File size: 2,005 Bytes
21b8ce0
b5c263a
d8a3c53
606c0ce
 
 
 
 
d8a3c53
b0f00f9
ec06a49
98758c3
d8a3c53
3b38821
d8a3c53
 
 
 
 
 
 
 
4d0808e
d8a3c53
606c0ce
ec06a49
606c0ce
 
 
 
 
6c9b3f7
d8a3c53
606c0ce
 
 
 
 
 
827a15f
606c0ce
d8a3c53
 
 
606c0ce
d8a3c53
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
import spaces
import subprocess
import gradio as gr
from huggingface_hub import hf_hub_download
from llama_cpp import Llama
from llama_cpp_agent import LlamaCppAgent
from llama_cpp_agent import MessagesFormatterType
from llama_cpp_agent.providers import LlamaCppPythonProvider

subprocess.run('pip install llama-cpp-python==0.2.75 --no-build-isolation --no-cache-dir --upgrade --only-binary=:all: --extra-index-url=https://abetlen.github.io/llama-cpp-python/whl/cu124', env={'CMAKE_ARGS': "-DLLAMA_CUDA=on"}, shell=True)

hf_hub_download(repo_id="TheBloke/Mistral-7B-Instruct-v0.2-GGUF", filename="mistral-7b-instruct-v0.2.Q6_K.gguf",  local_dir = "./models")

@spaces.GPU(duration=120)
def respond(
    message,
    history: list[tuple[str, str]],
    system_message,
    max_tokens,
    temperature,
    top_p,
):
    llama_model = Llama(r"models/mistral-7b-instruct-v0.2.Q6_K.gguf", n_batch=1024, n_threads=0, n_gpu_layers=33, n_ctx=8192, verbose=False)

    provider = LlamaCppPythonProvider(llama_model)

    agent = LlamaCppAgent(
      provider,
      system_prompt=f"{system_message}",
      predefined_messages_formatter_type=MessagesFormatterType.MISTRAL,
      debug_output=True
    )

    settings = provider.get_provider_default_settings()
    settings.stream = True
    settings.max_tokens = max_tokens
    settings.temperature = temperature
    settings.top_p = top_p

    yield agent.get_chat_response(message, llm_sampling_settings=settings)

demo = gr.ChatInterface(
    respond,
    additional_inputs=[
        gr.Textbox(value="You are a helpful assistant.", label="System message"),
        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
        gr.Slider(
            minimum=0.1,
            maximum=1.0,
            value=0.95,
            step=0.05,
            label="Top-p (nucleus sampling)",
        ),
    ],
)

if __name__ == "__main__":
    demo.launch()