halimbahae/text-to-code
0
1import json2import os3import shutil4import requests5 6import gradio as gr7from huggingface_hub import Repository8from text_generation import Client9 10from share_btn import community_icon_html, loading_icon_html, share_js, share_btn_css11 12HF_TOKEN = os.environ.get("HF_TOKEN", None)13 14API_URL = "https://api-inference.huggingface.co/models/codellama/CodeLlama-13b-hf"15 16FIM_PREFIX = "<PRE> "17FIM_MIDDLE = " <MID>"18FIM_SUFFIX = " <SUF>"19 20FIM_INDICATOR = "<FILL_ME>"21 22EOS_STRING = "</s>"23EOT_STRING = "<EOT>"24 25theme = gr.themes.Monochrome(26 primary_hue="indigo",27 secondary_hue="blue",28 neutral_hue="slate",29 radius_size=gr.themes.sizes.radius_sm,30 font=[31 gr.themes.GoogleFont("Open Sans"),32 "ui-sans-serif",33 "system-ui",34 "sans-serif",35 ],36)37 38client = Client(39 API_URL,40 headers={"Authorization": f"Bearer {HF_TOKEN}"},41)42 43 44def generate(45 prompt, temperature=0.9, max_new_tokens=256, top_p=0.95, repetition_penalty=1.0,46):47 48 temperature = float(temperature)49 if temperature < 1e-2:50 temperature = 1e-251 top_p = float(top_p)52 fim_mode = False53 54 generate_kwargs = dict(55 temperature=temperature,56 max_new_tokens=max_new_tokens,57 top_p=top_p,58 repetition_penalty=repetition_penalty,59 do_sample=True,60 seed=42,61 )62 63 if FIM_INDICATOR in prompt:64 fim_mode = True65 try:66 prefix, suffix = prompt.split(FIM_INDICATOR)67 except:68 raise ValueError(f"Only one {FIM_INDICATOR} allowed in prompt!")69 prompt = f"{FIM_PREFIX}{prefix}{FIM_SUFFIX}{suffix}{FIM_MIDDLE}"70 71 72 stream = client.generate_stream(prompt, **generate_kwargs)73 74 75 if fim_mode:76 output = prefix77 else:78 output = prompt79 80 previous_token = ""81 for response in stream:82 if any([end_token in response.token.text for end_token in [EOS_STRING, EOT_STRING]]):83 if fim_mode:84 output += suffix85 yield output86 return output87 print("output", output)88 else:89 return output90 else:91 output += response.token.text92 previous_token = response.token.text93 yield output94 return output95 96 97examples = [98 "X_train, y_train, X_test, y_test = train_test_split(X, y, test_size=0.1)\n\n# Train a logistic regression model, predict the labels on the test set and compute the accuracy score",99 "// Returns every other value in the array as a new array.\nfunction everyOther(arr) {",100 "Poor English: She no went to the market. Corrected English:",101 "def alternating(list1, list2):\n results = []\n for i in range(min(len(list1), len(list2))):\n results.append(list1[i])\n results.append(list2[i])\n if len(list1) > len(list2):\n <FILL_ME>\n else:\n results.extend(list2[i+1:])\n return results",102 "def remove_non_ascii(s: str) -> str:\n \"\"\" <FILL_ME>\nprint(remove_non_ascii('afkdj$$('))",103]104 105 106def process_example(args):107 for x in generate(args):108 pass109 return x110 111 112css = ".generating {visibility: hidden}"113 114monospace_css = """115#q-input textarea {116 font-family: monospace, 'Consolas', Courier, monospace;117}118"""119 120 121css += share_btn_css + monospace_css + ".gradio-container {color: black}"122 123description = """124<div style="text-align: center;">125 <h1> 🦙 Code Llama Playground</h1>126</div>127<div style="text-align: left;">128 <p>This is a demo to generate text and code with the following <a href="https://huggingface.co/codellama/CodeLlama-13b-hf">Code Llama model (13B)</a>. Please note that this model is not designed for instruction purposes but for code completion. If you're looking for instruction or want to chat with a fine-tuned model, you can use <a href="https://huggingface.co/spaces/codellama/codellama-13b-chat">this demo instead</a>. You can learn more about the model in the <a href="https://huggingface.co/blog/codellama/">blog post</a> or <a href="https://huggingface.co/papers/2308.12950">paper</a></p>129 <p>For a chat demo of the largest Code Llama model (34B parameters), you can now <a href="https://huggingface.co/chat/">select Code Llama in Hugging Chat!</a></p>130</div>131"""132 133with gr.Blocks(theme=theme, analytics_enabled=False, css=css) as demo:134 with gr.Column():135 gr.Markdown(description)136 with gr.Row():137 with gr.Column():138 instruction = gr.Textbox(139 placeholder="Enter your code here",140 lines=5,141 label="Input",142 elem_id="q-input",143 )144 submit = gr.Button("Generate", variant="primary")145 output = gr.Code(elem_id="q-output", lines=30, label="Output")146 with gr.Row():147 with gr.Column():148 with gr.Accordion("Advanced settings", open=False):149 with gr.Row():150 column_1, column_2 = gr.Column(), gr.Column()151 with column_1:152 temperature = gr.Slider(153 label="Temperature",154 value=0.1,155 minimum=0.0,156 maximum=1.0,157 step=0.05,158 interactive=True,159 info="Higher values produce more diverse outputs",160 )161 max_new_tokens = gr.Slider(162 label="Max new tokens",163 value=256,164 minimum=0,165 maximum=8192,166 step=64,167 interactive=True,168 info="The maximum numbers of new tokens",169 )170 with column_2:171 top_p = gr.Slider(172 label="Top-p (nucleus sampling)",173 value=0.90,174 minimum=0.0,175 maximum=1,176 step=0.05,177 interactive=True,178 info="Higher values sample more low-probability tokens",179 )180 repetition_penalty = gr.Slider(181 label="Repetition penalty",182 value=1.05,183 minimum=1.0,184 maximum=2.0,185 step=0.05,186 interactive=True,187 info="Penalize repeated tokens",188 )189 190 gr.Examples(191 examples=examples,192 inputs=[instruction],193 cache_examples=False,194 fn=process_example,195 outputs=[output],196 )197 198 submit.click(199 generate,200 inputs=[instruction, temperature, max_new_tokens, top_p, repetition_penalty],201 outputs=[output],202 )203demo.queue(concurrency_count=16).launch(debug=True)