slices:
  - sources:
      - model: udkai/Turdus
        layer_range: [0, 32]
      - model: flemmingmiguel/MBX-7B
        layer_range: [0, 32]
merge_method: slerp
base_model: udkai/Turdus
parameters:
  t:
    - filter: self_attn
      value: [0, 0.5, 0.3, 0.7, 1]
    - filter: mlp
      value: [1, 0.5, 0.7, 0.3, 0]
    - value: 0.5
dtype: bfloat16

💻 Usage

python

!pip install -qU transformers accelerate

from transformers import AutoTokenizer
import transformers
import torch

# Load tokenizer and model
model = "Kumar955/Hemanth-llm"
tokenizer = AutoTokenizer.from_pretrained(model)

# Define the messages from the conversation
messages = [{"role": "user", "content": "What is a large language model?"}]

# Define the chat template for formatting the conversation
chat_template = """<s><|user|>{{ user_message }}<|assistant|>"""

# Extract the user message content
user_message = messages[0]["content"]

# Format the prompt using the chat template
prompt = chat_template.replace("{{ user_message }}", user_message)

# Load the pipeline with the specified model
pipeline = pipeline(
    "text-generation",
    model=model,
    torch_dtype=torch.float16,
    device_map="auto",
)

# Generate output with the model
outputs = pipeline(prompt, max_new_tokens=256, do_sample=True, temperature=0.7, top_k=50, top_p=0.95)

# Print the generated response
print(outputs[0]["generated_text"])