koesn commited on
Commit
c969a25
1 Parent(s): 003b21e

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +99 -0
README.md CHANGED
@@ -1,3 +1,102 @@
1
  ---
2
  license: apache-2.0
 
 
3
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  ---
2
  license: apache-2.0
3
+ language:
4
+ - en
5
  ---
6
+ # NeuralMaxime-7B-slerp-GGUF
7
+
8
+
9
+ ## Description
10
+
11
+ This repo contains GGUF format model files for NeuralMaxime-7B-slerp-GGUF.
12
+
13
+ ## Files Provided
14
+
15
+ | Name | Quant | Bits | File Size | Remark |
16
+ | --------------------------------- | ------ | ---- | --------- | -------------------------------- |
17
+ | neuralmaxime-7b-slerp.IQ3_S.gguf | IQ3_S | 3 | 3.18 GB | 3.44 bpw quantization |
18
+ | neuralmaxime-7b-slerp.IQ3_M.gguf | IQ3_M | 3 | 3.28 GB | 3.66 bpw quantization mix |
19
+ | neuralmaxime-7b-slerp.Q4_0.gguf | Q4_0 | 4 | 4.11 GB | 3.56G, +0.2166 ppl |
20
+ | neuralmaxime-7b-slerp.IQ4_NL.gguf | IQ4_NL | 4 | 4.16 GB | 4.25 bpw non-linear quantization |
21
+ | neuralmaxime-7b-slerp.Q4_K_M.gguf | Q4_K_M | 4 | 4.37 GB | 3.80G, +0.0532 ppl |
22
+ | neuralmaxime-7b-slerp.Q5_K_M.gguf | Q5_K_M | 5 | 5.13 GB | 4.45G, +0.0122 ppl |
23
+ | neuralmaxime-7b-slerp.Q6_K.gguf | Q6_K | 6 | 5.94 GB | 5.15G, +0.0008 ppl |
24
+ | neuralmaxime-7b-slerp.Q8_0.gguf | Q8_0 | 8 | 7.70 GB | 6.70G, +0.0004 ppl |
25
+
26
+ ## Parameters
27
+
28
+ | path | type | architecture | rope_theta | sliding_win | max_pos_embed |
29
+ | ----------------------------- | ------- | ------------------ | ---------- | ----------- | ------------- |
30
+ | Kukedlc/NeuralMaxime-7B-slerp | mistral | MistralForCausalLM | 10000.0 | 4096 | 32768 |
31
+
32
+ ## Benchmarks
33
+
34
+ ![](https://i.ibb.co/g7sqr1r/Neural-Maxime-7-B-slerp.png)
35
+
36
+ # Original Model Card
37
+
38
+ ---
39
+ tags:
40
+ - merge
41
+ - mergekit
42
+ - lazymergekit
43
+ - mlabonne/AlphaMonarch-7B
44
+ - mlabonne/NeuralMonarch-7B
45
+ base_model:
46
+ - mlabonne/AlphaMonarch-7B
47
+ - mlabonne/NeuralMonarch-7B
48
+ license: apache-2.0
49
+ ---
50
+
51
+ # NeuralMaxime-7B-slerp
52
+
53
+ ![](https://raw.githubusercontent.com/kukedlc87/imagenes/main/DALL%C2%B7E%202024-02-18%2015.45.07%20-%20Visualize%20a%20highly%20sophisticated%2C%20high-definition%20robot%20named%20Neural%20Maxime.%20This%20language%20model%20robot%20is%20distinguished%20by%20its%20innovative%20design%2C%20feat.webp)
54
+ NeuralMaxime-7B-slerp is a merge of the following models using [LazyMergekit](https://colab.research.google.com/drive/1obulZ1ROXHjYLn6PPZJwRR6GzgQogxxb?usp=sharing):
55
+ * [mlabonne/AlphaMonarch-7B](https://huggingface.co/mlabonne/AlphaMonarch-7B)
56
+ * [mlabonne/NeuralMonarch-7B](https://huggingface.co/mlabonne/NeuralMonarch-7B)
57
+
58
+ ## 🧩 Configuration
59
+
60
+ ```yaml
61
+ slices:
62
+ - sources:
63
+ - model: mlabonne/AlphaMonarch-7B
64
+ layer_range: [0, 32]
65
+ - model: mlabonne/NeuralMonarch-7B
66
+ layer_range: [0, 32]
67
+ merge_method: slerp
68
+ base_model: mlabonne/AlphaMonarch-7B
69
+ parameters:
70
+ t:
71
+ - filter: self_attn
72
+ value: [0, 0.5, 0.3, 0.7, 1]
73
+ - filter: mlp
74
+ value: [1, 0.5, 0.7, 0.3, 0]
75
+ - value: 0.5
76
+ dtype: bfloat16
77
+ ```
78
+
79
+ ## 💻 Usage
80
+
81
+ ```python
82
+ !pip install -qU transformers accelerate
83
+
84
+ from transformers import AutoTokenizer
85
+ import transformers
86
+ import torch
87
+
88
+ model = "Kukedlc/NeuralMaxime-7B-slerp"
89
+ messages = [{"role": "user", "content": "What is a large language model?"}]
90
+
91
+ tokenizer = AutoTokenizer.from_pretrained(model)
92
+ prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
93
+ pipeline = transformers.pipeline(
94
+ "text-generation",
95
+ model=model,
96
+ torch_dtype=torch.float16,
97
+ device_map="auto",
98
+ )
99
+
100
+ outputs = pipeline(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
101
+ print(outputs[0]["generated_text"])
102
+ ```