1-Gutenberg-Doppel-70B.Q5_0.gguf](https://huggingface.co/RichardErkhov/nbeerbower_-_Llama3.1-Gutenberg-Doppel-70B-gguf/tree/main/) | Q5_0 | 45.32GB |
33
+ | [Llama3.1-Gutenberg-Doppel-70B.Q5_K_S.gguf](https://huggingface.co/RichardErkhov/nbeerbower_-_Llama3.1-Gutenberg-Doppel-70B-gguf/tree/main/) | Q5_K_S | 45.32GB |
34
+ | [Llama3.1-Gutenberg-Doppel-70B.Q5_K.gguf](https://huggingface.co/RichardErkhov/nbeerbower_-_Llama3.1-Gutenberg-Doppel-70B-gguf/tree/main/) | Q5_K | 46.52GB |
35
+ | [Llama3.1-Gutenberg-Doppel-70B.Q5_K_M.gguf](https://huggingface.co/RichardErkhov/nbeerbower_-_Llama3.1-Gutenberg-Doppel-70B-gguf/tree/main/) | Q5_K_M | 46.52GB |
36
+ | [Llama3.1-Gutenberg-Doppel-70B.Q5_1.gguf](https://huggingface.co/RichardErkhov/nbeerbower_-_Llama3.1-Gutenberg-Doppel-70B-gguf/tree/main/) | Q5_1 | 49.36GB |
37
+ | [Llama3.1-Gutenberg-Doppel-70B.Q6_K.gguf](https://huggingface.co/RichardErkhov/nbeerbower_-_Llama3.1-Gutenberg-Doppel-70B-gguf/tree/main/) | Q6_K | 53.91GB |
38
+ | [Llama3.1-Gutenberg-Doppel-70B.Q8_0.gguf](https://huggingface.co/RichardErkhov/nbeerbower_-_Llama3.1-Gutenberg-Doppel-70B-gguf/tree/main/) | Q8_0 | 69.83GB |
39
+
40
+
41
+
42
+
43
+ Original model description:
44
+ ---
45
+ license: llama3.1
46
+ library_name: transformers
47
+ base_model:
48
+ - mlabonne/Hermes-3-Llama-3.1-70B-lorablated
49
+ datasets:
50
+ - jondurbin/gutenberg-dpo-v0.1
51
+ - nbeerbower/gutenberg2-dpo
52
+ model-index:
53
+ - name: Llama3.1-Gutenberg-Doppel-70B
54
+ results:
55
+ - task:
56
+ type: text-generation
57
+ name: Text Generation
58
+ dataset:
59
+ name: IFEval (0-Shot)
60
+ type: HuggingFaceH4/ifeval
61
+ args:
62
+ num_few_shot: 0
63
+ metrics:
64
+ - type: inst_level_strict_acc and prompt_level_strict_acc
65
+ value: 70.92
66
+ name: strict accuracy
67
+ source:
68
+ url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=nbeerbower/Llama3.1-Gutenberg-Doppel-70B
69
+ name: Open LLM Leaderboard
70
+ - task:
71
+ type: text-generation
72
+ name: Text Generation
73
+ dataset:
74
+ name: BBH (3-Shot)
75
+ type: BBH
76
+ args:
77
+ num_few_shot: 3
78
+ metrics:
79
+ - type: acc_norm
80
+ value: 52.56
81
+ name: normalized accuracy
82
+ source:
83
+ url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=nbeerbower/Llama3.1-Gutenberg-Doppel-70B
84
+ name: Open LLM Leaderboard
85
+ - task:
86
+ type: text-generation
87
+ name: Text Generation
88
+ dataset:
89
+ name: MATH Lvl 5 (4-Shot)
90
+ type: hendrycks/competition_math
91
+ args:
92
+ num_few_shot: 4
93
+ metrics:
94
+ - type: exact_match
95
+ value: 13.75
96
+ name: exact match
97
+ source:
98
+ url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=nbeerbower/Llama3.1-Gutenberg-Doppel-70B
99
+ name: Open LLM Leaderboard
100
+ - task:
101
+ type: text-generation
102
+ name: Text Generation
103
+ dataset:
104
+ name: GPQA (0-shot)
105
+ type: Idavidrein/gpqa
106
+ args:
107
+ num_few_shot: 0
108
+ metrics:
109
+ - type: acc_norm
110
+ value: 12.64
111
+ name: acc_norm
112
+ source:
113
+ url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=nbeerbower/Llama3.1-Gutenberg-Doppel-70B
114
+ name: Open LLM Leaderboard
115
+ - task:
116
+ type: text-generation
117
+ name: Text Generation
118
+ dataset:
119
+ name: MuSR (0-shot)
120
+ type: TAUR-Lab/MuSR
121
+ args:
122
+ num_few_shot: 0
123
+ metrics:
124
+ - type: acc_norm
125
+ value: 22.68
126
+ name: acc_norm
127
+ source:
128
+ url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=nbeerbower/Llama3.1-Gutenberg-Doppel-70B
129
+ name: Open LLM Leaderboard
130
+ - task:
131
+ type: text-generation
132
+ name: Text Generation
133
+ dataset:
134
+ name: MMLU-PRO (5-shot)
135
+ type: TIGER-Lab/MMLU-Pro
136
+ config: main
137
+ split: test
138
+ args:
139
+ num_few_shot: 5
140
+ metrics:
141
+ - type: acc
142
+ value: 41.52
143
+ name: accuracy
144
+ source:
145
+ url: https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=nbeerbower/Llama3.1-Gutenberg-Doppel-70B
146
+ name: Open LLM Leaderboard
147
+ ---
148
+ ![image/png](https://huggingface.co/nbeerbower/Mistral-Small-Gutenberg-Doppel-22B/resolve/main/doppel-header?download=true)
149
+
150
+ # Llama3.1-Gutenberg-Doppel-70B
151
+
152
+ [mlabonne/Hermes-3-Llama-3.1-70B-lorablated](https://huggingface.co/mlabonne/Hermes-3-Llama-3.1-70B-lorablated) finetuned on [jondurbin/gutenberg-dpo-v0.1](https://huggingface.co/datasets/jondurbin/gutenberg-dpo-v0.1) and [nbeerbower/gutenberg2-dpo](https://huggingface.co/datasets/nbeerbower/gutenberg2-dpo).
153
+
154
+ ### Method
155
+
156
+ [ORPO tuned](https://mlabonne.github.io/blog/posts/2024-04-19_Fine_tune_Llama_3_with_ORPO.html) with 2x H100 for 3 epochs.
157
+
158
+ Thank you [Schneewolf Labs](https://schneewolflabs.com/) for the compute.
159
+ # [Open LLM Leaderboard Evaluation Results](https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard)
160
+ Detailed results can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/details_nbeerbower__Llama3.1-Gutenberg-Doppel-70B)
161
+
162
+ | Metric |Value|
163
+ |-------------------|----:|
164
+ |Avg. |35.68|
165
+ |IFEval (0-Shot) |70.92|
166
+ |BBH (3-Shot) |52.56|
167
+ |MATH Lvl 5 (4-Shot)|13.75|
168
+ |GPQA (0-shot) |12.64|
169
+ |MuSR (0-shot) |22.68|
170
+ |MMLU-PRO (5-shot) |41.52|
171
+
172
+
173
+