neo_7b-slerp is a merge of the following models using
LazyMergekit:
1slices:
2 - sources:
3 - model: m-a-p/neo_7b
4 layer_range: [0, 1]
5 - model: m-a-p/neo_7b
6 layer_range: [1, 2]
7 - sources:
8 - model: m-a-p/neo_7b
9 layer_range: [2, 3]
10 - model: m-a-p/neo_7b
11 layer_range: [3, 4]
12 - sources:
13 - model: m-a-p/neo_7b
14 layer_range: [4, 5]
15 - model: m-a-p/neo_7b
16 layer_range: [5,6]
17 - sources:
18 - model: m-a-p/neo_7b
19 layer_range: [6, 7]
20 - model: m-a-p/neo_7b
21 layer_range: [7, 8]
22 - sources:
23 - model: m-a-p/neo_7b
24 layer_range: [8, 9]
25 - model: m-a-p/neo_7b
26 layer_range: [9, 10]
27 - sources:
28 - model: m-a-p/neo_7b
29 layer_range: [10, 11]
30 - model: m-a-p/neo_7b
31 layer_range: [11, 12]
32 - sources:
33 - model: m-a-p/neo_7b
34 layer_range: [12, 13]
35 - model: m-a-p/neo_7b
36 layer_range: [13, 14]
37 - sources:
38 - model: m-a-p/neo_7b
39 layer_range: [14, 15]
40 - model: m-a-p/neo_7b
41 layer_range: [15, 16]
42 - sources:
43 - model: m-a-p/neo_7b
44 layer_range: [16, 17]
45 - model: m-a-p/neo_7b
46 layer_range: [17, 18]
47 - sources:
48 - model: m-a-p/neo_7b
49 layer_range: [18, 19]
50 - model: m-a-p/neo_7b
51 layer_range: [19, 20]
52 - sources:
53 - model: m-a-p/neo_7b
54 layer_range: [20, 21]
55 - model: m-a-p/neo_7b
56 layer_range: [21, 22]
57 - sources:
58 - model: m-a-p/neo_7b
59 layer_range: [22, 23]
60 - model: m-a-p/neo_7b
61 layer_range: [23, 24]
62 - sources:
63 - model: m-a-p/neo_7b
64 layer_range: [24, 25]
65 - model: m-a-p/neo_7b
66 layer_range: [25, 26]
67 - sources:
68 - model: m-a-p/neo_7b
69 layer_range: [26, 27]
70 - model: m-a-p/neo_7b
71 layer_range: [27, 28]
72merge_method: slerp
73base_model: m-a-p/neo_7b
74parameters:
75 t: 0.5
76dtype: bfloat16
1!pip install -qU transformers accelerate
2
3from transformers import AutoTokenizer
4import transformers
5import torch
6
7model = "DewEfresh/neo_7b-slerp"
8messages = [{"role": "user", "content": "What is a large language model?"}]
9
10tokenizer = AutoTokenizer.from_pretrained(model)
11prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
12pipeline = transformers.pipeline(
13 "text-generation",
14 model=model,
15 torch_dtype=torch.float16,
16 device_map="auto",
17)
18
19outputs = pipeline(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
20print(outputs[0]["generated_text"])
