Upload README.md with huggingface_hub
Browse files
README.md
ADDED
@@ -0,0 +1,171 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
base_model: M4-ai/NeuralReyna-Mini-1.8B-v0.3
|
3 |
+
datasets:
|
4 |
+
- argilla/OpenHermes2.5-dpo-binarized-alpha
|
5 |
+
- Locutusque/Hercules-v3.0
|
6 |
+
inference: false
|
7 |
+
language:
|
8 |
+
- en
|
9 |
+
license: apache-2.0
|
10 |
+
model-index:
|
11 |
+
- name: NeuralReyna-Mini-1.8B-v0.3
|
12 |
+
results:
|
13 |
+
- dataset:
|
14 |
+
args:
|
15 |
+
num_few_shot: 25
|
16 |
+
config: ARC-Challenge
|
17 |
+
name: AI2 Reasoning Challenge (25-Shot)
|
18 |
+
split: test
|
19 |
+
type: ai2_arc
|
20 |
+
metrics:
|
21 |
+
- name: normalized accuracy
|
22 |
+
type: acc_norm
|
23 |
+
value: 35.58
|
24 |
+
source:
|
25 |
+
name: Open LLM Leaderboard
|
26 |
+
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=M4-ai/NeuralReyna-Mini-1.8B-v0.3
|
27 |
+
task:
|
28 |
+
name: Text Generation
|
29 |
+
type: text-generation
|
30 |
+
- dataset:
|
31 |
+
args:
|
32 |
+
num_few_shot: 10
|
33 |
+
name: HellaSwag (10-Shot)
|
34 |
+
split: validation
|
35 |
+
type: hellaswag
|
36 |
+
metrics:
|
37 |
+
- name: normalized accuracy
|
38 |
+
type: acc_norm
|
39 |
+
value: 61.13
|
40 |
+
source:
|
41 |
+
name: Open LLM Leaderboard
|
42 |
+
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=M4-ai/NeuralReyna-Mini-1.8B-v0.3
|
43 |
+
task:
|
44 |
+
name: Text Generation
|
45 |
+
type: text-generation
|
46 |
+
- dataset:
|
47 |
+
args:
|
48 |
+
num_few_shot: 5
|
49 |
+
config: all
|
50 |
+
name: MMLU (5-Shot)
|
51 |
+
split: test
|
52 |
+
type: cais/mmlu
|
53 |
+
metrics:
|
54 |
+
- name: accuracy
|
55 |
+
type: acc
|
56 |
+
value: 44.22
|
57 |
+
source:
|
58 |
+
name: Open LLM Leaderboard
|
59 |
+
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=M4-ai/NeuralReyna-Mini-1.8B-v0.3
|
60 |
+
task:
|
61 |
+
name: Text Generation
|
62 |
+
type: text-generation
|
63 |
+
- dataset:
|
64 |
+
args:
|
65 |
+
num_few_shot: 0
|
66 |
+
config: multiple_choice
|
67 |
+
name: TruthfulQA (0-shot)
|
68 |
+
split: validation
|
69 |
+
type: truthful_qa
|
70 |
+
metrics:
|
71 |
+
- type: mc2
|
72 |
+
value: 41.99
|
73 |
+
source:
|
74 |
+
name: Open LLM Leaderboard
|
75 |
+
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=M4-ai/NeuralReyna-Mini-1.8B-v0.3
|
76 |
+
task:
|
77 |
+
name: Text Generation
|
78 |
+
type: text-generation
|
79 |
+
- dataset:
|
80 |
+
args:
|
81 |
+
num_few_shot: 5
|
82 |
+
config: winogrande_xl
|
83 |
+
name: Winogrande (5-shot)
|
84 |
+
split: validation
|
85 |
+
type: winogrande
|
86 |
+
metrics:
|
87 |
+
- name: accuracy
|
88 |
+
type: acc
|
89 |
+
value: 60.93
|
90 |
+
source:
|
91 |
+
name: Open LLM Leaderboard
|
92 |
+
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=M4-ai/NeuralReyna-Mini-1.8B-v0.3
|
93 |
+
task:
|
94 |
+
name: Text Generation
|
95 |
+
type: text-generation
|
96 |
+
- dataset:
|
97 |
+
args:
|
98 |
+
num_few_shot: 5
|
99 |
+
config: main
|
100 |
+
name: GSM8k (5-shot)
|
101 |
+
split: test
|
102 |
+
type: gsm8k
|
103 |
+
metrics:
|
104 |
+
- name: accuracy
|
105 |
+
type: acc
|
106 |
+
value: 6.75
|
107 |
+
source:
|
108 |
+
name: Open LLM Leaderboard
|
109 |
+
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=M4-ai/NeuralReyna-Mini-1.8B-v0.3
|
110 |
+
task:
|
111 |
+
name: Text Generation
|
112 |
