| import gradio as gr | |
| from llama_cpp import Llama | |
| llm = Llama(model_path="vicuna-AlekseyKorshuk-7B-GPTQ-4bit-128g.GGML.bin", n_ctx=2048) | |
| def generate_text(llm, prompt, max_tokens=468, temperature=0.1, top_p=0.5): | |
| output = llm(prompt, max_tokens=max_tokens, temperature=temperature, top_p=top_p, echo=False, stop=["#"]) | |
| text = output['choices'][0]['text'] | |
| return text | |
| input_text = gr.inputs.Textbox(lines= 10, label="Enter your input text") | |
| output_text = gr.outputs.Textbox(label="Output text") | |
| description = "llama.cpp implementation in python [https://github.com/abetlen/llama-cpp-python]" | |
| examples = [ | |
| ["What is the capital of France? ", "The capital of France is Paris."], | |
| ["Who wrote the novel 'Pride and Prejudice'?", "The novel 'Pride and Prejudice' was written by Jane Austen."], | |
| ["What is the square root of 64?", "The square root of 64 is 8."] | |
| ] | |
| gr.Interface(fn=generate_text, inputs=input_text, outputs=output_text, title="Vicuna Language Model", description=description, examples=examples).launch() | |