cahlen/cdg-noir-interrogation-ex10

Jack 'Hardcase' Harris & Ralph 'Ricky' Reynolds: The informant's dubious alibi for the night of the theft - Generated by Conversation Dataset Generator

This dataset was generated using the Conversation Dataset Generator script available at https://github.com/cahlen/conversation-dataset-generator.

Generation Parameters

• Number of Conversations Requested: 10
• Number of Conversations Successfully Generated: 10
• Total Turns: 106
• Model ID: meta-llama/Meta-Llama-3-8B-Instruct
• Creative Brief Used: A hardboiled detective interrogating a nervous informant about a stolen artifact in a smoky, rain-slicked alley.
• Persona 1 Search Term Used: N/A
• Persona 2 Search Term Used: N/A

Note: When using Creative Brief mode, Persona 1 (Jack 'Hardcase' Harris) and Persona 2 (Ralph 'Ricky' Reynolds) were determined once from the brief and kept consistent. The topic, scenario, and style may have been automatically varied for each example based on the original brief, aiming for diversity. The parameters shown below reflect the configuration used for the last successfully generated example.

• Topic: The informant's dubious alibi for the night of the theft
• Scenario: In a cramped, smoke-filled backroom of a seedy dive bar
• Style: Confrontational, with a focus on the detective's aggressive questioning and the informant's growing desperation
• Included Points: pocket watch, thief, alibi, witness, rain, cigarette smoke, alleyway, nervous laughter

Personas

Jack 'Hardcase' Harris Description: A grizzled, hard-boiled detective with a sharp tongue and a penchant for sarcasm. Speaks in a gravelly voice, often with a hint of menace. Has a tendency to interrupt and lean in close. Has a cigarette perpetually dangling from his lip. -> maps to role: human

Ralph 'Ricky' Reynolds Description: A nervous, fidgety informant with a habit of stammering and sweating. Speaks quickly and hesitantly, often avoiding eye contact. Has a tendency to glance around nervously, as if worried about being overheard. -> maps to role: gpt

Usage

To use this dataset:

1. Clone the repository:

git lfs install
git clone https://huggingface.co/datasets/cahlen/cdg-noir-interrogation-ex10

2. Load in Python:

from datasets import load_dataset

dataset = load_dataset("cahlen/cdg-noir-interrogation-ex10")

# Access the data (e.g., the training split)
print(dataset['train'][0])

LoRA Training Example (Basic)

Below is a basic example of how you might use this dataset to fine-tune a small model like google/gemma-2b-it using LoRA with the PEFT and TRL libraries.

Note: This requires installing additional libraries: pip install -U transformers datasets accelerate peft trl bitsandbytes torch

import torch
from datasets import load_dataset
from peft import LoraConfig, get_peft_model, prepare_model_for_kbit_training
from transformers import AutoModelForCausalLM, AutoTokenizer, TrainingArguments, BitsAndBytesConfig
from trl import SFTTrainer

# 1. Load the dataset
dataset_id = "cahlen/cdg-noir-interrogation-ex10"
dataset = load_dataset(dataset_id)

# 2. Load Base Model & Tokenizer (using a small model like Gemma 2B)
model_id = "google/gemma-2b-it"

# Quantization Config (optional, for efficiency)
quantization_config = BitsAndBytesConfig(
    load_in_4bit=True,
    bnb_4bit_quant_type="nf4",
    bnb_4bit_compute_dtype=torch.bfloat16 # or torch.float16
)

# Tokenizer
tokenizer = AutoTokenizer.from_pretrained(model_id, trust_remote_code=True)
# Set padding token if necessary (Gemma's is <pad>)
if tokenizer.pad_token is None:
    tokenizer.pad_token = tokenizer.eos_token
    tokenizer.pad_token_id = tokenizer.eos_token_id

# Model
model = AutoModelForCausalLM.from_pretrained(
    model_id,
    quantization_config=quantization_config,
    device_map="auto", # Automatically place model shards
    trust_remote_code=True
)

# Prepare model for k-bit training if using quantization
model = prepare_model_for_kbit_training(model)

# 3. LoRA Configuration
lora_config = LoraConfig(
    r=8, # Rank
    lora_alpha=16,
    target_modules=["q_proj", "k_proj", "v_proj", "o_proj"], # Adjust based on model architecture
    lora_dropout=0.05,
    bias="none",
    task_type="CAUSAL_LM"
)

model = get_peft_model(model, lora_config)
model.print_trainable_parameters()

# 4. Training Arguments (minimal example)
training_args = TrainingArguments(
    output_dir="./lora-adapter-Jack 'Hardcase' Harris-Ralph 'Ricky' Reynolds", # Choose a directory
    per_device_train_batch_size=1,
    gradient_accumulation_steps=4,
    learning_rate=2e-4,
    num_train_epochs=1, # Use 1 epoch for a quick demo
    logging_steps=10,
    save_steps=50, # Save adapter periodically
    fp16=False, # Use bf16 if available, otherwise fp16
    bf16=torch.cuda.is_bf16_supported(),
    optim="paged_adamw_8bit", # Use paged optimizer for efficiency
    report_to="none" # Disable wandb/tensorboard for simple example
)

# 5. Create SFTTrainer
trainer = SFTTrainer(
    model=model,
    train_dataset=dataset['train'], # Assumes 'train' split exists
    peft_config=lora_config,
    tokenizer=tokenizer,
    args=training_args,
    max_seq_length=512, # Adjust as needed
    dataset_text_field="content", # Use content field directly 
    packing=True, # Pack sequences for efficiency
)

# 6. Train
print("Starting LoRA training...")
trainer.train()

### 7. Save the LoRA adapter
# Use a fixed string for the example output directory
trainer.save_model("./lora-adapter-output-directory") 
print(f"LoRA adapter saved to ./lora-adapter-output-directory")

Dataset Format (JSON Lines source)

Each row in the dataset contains the following keys:

• conversation_id: Unique identifier for the conversation
• turn_number: The sequential number of the turn within a conversation
• role: Either 'human' or 'gpt' indicating who is speaking
• speaker_name: The actual name of the speaker (e.g., 'Jack 'Hardcase' Harris' or 'Ralph 'Ricky' Reynolds')
• topic: The conversation topic
• scenario: The scenario in which the conversation takes place
• style: The stylistic direction for the conversation
• include_points: Specific points to include in the conversation
• content: The actual text content of the turn