dtype: bfloat16
merge_method: linear
slices:
 - sources:
      - layer_range: [0, 16]
        model: codellama/CodeLlama-7b-Instruct-hf
        parameters:
          weight: 0.25
      - layer_range: [16, 32]
        model: codellama/CodeLlama-7b-Instruct-hf
        parameters:
          weight: 0.25

💻 Usage

python

!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

model = "JoPmt/CodeLlemur-2B-Instruct-line"
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"])

RichardErkhov/JoPmt_-_CodeLlemur-3.5B-Instruct-line-8bits

Original model description:

CodeLlemur-2B-Instruct-line

🧩 Configuration

💻 Usage

RichardErkhov/
JoPmt_-_CodeLlemur-3.5B-Instruct-line-8bits