metadata
base_model:
- Jebadiah/Aria-coder-7b
tags:
- merge
- mergekit
- lazymergekit
- Jebadiah/Aria-coder-7b
Aria-tree-35-coder-7b
Aria-tree-35-coder-7b is a merge of the following models using LazyMergekit:
🧩 Configuration
name: Jebadiah/Aria-tree-35-coder-7b
models:
- model: Jebadiah/Aria-coder-7b
parameters:
density: 1.0
weight: 1.0
merge_method: passthrough
layers:
- source:
layer_range: [0, 15] # layers before the middle
target:
layer_range: [0, 15]
- source:
layer_range: [16, 16] # duplicate layer 16
target:
layer_range: [16, 18] # copy to positions 16, 17 and 18
- source:
layer_range: [17, 17] # duplicate layer 17
target:
layer_range: [19, 20] # copy to positions 19 and 20
- source:
layer_range: [18, 31] # layers after the middle (shifted by 3)
target:
layer_range: [21, 34]
dtype: float16
💻 Usage
!pip install -qU transformers accelerate
from transformers import AutoTokenizer
import transformers
import torch
model = "Jebadiah/Aria-tree-35-coder-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"])