Commit
•
b7f0977
1
Parent(s):
6463fb7
uploaded readme
Browse files
README.md
ADDED
@@ -0,0 +1,115 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
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 |
+
NeuralOmniWestBeaglake-7B - GGUF
|
11 |
+
- Model creator: https://huggingface.co/paulml/
|
12 |
+
- Original model: https://huggingface.co/paulml/NeuralOmniWestBeaglake-7B/
|
13 |
+
|
14 |
+
|
15 |
+
| Name | Quant method | Size |
|
16 |
+
| ---- | ---- | ---- |
|
17 |
+
| [NeuralOmniWestBeaglake-7B.Q2_K.gguf](https://huggingface.co/RichardErkhov/paulml_-_NeuralOmniWestBeaglake-7B-gguf/blob/main/NeuralOmniWestBeaglake-7B.Q2_K.gguf) | Q2_K | 2.53GB |
|
18 |
+
| [NeuralOmniWestBeaglake-7B.IQ3_XS.gguf](https://huggingface.co/RichardErkhov/paulml_-_NeuralOmniWestBeaglake-7B-gguf/blob/main/NeuralOmniWestBeaglake-7B.IQ3_XS.gguf) | IQ3_XS | 2.81GB |
|
19 |
+
| [NeuralOmniWestBeaglake-7B.IQ3_S.gguf](https://huggingface.co/RichardErkhov/paulml_-_NeuralOmniWestBeaglake-7B-gguf/blob/main/NeuralOmniWestBeaglake-7B.IQ3_S.gguf) | IQ3_S | 2.96GB |
|
20 |
+
| [NeuralOmniWestBeaglake-7B.Q3_K_S.gguf](https://huggingface.co/RichardErkhov/paulml_-_NeuralOmniWestBeaglake-7B-gguf/blob/main/NeuralOmniWestBeaglake-7B.Q3_K_S.gguf) | Q3_K_S | 2.95GB |
|
21 |
+
| [NeuralOmniWestBeaglake-7B.IQ3_M.gguf](https://huggingface.co/RichardErkhov/paulml_-_NeuralOmniWestBeaglake-7B-gguf/blob/main/NeuralOmniWestBeaglake-7B.IQ3_M.gguf) | IQ3_M | 3.06GB |
|
22 |
+
| [NeuralOmniWestBeaglake-7B.Q3_K.gguf](https://huggingface.co/RichardErkhov/paulml_-_NeuralOmniWestBeaglake-7B-gguf/blob/main/NeuralOmniWestBeaglake-7B.Q3_K.gguf) | Q3_K | 3.28GB |
|
23 |
+
| [NeuralOmniWestBeaglake-7B.Q3_K_M.gguf](https://huggingface.co/RichardErkhov/paulml_-_NeuralOmniWestBeaglake-7B-gguf/blob/main/NeuralOmniWestBeaglake-7B.Q3_K_M.gguf) | Q3_K_M | 3.28GB |
|
24 |
+
| [NeuralOmniWestBeaglake-7B.Q3_K_L.gguf](https://huggingface.co/RichardErkhov/paulml_-_NeuralOmniWestBeaglake-7B-gguf/blob/main/NeuralOmniWestBeaglake-7B.Q3_K_L.gguf) | Q3_K_L | 3.56GB |
|
25 |
+
| [NeuralOmniWestBeaglake-7B.IQ4_XS.gguf](https://huggingface.co/RichardErkhov/paulml_-_NeuralOmniWestBeaglake-7B-gguf/blob/main/NeuralOmniWestBeaglake-7B.IQ4_XS.gguf) | IQ4_XS | 3.67GB |
|
26 |
+
| [NeuralOmniWestBeaglake-7B.Q4_0.gguf](https://huggingface.co/RichardErkhov/paulml_-_NeuralOmniWestBeaglake-7B-gguf/blob/main/NeuralOmniWestBeaglake-7B.Q4_0.gguf) | Q4_0 | 3.83GB |
