neo_7b-slerp
neo_7b-slerp is a merge of the following models using LazyMergekit:
🧩 Configuration
slices:
- sources:
- model: m-a-p/neo_7b
layer_range: [0, 1]
- model: m-a-p/neo_7b
layer_range: [1, 2]
- sources:
- model: m-a-p/neo_7b
layer_range: [2, 3]
- model: m-a-p/neo_7b
layer_range: [3, 4]
- sources:
- model: m-a-p/neo_7b
layer_range: [4, 5]
- model: m-a-p/neo_7b
layer_range: [5,6]
- sources:
- model: m-a-p/neo_7b
layer_range: [6, 7]
- model: m-a-p/neo_7b
layer_range: [7, 8]
- sources:
- model: m-a-p/neo_7b
layer_range: [8, 9]
- model: m-a-p/neo_7b
layer_range: [9, 10]
- sources:
- model: m-a-p/neo_7b
layer_range: [10, 11]
- model: m-a-p/neo_7b
layer_range: [11, 12]
- sources:
- model: m-a-p/neo_7b
layer_range: [12, 13]
- model: m-a-p/neo_7b
layer_range: [13, 14]
- sources:
- model: m-a-p/neo_7b
layer_range: [14, 15]
- model: m-a-p/neo_7b
layer_range: [15, 16]
- sources:
- model: m-a-p/neo_7b
layer_range: [16, 17]
- model: m-a-p/neo_7b
layer_range: [17, 18]
- sources:
- model: m-a-p/neo_7b
layer_range: [18, 19]
- model: m-a-p/neo_7b
layer_range: [19, 20]
- sources:
- model: m-a-p/neo_7b
layer_range: [20, 21]
- model: m-a-p/neo_7b
layer_range: [21, 22]
- sources:
- model: m-a-p/neo_7b
layer_range: [22, 23]
- model: m-a-p/neo_7b
layer_range: [23, 24]
- sources:
- model: m-a-p/neo_7b
layer_range: [24, 25]
- model: m-a-p/neo_7b
layer_range: [25, 26]
- sources:
- model: m-a-p/neo_7b
layer_range: [26, 27]
- model: m-a-p/neo_7b
layer_range: [27, 28]
merge_method: slerp
base_model: m-a-p/neo_7b
parameters:
t: 0.5
dtype: bfloat16
💻 Usage
!pip install -qU transformers accelerate
from transformers import AutoTokenizer
import transformers
import torch
model = "DewEfresh/neo_7b-slerp"
messages = [{"role": "user", "content": "What is a large language model?"}]
tokenizer = AutoTokenizer.from_pretrained(model)
prompt = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
pipeline = transformers.pipeline(
"text-generation",
model=model,
torch_dtype=torch.float16,
device_map="auto",
)
outputs = pipeline(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)
print(outputs[0]["generated_text"])