Elysium2-11b-Slerp is a merge of the following models using
LazyMergekit:
1slices:
2 - sources:
3 - model: powermove72/GreenScorpius-xb-Passthrough
4 layer_range: [0, 50]
5 - model: powermove72/Trinity_Notus-xb
6 layer_range: [0, 50]
7merge_method: slerp
8base_model: powermove72/GreenScorpius-xb-Passthrough
9parameters:
10 t:
11 - filter: self_attn
12 value: [0, 0.5, 0.3, 0.7, 1]
13 - filter: mlp
14 value: [1, 0.5, 0.7, 0.3, 0]
15 - value: 0.5
16dtype: bfloat16
17 ```
18
19## 💻 Usage
20
21```python
22!pip install -qU transformers accelerate
23
24from transformers import AutoTokenizer
25import transformers
26import torch
27
28model = "powermove72/Elysium2-11b-Slerp"
29messages = [{"role": "user", "content": "What is a large language model?"}]
30
31tokenizer = AutoTokenizer.from_pretrained(model)
32prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
33pipeline = transformers.pipeline(
34 "text-generation",
35 model=model,
36 torch_dtype=torch.float16,
37 device_map="auto",
38)
39
40outputs = pipeline(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
41print(outputs[0]["generated_text"])
