Mistral 7B Merges
Collection
Merges that may or may not be worth using. All credit goes to Maxime Labonne's course, https://github.com/mlabonne/llm-course, + mergekit
•
6 items
•
Updated
WONMSeverusDevil-TIES-7B is a merge of the following models using LazyMergekit:
# Open-LLM Benchmark Results:
WONMSeverusDevil-TIES-7B LLM AutoEval📑
| Model |AGIEval|GPT4All|TruthfulQA|Bigbench|Average|
|------------------------|------:|------:|---------:|-------:|------:|
|WONMSeverusDevil-TIES-7B| 45.26| 77.07| 72.47| 48.85| 60.91|
models:
- model: FelixChao/WestSeverus-7B-DPO-v2
parameters:
density: [1, 0.7, 0.1] # density gradient
weight: 1.0
- model: jsfs11/WestOrcaNeuralMarco-DPO-v2-DARETIES-7B
parameters:
density: 0.65
weight: [0, 0.3, 0.7, 1] # weight gradient
- model: mlabonne/Daredevil-7B
parameters:
density: 0.33
weight:
- filter: mlp
value: 0.5
- value: 0
merge_method: ties
base_model: mistralai/Mistral-7B-v0.1
parameters:
normalize: true
int8_mask: true
dtype: float16
!pip install -qU transformers accelerate
from transformers import AutoTokenizer
import transformers
import torch
model = "jsfs11/WONMSeverusDevil-TIES-7B"
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"])