+
type: text-generation
|
113 |
+
model_creator: M4-ai
|
114 |
+
model_name: NeuralReyna-Mini-1.8B-v0.3
|
115 |
+
pipeline_tag: text-generation
|
116 |
+
quantized_by: afrideva
|
117 |
+
tags:
|
118 |
+
- gguf
|
119 |
+
- ggml
|
120 |
+
- quantized
|
121 |
+
- q2_k
|
122 |
+
- q3_k_m
|
123 |
+
- q4_k_m
|
124 |
+
- q5_k_m
|
125 |
+
- q6_k
|
126 |
+
- q8_0
|
127 |
+
---
|
128 |
+
# M4-ai/NeuralReyna-Mini-1.8B-v0.3-GGUF
|
129 |
+
|
130 |
+
Quantized GGUF model files for [NeuralReyna-Mini-1.8B-v0.3](https://huggingface.co/M4-ai/NeuralReyna-Mini-1.8B-v0.3) from [M4-ai](https://huggingface.co/M4-ai)
|
131 |
+
|
132 |
+
|
133 |
+
| Name | Quant method | Size |
|
134 |
+
| ---- | ---- | ---- |
|
135 |
+
| [neuralreyna-mini-1.8b-v0.3.fp16.gguf](https://huggingface.co/afrideva/NeuralReyna-Mini-1.8B-v0.3-GGUF/resolve/main/neuralreyna-mini-1.8b-v0.3.fp16.gguf) | fp16 | 3.68 GB |
|
136 |
+
| [neuralreyna-mini-1.8b-v0.3.q2_k.gguf](https://huggingface.co/afrideva/NeuralReyna-Mini-1.8B-v0.3-GGUF/resolve/main/neuralreyna-mini-1.8b-v0.3.q2_k.gguf) | q2_k | 846.57 MB |
|
137 |
+
| [neuralreyna-mini-1.8b-v0.3.q3_k_m.gguf](https://huggingface.co/afrideva/NeuralReyna-Mini-1.8B-v0.3-GGUF/resolve/main/neuralreyna-mini-1.8b-v0.3.q3_k_m.gguf) | q3_k_m | 1.02 GB |
|
138 |
+
| [neuralreyna-mini-1.8b-v0.3.q4_k_m.gguf](https://huggingface.co/afrideva/NeuralReyna-Mini-1.8B-v0.3-GGUF/resolve/main/neuralreyna-mini-1.8b-v0.3.q4_k_m.gguf) | q4_k_m | 1.22 GB |
|
139 |
+
| [neuralreyna-mini-1.8b-v0.3.q5_k_m.gguf](https://huggingface.co/afrideva/NeuralReyna-Mini-1.8B-v0.3-GGUF/resolve/main/neuralreyna-mini-1.8b-v0.3.q5_k_m.gguf) | q5_k_m | 1.38 GB |
|
140 |
+
| [neuralreyna-mini-1.8b-v0.3.q6_k.gguf](https://huggingface.co/afrideva/NeuralReyna-Mini-1.8B-v0.3-GGUF/resolve/main/neuralreyna-mini-1.8b-v0.3.q6_k.gguf) | q6_k | 1.58 GB |
|
141 |
+
| [neuralreyna-mini-1.8b-v0.3.q8_0.gguf](https://huggingface.co/afrideva/NeuralReyna-Mini-1.8B-v0.3-GGUF/resolve/main/neuralreyna-mini-1.8b-v0.3.q8_0.gguf) | q8_0 | 1.96 GB |
|
142 |
+
|
143 |
+
|
144 |
+
|
145 |
+
## Original Model Card:
|
146 |
+
# NeuralReyna-Mini-1.8B-v0.3
|
147 |
+
![Reyna image](https://th.bing.com/th/id/OIG3.8IBxuT77hh6Y_r1DZ6WK?dpr=2.6&pid=ImgDetMain)
|
148 |
+
# Description
|
149 |
+
Taken aloobun/Reyna-Mini-1.8B-v0.2 and further fine-tuned it using DPO using the argilla/OpenHermes2.5-dpo-binarized-alpha.
|
150 |
+
|
151 |
+
This model has capabilities in coding, math, science, roleplay, and function calling.
|
152 |
+
|
153 |
+
This model was trained on OpenAI's ChatML prompt format.
|
154 |
+
# Evaluation
|
155 |
+
Coming soon
|
156 |
+
|
157 |
+
# Contributions
|
158 |
+
|
159 |
+
Thanks to @aloobun and @Locutusque for their contributions to this model.
|
160 |
+
# [Open LLM Leaderboard Evaluation Results](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)
|
161 |
+
Detailed results can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/details_M4-ai__NeuralReyna-Mini-1.8B-v0.3)
|
162 |
+
|
163 |
+
| Metric |Value|
|
164 |
+
|---------------------------------|----:|
|
165 |
+
|Avg. |41.77|
|
166 |
+
|AI2 Reasoning Challenge (25-Shot)|35.58|
|
167 |
+
|HellaSwag (10-Shot) |61.13|
|
168 |
+
|MMLU (5-Shot) |44.22|
|
169 |
+
|TruthfulQA (0-shot) |41.99|
|
170 |
+
|Winogrande (5-shot) |60.93|
|
171 |
+
|GSM8k (5-shot) | 6.75|
|