|
27 |
+
| [NeuralOmniWestBeaglake-7B.IQ4_NL.gguf](https://huggingface.co/RichardErkhov/paulml_-_NeuralOmniWestBeaglake-7B-gguf/blob/main/NeuralOmniWestBeaglake-7B.IQ4_NL.gguf) | IQ4_NL | 3.87GB |
|
28 |
+
| [NeuralOmniWestBeaglake-7B.Q4_K_S.gguf](https://huggingface.co/RichardErkhov/paulml_-_NeuralOmniWestBeaglake-7B-gguf/blob/main/NeuralOmniWestBeaglake-7B.Q4_K_S.gguf) | Q4_K_S | 3.86GB |
|
29 |
+
| [NeuralOmniWestBeaglake-7B.Q4_K.gguf](https://huggingface.co/RichardErkhov/paulml_-_NeuralOmniWestBeaglake-7B-gguf/blob/main/NeuralOmniWestBeaglake-7B.Q4_K.gguf) | Q4_K | 4.07GB |
|
30 |
+
| [NeuralOmniWestBeaglake-7B.Q4_K_M.gguf](https://huggingface.co/RichardErkhov/paulml_-_NeuralOmniWestBeaglake-7B-gguf/blob/main/NeuralOmniWestBeaglake-7B.Q4_K_M.gguf) | Q4_K_M | 4.07GB |
|
31 |
+
| [NeuralOmniWestBeaglake-7B.Q4_1.gguf](https://huggingface.co/RichardErkhov/paulml_-_NeuralOmniWestBeaglake-7B-gguf/blob/main/NeuralOmniWestBeaglake-7B.Q4_1.gguf) | Q4_1 | 4.24GB |
|
32 |
+
| [NeuralOmniWestBeaglake-7B.Q5_0.gguf](https://huggingface.co/RichardErkhov/paulml_-_NeuralOmniWestBeaglake-7B-gguf/blob/main/NeuralOmniWestBeaglake-7B.Q5_0.gguf) | Q5_0 | 4.65GB |
|
33 |
+
| [NeuralOmniWestBeaglake-7B.Q5_K_S.gguf](https://huggingface.co/RichardErkhov/paulml_-_NeuralOmniWestBeaglake-7B-gguf/blob/main/NeuralOmniWestBeaglake-7B.Q5_K_S.gguf) | Q5_K_S | 4.65GB |
|
34 |
+
| [NeuralOmniWestBeaglake-7B.Q5_K.gguf](https://huggingface.co/RichardErkhov/paulml_-_NeuralOmniWestBeaglake-7B-gguf/blob/main/NeuralOmniWestBeaglake-7B.Q5_K.gguf) | Q5_K | 4.78GB |
|
35 |
+
| [NeuralOmniWestBeaglake-7B.Q5_K_M.gguf](https://huggingface.co/RichardErkhov/paulml_-_NeuralOmniWestBeaglake-7B-gguf/blob/main/NeuralOmniWestBeaglake-7B.Q5_K_M.gguf) | Q5_K_M | 4.78GB |
|
36 |
+
| [NeuralOmniWestBeaglake-7B.Q5_1.gguf](https://huggingface.co/RichardErkhov/paulml_-_NeuralOmniWestBeaglake-7B-gguf/blob/main/NeuralOmniWestBeaglake-7B.Q5_1.gguf) | Q5_1 | 5.07GB |
|
37 |
+
| [NeuralOmniWestBeaglake-7B.Q6_K.gguf](https://huggingface.co/RichardErkhov/paulml_-_NeuralOmniWestBeaglake-7B-gguf/blob/main/NeuralOmniWestBeaglake-7B.Q6_K.gguf) | Q6_K | 5.53GB |
|
38 |
+
| [NeuralOmniWestBeaglake-7B.Q8_0.gguf](https://huggingface.co/RichardErkhov/paulml_-_NeuralOmniWestBeaglake-7B-gguf/blob/main/NeuralOmniWestBeaglake-7B.Q8_0.gguf) | Q8_0 | 7.17GB |
|
39 |
+
|
40 |
+
|
41 |
+
|
42 |
+
