neo4j/text2cypher-gemma-2-9b-it-finetuned-2024v1

Model Card for Model ID

Model Details

Model Description

This model serves as a demonstration of how fine-tuning foundational models using the Neo4j-Text2Cypher(2024) Dataset (link) can enhance performance on the Text2Cypher task.
Please note, this is part of ongoing research and exploration, aimed at highlighting the dataset's potential rather than a production-ready solution.

Base model: google/gemma-2-9b-it
Dataset: neo4j/text2cypher-2024v1

An overview of the finetuned models and benchmarking results are shared at Link1 and Link2

Have ideas or insights? Contact us: Neo4j/Team-GenAI

Bias, Risks, and Limitations

We need to be cautious about a few risks:

• In our evaluation setup, the training and test sets come from the same data distribution (sampled from a larger dataset). If the data distribution changes, the results may not follow the same pattern.
• The datasets used were gathered from publicly available sources. Over time, foundational models may access both the training and test sets, potentially achieving similar or even better results.

Also check the related blogpost:Link

Training Details

Training Procedure

Used RunPod with following setup:

• 1 x A100 PCIe
• 31 vCPU 117 GB RAM
• runpod/pytorch:2.4.0-py3.11-cuda12.4.1-devel-ubuntu22.04
• On-Demand - Secure Cloud
• 60 GB Disk
• 60 GB Pod Volume

  Training Hyperparameters

  • lora_config = LoraConfig( r=64, lora_alpha=64, target_modules=target_modules, lora_dropout=0.05, bias="none", task_type="CAUSAL_LM", )
  • sft_config = SFTConfig( dataset_text_field=dataset_text_field, per_device_train_batch_size=4, gradient_accumulation_steps=8, dataset_num_proc=16, max_seq_length=1600, logging_dir="./logs", num_train_epochs=1, learning_rate=2e-5, save_steps=5, save_total_limit=1, logging_steps=5, output_dir="outputs", optim="paged_adamw_8bit", save_strategy="steps", )
  • bnb_config = BitsAndBytesConfig( load_in_4bit=True, bnb_4bit_use_double_quant=True, bnb_4bit_quant_type="nf4", bnb_4bit_compute_dtype=torch.bfloat16, )

  Framework versions

  • PEFT 0.12.0

  Example Cypher generation

  from peft import PeftModel, PeftConfig
  import torch
  from transformers import (
      AutoModelForCausalLM,
      AutoTokenizer,
      BitsAndBytesConfig,
  )
  
  
  instruction = (
      "Generate Cypher statement to query a graph database. "
      "Use only the provided relationship types and properties in the schema. \n"
      "Schema: {schema} \n Question: {question}  \n Cypher output: "
  )
  
  
  
  def prepare_chat_prompt(question, schema) -> list[dict]:
      chat = [
          {
              "role": "user",
              "content": instruction.format(
                  schema=schema, question=question
              ),
          }
      ]
      return chat
  
  def _postprocess_output_cypher(output_cypher: str) -> str:
      # Remove any explanation. E.g.  MATCH...\n\n**Explanation:**\n\n -> MATCH...
      # Remove cypher indicator. E.g.```cypher\nMATCH...```` --> MATCH...
      # Note: Possible to have both:
      #   E.g. ```cypher\nMATCH...````\n\n**Explanation:**\n\n --> MATCH...
      partition_by = "**Explanation:**"
      output_cypher, _, _ = output_cypher.partition(partition_by)
      output_cypher = output_cypher.strip("`\n")
      output_cypher = output_cypher.lstrip("cypher\n")
      output_cypher = output_cypher.strip("`\n ")
      return output_cypher
  
  # Model
  model_name = "neo4j/text2cypher-gemma-2-9b-it-finetuned-2024v1"
  bnb_config = BitsAndBytesConfig(
      load_in_4bit=True,
      bnb_4bit_use_double_quant=True,
      bnb_4bit_quant_type="nf4",
      bnb_4bit_compute_dtype=torch.bfloat16,
  )
  tokenizer = AutoTokenizer.from_pretrained(model_name)
  model = AutoModelForCausalLM.from_pretrained(
      model_name,
      quantization_config=bnb_config,
      torch_dtype=torch.bfloat16,
      attn_implementation="eager",
      low_cpu_mem_usage=True,
  )
  
  # Question
  question = "What are the movies of Tom Hanks?"
  schema = "(:Actor)-[:ActedIn]->(:Movie)"
  new_message = prepare_chat_prompt(question=question, schema=schema)
  prompt = tokenizer.apply_chat_template(new_message, add_generation_prompt=True, tokenize=False)
  inputs = tokenizer(prompt, return_tensors="pt", padding=True)
  
  # