README.md

metadata

license: apache-2.0
datasets:
  - databricks/databricks-dolly-15k
language:
  - en
base_model: TinyLlama/TinyLlama-1.1B-Chat-v1.0
pipeline_tag: text-generation

TinyLlama Dolly 15

The TinyLlama Dolly 15k is a specialized large language model finely tuned on the Databricks Dolly 15k dataset. This dataset, composed of 15,000 high-quality, human-curated prompts and responses, helps the model excel in understanding and generating contextually relevant and coherent text. By leveraging this diverse and comprehensive dataset, TinyLlama Dolly 15k enhances its ability to engage in nuanced conversations and provide accurate, context-aware responses across a range of topics. The result is a powerful tool for applications requiring advanced natural language understanding and generation.

How to use

Below is a snippet that can be used to test the model.

from transformers import pipeline

pipe = pipeline("text-generation", model="Bhooyas/tinyllama-dolly-15k", device_map="auto")

messages = [
    {
        "role": "user",
        "context": "The TinyLlama Dolly 15k is a specialized large language model finely tuned on the Databricks Dolly 15k dataset. This dataset, composed of 15,000 high-quality, human-curated prompts and responses, helps the model excel in understanding and generating contextually relevant and coherent text. By leveraging this diverse and comprehensive dataset, TinyLlama Dolly 15k enhances its ability to engage in nuanced conversations and provide accurate, context-aware responses across a range of topics. The result is a powerful tool for applications requiring advanced natural language understanding and generation.",
        "instruction": "What is TinyLlama Dolly 15k?"
    }
]

prompt = pipe.tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
outputs = pipe(prompt, max_new_tokens=256)
print(outputs[0]["generated_text"])