|
43 |
+
Original model description:
|
44 |
+
---
|
45 |
+
tags:
|
46 |
+
- merge
|
47 |
+
- mergekit
|
48 |
+
- lazymergekit
|
49 |
+
- shadowml/WestBeagle-7B
|
50 |
+
- shadowml/Beaglake-7B
|
51 |
+
- mlabonne/NeuralOmniBeagle-7B
|
52 |
+
base_model:
|
53 |
+
- shadowml/WestBeagle-7B
|
54 |
+
- shadowml/Beaglake-7B
|
55 |
+
- mlabonne/NeuralOmniBeagle-7B
|
56 |
+
license: cc-by-nc-4.0
|
57 |
+
---
|
58 |
+
|
59 |
+
# NeuralOmniWestBeaglake-7B
|
60 |
+
|
61 |
+
NeuralOmniWestBeaglake-7B is a merge of the following models using [LazyMergekit](https://colab.research.google.com/drive/1obulZ1ROXHjYLn6PPZJwRR6GzgQogxxb?usp=sharing):
|
62 |
+
* [shadowml/WestBeagle-7B](https://huggingface.co/shadowml/WestBeagle-7B)
|
63 |
+
* [shadowml/Beaglake-7B](https://huggingface.co/shadowml/Beaglake-7B)
|
64 |
+
* [mlabonne/NeuralOmniBeagle-7B](https://huggingface.co/mlabonne/NeuralOmniBeagle-7B)
|
65 |
+
|
66 |
+
## 🧩 Configuration
|
67 |
+
|
68 |
+
```yaml
|
69 |
+
models:
|
70 |
+
- model: mistralai/Mistral-7B-v0.1
|
71 |
+
# no parameters necessary for base model
|
72 |
+
- model: shadowml/WestBeagle-7B
|
73 |
+
parameters:
|
74 |
+
density: 0.65
|
75 |
+
weight: 0.4
|
76 |
+
- model: shadowml/Beaglake-7B
|
77 |
+
parameters:
|
78 |
+
density: 0.6
|
79 |
+
weight: 0.35
|
80 |
+
- model: mlabonne/NeuralOmniBeagle-7B
|
81 |
+
parameters:
|
82 |
+
density: 0.6
|
83 |
+
weight: 0.45
|
84 |
+
merge_method: dare_ties
|
85 |
+
base_model: mistralai/Mistral-7B-v0.1
|
86 |
+
parameters:
|
87 |
+
int8_mask: true
|
88 |
+
dtype: float16
|
89 |
+
```
|
90 |
+
|
91 |
+
## 💻 Usage
|
92 |
+
|
93 |
+
```python
|
94 |
+
!pip install -qU transformers accelerate
|
95 |
+
|
96 |
+
from transformers import AutoTokenizer
|
97 |
+
import transformers
|
98 |
+
import torch
|
99 |
+
|
100 |
+
model = "paulml/NeuralOmniWestBeaglake-7B"
|
101 |
+
messages = [{"role": "user", "content": "What is a large language model?"}]
|
102 |
+
|
103 |
+
tokenizer = AutoTokenizer.from_pretrained(model)
|
104 |
+
prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
|
105 |
+
pipeline = transformers.pipeline(
|
106 |
+
"text-generation",
|
107 |
+
model=model,
|
108 |
+
torch_dtype=torch.float16,
|
109 |
+
device_map="auto",
|
110 |
+
)
|
111 |
+
|
112 |
+
outputs = pipeline(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
|
113 |
+
print(outputs[0]["generated_text"])
|
114 |
+
```
|
115 |
+
|