Spaces:
Build error
Build error
| """ | |
| Try out gradio.Chatinterface. | |
| colab gradio-chatinterface. | |
| %%writefile reuirements.txt | |
| gradio | |
| transformers | |
| sentencepiece | |
| torch | |
| """ | |
| # pylint: disable=line-too-long, missing-module-docstring, missing-function-docstring | |
| # import torch | |
| import gradio as gr | |
| from examples_list import examples_list | |
| from transformers import AutoModel, AutoTokenizer # AutoModelForCausalLM, | |
| # device = "cuda" if torch.cuda.is_available() else "cpu" | |
| # tokenizer = AutoTokenizer.from_pretrained("stabilityai/StableBeluga2", use_fast=False) | |
| # model = AutoModelForCausalLM.from_pretrained("stabilityai/StableBeluga2", torch_dtype=torch.float16, low_cpu_mem_usage=True, device_map="auto") | |
| # system_prompt = "### System:\nYou are Stable Beluga, an AI that follows instructions extremely well. Help as much as you can. Remember, be safe, and don't do anything illegal.\n\n" | |
| # pipeline = pipeline(task="text-generation", model="meta-llama/Llama-2-7b") | |
| _ = """ | |
| tokenizer = AutoTokenizer.from_pretrained( | |
| "THUDM/chatglm2-6b-int4", trust_remote_code=True | |
| ) | |
| chat_model = AutoModel.from_pretrained( | |
| "THUDM/chatglm2-6b-int4", trust_remote_code=True # 3.92G | |
| ).float() | |
| """ | |
| def stream_chat(): | |
| """samples: | |
| Sure [('test me', 'Sure')] | |
| Sure, [('test me', 'Sure,')] | |
| Sure, I [('test me', 'Sure, I')] | |
| Sure, I' [('test me', "Sure, I'")] | |
| Sure, I'd [('test me', "Sure, I'd")] | |
| """ | |
| resp = "" | |
| for elm in range(10): | |
| resp += str(elm) | |
| from time import sleep | |
| sleep(0.1) | |
| yield elm | |
| def chat(message="", history=[]): | |
| # prompt = f"{system_prompt}### User: {message}\n\n### Assistant:\n" | |
| # inputs = tokenizer(prompt, return_tensors="pt").to(device=device) | |
| # output = model.generate(**inputs, do_sample=True, top_p=0.95, top_k=0, max_new_tokens=256) | |
| # return tokenizer.decode(output[0], skip_special_tokens=True) | |
| _ = """ | |
| for response, _ in chat_model.stream_chat( | |
| tokenizer, message, history, max_length=2048, top_p=0.7, temperature=0.95 | |
| ): | |
| yield response | |
| """ | |
| g = update_chatbot() | |
| g.send(None) | |
| for response in stream_chat(): | |
| # yield response | |
| g.send(response) | |
| return history | |
| def update_chatbot(): | |
| while 1: | |
| message = yield | |
| print(f"{message=}") | |
| def chat1(message, history): | |
| # prompt = f"{system_prompt}### User: {message}\n\n### Assistant:\n" | |
| # inputs = tokenizer(prompt, return_tensors="pt").to(device=device) | |
| # output = model.generate(**inputs, do_sample=True, top_p=0.95, top_k=0, max_new_tokens=256) | |
| # return tokenizer.decode(output[0], skip_special_tokens=True) | |
| for response, _ in chat_model.stream_chat( | |
| tokenizer, message, history, max_length=2048, top_p=0.7, temperature=0.95 | |
| ): | |
| yield response, _ | |
| with gr.Blocks(theme=gr.themes.Glass(text_size="sm", spacing_size="sm"),) as block: | |
| chatbot = gr.Chatbot() | |
| msg = gr.Textbox() | |
| # gr.ChatInterface( | |
| block( | |
| chat, | |
| [msg, chatbot], | |
| [chatbot], | |
| # title="gradio-chatinterface-tryout", | |
| # examples=examples_list, | |
| ).queue(max_size=2).launch() | |