app.py

import gradio as gr
import os

from peft import PeftConfig, PeftModel
from transformers import AutoModelForCausalLM, AutoTokenizer

base_model_name = "google/gemma-7b"
#adapter_model_name = "samidh/cope-g2b-2c-hs-skr-s1.5.9-sx-sk-s5.d25"
#adapter_model_name = "samidh/cope-g2b-2c-hs-skr-s1.5.9-sx-sk-s1.5.l1e4-e10-d25"
#adapter_model_name = "samidh/cope-g2b-2c-hs-s1.f5.9.l5e5-e10-d25-r8"
#adapter_model_name = "samidh/cope-g2b-2c-hs.s1.5.9-sx.s1.5.9o-hr.s5-sh.s5.l1e4-e10-d25-r8"
#adapter_model_name = "samidh/cope-ap-g2b-2c-hs.s1.5.9-sx.s1.5.9o-vl.s1.5.9-hr.s5-sh.s5.l1e4-e5-d25-r8"
#adapter_model_name = "samidh/cope-g2b-2c-hs.s1.5pcf.9.l5e5-e10-d25-r8"
#adapter_model_name = "samidh/cope-g7bq-2c-hs.s1.5fpc.9-sx.s1.5.9o-VL.s1.5.9-HR.s5-SH.s5-l1e4-e5-d25-r8"
adapter_model_name = "samidh/cope-g7bq-2c-hs.s1.5fpc.9-sx.s1.5.9o-VL.s1.5.9-HR.s5-SH.s5-l5e5-e3-d25-r8"

model = AutoModelForCausalLM.from_pretrained(base_model_name, token=os.environ['HF_TOKEN'])
model = PeftModel.from_pretrained(model, adapter_model_name, token=os.environ['HF_TOKEN'])
model.merge_and_unload()

tokenizer = AutoTokenizer.from_pretrained(base_model_name)

#inputs = tokenizer.encode("This movie was really", return_tensors="pt")

#outputs = model.generate(inputs, max_new_tokens=1)
#print(tokenizer.decode(outputs[0], skip_special_tokens=True)[-1])

#gr.load("models/samidh/cope-gemma2b-hs-2c-skr-s1.5.9.d25", hf_token=os.environ['HF_TOKEN']).launch()

PROMPT = """
INSTRUCTIONS
============
Examine the given POLICY and determine if the given CONTENT meets the criteria for ANY of the LABELS. Answer "1" if yes, and "0" if no.

POLICY
======

{policy}


CONTENT
=======

{content}


ANSWER
======

"""

DEFAULT_POLICY = """# Policy

## Overview

This policy guides the classification of content related to expressions that may be sensitive or controversial, aiming to foster a community where open dialogue is encouraged while maintaining respect for all individuals.

## Definition of Terms

- Protected Classes: Individuals or groups of individuals defined by sharing a race, ethnicity, religion, disability, sex, gender identity, or sexual orientation. The scope is narrowed to focus on characteristics commonly recognized in anti-discrimination laws. Religions themselves, as concepts ARE NOT protected - only their members are.
- Non-Protected Classes: Groups defined by other characteristics, such as age, socioeconomic status, and education, which are typically excluded partially or entirely from anti-discrimination laws.

## Interpretation of Language

## Definition of Labels

### (HS): Hate Speech

#### Includes

- Explicit Violence Advocacy: Direct and clear threats of or praise for violence or harm against individuals or groups based on their membership in a particular protected class. Stating an intention to defend against aggression DOES NOT qualify
- Intentional Dehumanization: Statements that clearly depict individuals or groups as inherently ""other"", alien, animalistic, unintelligent, immoral, unclean, or less-than-fully-human based on their membership in a particular protected class in a way that justifies harm or discrimination.
- Targeted Use of Derogatory Slurs: Targeting another person or group of people using a one-word name for a particular protected class that has an inherent negative connotation (e.g. Nigger, Kike, Cunt, Retard). Multi-word terms are never slurs.
- Explicit Discrimination Advocacy: Direct and clear calls for exclusion, segregation, or discrimination against individuals or groups based on their membership in a particular protected class, with a clear intent to promote inequality.
- Direct Hateful Insults: Content that directly addresses another person or group of people the second person (e.g. ""You over there"") and insults them based on their membership in a particular protected class

#### Excludes

- Artistic and Educational Content: Expressions intended for artistic, educational, or documentary purposes that discuss sensitive topics but do not advocate for violence or discrimination against individuals or groups based on their membership in a particular protected class.
- Political and Social Commentary: Commentary on political issues, social issues, and political ideologies that does not directly incite violence or discrimination against individuals or groups based on their membership in a particular protected class.
- Rebutting Hateful Language: Content that rebuts, condemns, questions, criticizes, or mocks a different person's hateful language or ideas OR that insults the person advocating those hateful
- Quoting Hateful Language: Content in which the author quotes someone else's hateful language or ideas while discussing, explaining, or neutrally factually presenting those ideas.
- Describing Sectarian Violence: Content that describes, but does not endorse or praise, violent physical injury against a specifically named race, ethnicity, nationality, sexual orientation, or religious community by another specifically named race, ethnicity, nationality, sexual orientation, or religious community
"""

DEFAULT_CONTENT = "LLMs steal our jobs."

# Function to make predictions
def predict(content, policy):
    input_text = PROMPT.format(policy=policy, content=content)
    inputs = tokenizer.encode(input_text, return_tensors="pt")
    outputs = model.generate(inputs, max_new_tokens=1)
    decoded_output = tokenizer.decode(outputs[0], skip_special_tokens=True)
    if int(decoded_output[-1]) == 0:
        return f'NON-Violating ({decoded_output[-1]})'
    else:
        return f'VIOLATING ({decoded_output[-1]})'


with gr.Blocks() as iface:
    gr.Markdown("# CoPE Alpha Preview")
    gr.Markdown("See if the given content violates your given policy.")
    
    with gr.Row():
        content_input = gr.Textbox(label="Content", lines=2, value=DEFAULT_CONTENT)
        policy_input = gr.Textbox(label="Policy", lines=10, value=DEFAULT_POLICY)
    
    submit_btn = gr.Button("Submit")
    output = gr.Label(label="Label")
    
    gr.Markdown("""
    ## About CoPE
    
    CoPE (the COntent Policy Evaluation engine) is a small language model capable of accurate content policy labeling. This is a **preview** of our alpha release and is strictly for **research** purposes. This should **NOT** be used for any production use cases.
    
    ## How to Use
                                                                                                                                                                                                                                                                       
    1. Enter your content in the "Content" box.
    2. Specify your policy in the "Policy" box.
    3. Click "Submit" to see the results.

    **Note**: Inference times are **very slow** (30-45 seconds) since this is built on dev infra and not yet optimized for live systems. Please be patient while testing!
    
    ## More Info

    - [Give us feedback](https://forms.gle/BHpt6BpH2utaf4ez9) to help us improve
    - [Read our FAQ](https://docs.google.com/document/d/1Cp3GJ5k2I-xWZ4GK9WI7Xv8TpKdHmjJ3E9RbzP5Cc_Y/edit) to learn more about CoPE
    - [Join our mailing list](https://forms.gle/PCABrZdhTuXE9w9ZA) to keep in touch
    """)
    
    submit_btn.click(predict, inputs=[content_input, policy_input], outputs=output)

# Launch the app
iface.launch()