File size: 4,693 Bytes
aae630c
 
a663068
 
 
407802b
a5deebe
aae630c
 
 
 
a663068
 
407802b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
fc5bd6e
d25f823
aae630c
a663068
 
 
 
 
 
 
 
 
 
 
 
a5deebe
 
 
 
 
 
 
 
aae630c
a5deebe
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
8cc88b0
a5deebe
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
eda9136
a5deebe
 
 
 
eda9136
a5deebe
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
3548597
aae630c
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
import gradio as gr
from huggingface_hub import InferenceClient
import transformers
from transformers import AutoTokenizer,GenerationConfig
import torch
from peft import PeftModel
import spaces

"""
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("HuggingFaceH4/zephyr-7b-beta")
base_model = "Neko-Institute-of-Science/LLaMA-65B-HF"
lora_weights = "./"
#lora_weights = LoraConfig(
#    auto_mapping=None,
#    base_model_name_or_path="Neko-Institute-of-Science/LLaMA-65B-HF",
#   bias=None,
#    fan_in_fan_out=False,
#    inference_mode=True,
#    init_lora_weights=True,
#    layers_pattern=None,
#    layers_to_transform=None,
#    lora_alpha=16,
#    lora_dropout=0.05,
#    modules_to_save=None,
#    peft_type="LORA",
#    revision=None,
#    target_modules=["q_proj","k_proj","v_proj","o_proj","gate_proj","up_proj","down_proj"],
#    task_type="CAUSAL_LM",
#)

cache_dir = "/data"

PROMPT_DICT = {
    "prompt_input": (
        "Below is an instruction that describes a task, paired with further context. "
        "Write a response that appropriately completes the request.\n\n"
        "Instruction:\n{instruction}\n\n Input:\n{input}\n\n Response:"
    ),
    "prompt_no_input": (
        "Below is an instruction that describes a task. "
        "Write a response that appropriately completes the request.\n\n"
        "Instruction:\n{instruction}\n\nResponse:"
    ),
}
from llama_rope_scaled_monkey_patch import replace_llama_rope_with_scaled_rope
replace_llama_rope_with_scaled_rope()
model = transformers.AutoModelForCausalLM.from_pretrained(
            base_model,
            torch_dtype=torch.float16,
            cache_dir=cache_dir,
            device_map="auto",
        )

model = PeftModel.from_pretrained(
            model,
            lora_weights,
            device_map="auto",
            cache_dir=cache_dir,
            torch_dtype=torch.float16,
        )
tokenizer =  AutoTokenizer.from_pretrained(base_model,use_fast=False,cache_dir=cache_dir)
tokenizer.pad_token = tokenizer.unk_token
def generate_prompt(instruction, input=None):
    if input:
        return PROMPT_DICT["prompt_input"].format(instruction=instruction,input=input)
    else:
        return PROMPT_DICT["prompt_no_input"].format(instruction=instruction)

@spaces.GPU
def respond(
    message,
    history: list[tuple[str, str]],
    system_message,
    max_tokens,
    temperature,
    top_p,
):
    ins_f = generate_prompt(message,None)
    inputs  =  tokenizer(ins_f, return_tensors="pt")
    input_ids = inputs["input_ids"].cuda()
    generation_config = GenerationConfig(
            temperature=0.1,
            top_p=0.75,
            top_k=40,
            do_sample=True,
            num_beams=1,
            max_new_tokens = 512
        )

    # Without streaming
    with torch.no_grad():
        generation_output = model.generate(
                input_ids=input_ids,
                generation_config=generation_config,
                return_dict_in_generate=True,
                output_scores=False,
                max_new_tokens=max_new_tokens,
            )
    s = generation_output.sequences[0]
    output = tokenizer.decode(s)
    response = output.split("Response:")[1].strip()
    yield response
    
    #messages = [{"role": "system", "content": system_message}]

    #for val in history:
    #    if val[0]:
    #        messages.append({"role": "user", "content": val[0]})
    #    if val[1]:
    #        messages.append({"role": "assistant", "content": val[1]})

    # messages.append({"role": "user", "content": message})

    #response = ""

    #for message in client.chat_completion(
    #    messages,
    #    max_tokens=max_tokens,
    #    stream=True,
    #    temperature=temperature,
    #    top_p=top_p,
    #):
    #    token = message.choices[0].delta.content

    #    response += token
    #    yield response


"""
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__":
    model.eval()
    demo.launch()