Spaces:
Sleeping
Sleeping
File size: 7,457 Bytes
61d946f 7c364eb 61d946f 7c364eb b9acf24 65baa5e 7c364eb 65baa5e 786bf8b 65baa5e c5d274e 65baa5e |
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 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 |
# import gradio as gr
# from huggingface_hub import InferenceClient
# """
# For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
# """
# client = InferenceClient("meta-llama/Meta-Llama-3-8B-Instruct")
# ## None type
# def respond(
# message: str,
# history: list[tuple[str, str]], # This will not be used
# system_message: str,
# max_tokens: int,
# temperature: float,
# top_p: float,
# ):
# messages = [{"role": "system", "content": system_message}]
# # Append only the latest user message
# messages.append({"role": "user", "content": message})
# response = ""
# try:
# # Generate response from the model
# for message in client.chat_completion(
# messages,
# max_tokens=max_tokens,
# stream=True,
# temperature=temperature,
# top_p=top_p,
# ):
# if message.choices[0].delta.content is not None:
# token = message.choices[0].delta.content
# response += token
# yield response
# except Exception as e:
# yield f"An error occurred: {e}"
# ],
# )
# if __name__ == "__main__":
# demo.launch()
##Running smothly CHATBOT
# import gradio as gr
# from huggingface_hub import InferenceClient
# """
# For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
# """
# client = InferenceClient("meta-llama/Meta-Llama-3-8B-Instruct")
# def respond(
# message: str,
# history: list[tuple[str, str]], # This will not be used
# system_message: str,
# max_tokens: int,
# temperature: float,
# top_p: float,
# ):
# # Build the messages list
# messages = [{"role": "system", "content": system_message}]
# messages.append({"role": "user", "content": message})
# response = ""
# try:
# # Generate response from the model
# for msg in client.chat_completion(
# messages=messages,
# max_tokens=max_tokens,
# stream=True,
# temperature=temperature,
# top_p=top_p,
# ):
# if msg.choices[0].delta.content is not None:
# token = msg.choices[0].delta.content
# response += token
# yield response
# except Exception as e:
# yield f"An error occurred: {e}"
# """
# For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
# """
# demo = gr.ChatInterface(
# respond,
# additional_inputs=[
# gr.Textbox(value="You are a friendly Chatbot.", 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()
# Use a pipeline as a high-level helper
import spaces
import os
import subprocess
from llama_cpp import Llama
from llama_cpp_agent import LlamaCppAgent, MessagesFormatterType
from llama_cpp_agent.providers import LlamaCppPythonProvider
from llama_cpp_agent.chat_history import BasicChatHistory
from llama_cpp_agent.chat_history.messages import Roles
import gradio as gr
from huggingface_hub import hf_hub_download
huggingface_token = os.getenv("HF_TOKEN")
# Download the Meta-Llama-3.1-8B-Instruct model
hf_hub_download(
repo_id="bartowski/Meta-Llama-3.1-8B-Instruct-GGUF",
filename="Meta-Llama-3.1-8B-Instruct-Q5_K_M.gguf",
local_dir="./models",
token=huggingface_token
)
llm = None
llm_model = None
@spaces.GPU(duration=120)
def respond(
message,
history: list[tuple[str, str]],
model,
system_message,
max_tokens,
temperature,
top_p,
top_k,
repeat_penalty,
):
chat_template = MessagesFormatterType.GEMMA_2
global llm
global llm_model
# Load model only if it's not already loaded or if a new model is selected
if llm is None or llm_model != model:
try:
llm = Llama(
model_path=f"models/{model}",
flash_attn=True,
n_gpu_layers=81, # Adjust based on available GPU resources
n_batch=1024,
n_ctx=8192,
)
llm_model = model
except Exception as e:
return f"Error loading model: {str(e)}"
provider = LlamaCppPythonProvider(llm)
agent = LlamaCppAgent(
provider,
system_prompt=f"{system_message}",
predefined_messages_formatter_type=chat_template,
debug_output=True
)
settings = provider.get_provider_default_settings()
settings.temperature = temperature
settings.top_k = top_k
settings.top_p = top_p
settings.max_tokens = max_tokens
settings.repeat_penalty = repeat_penalty
settings.stream = True
messages = BasicChatHistory()
# Add user and assistant messages to the history
for msn in history:
user = {'role': Roles.user, 'content': msn[0]}
assistant = {'role': Roles.assistant, 'content': msn[1]}
messages.add_message(user)
messages.add_message(assistant)
# Stream the response
try:
stream = agent.get_chat_response(
message,
llm_sampling_settings=settings,
chat_history=messages,
returns_streaming_generator=True,
print_output=False
)
outputs = ""
for output in stream:
outputs += output
yield outputs
except Exception as e:
yield f"Error during response generation: {str(e)}"
description = """<p align="center">Using the Meta-Llama-3.1-8B-Instruct Model</p>"""
demo = gr.ChatInterface(
respond,
additional_inputs=[
gr.Dropdown([
'Meta-Llama-3.1-8B-Instruct-Q5_K_M.gguf'
],
value="Meta-Llama-3.1-8B-Instruct-Q5_K_M.gguf",
label="Model"
),
gr.Textbox(value="You are a helpful assistant.", label="System message"),
gr.Slider(minimum=1, maximum=4096, value=2048, step=1, label="Max 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",
),
gr.Slider(
minimum=0,
maximum=100,
value=40,
step=1,
label="Top-k",
),
gr.Slider(
minimum=0.0,
maximum=2.0,
value=1.1,
step=0.1,
label="Repetition penalty",
),
],
retry_btn="Retry",
undo_btn="Undo",
clear_btn="Clear",
submit_btn="Send",
title="Chat with Meta-Llama-3.1-8B-Instruct using llama.cpp",
description=description,
chatbot=gr.Chatbot(
scale=1,
likeable=False,
show_copy_button=True
)
)
if __name__ == "__main__":
demo.launch() |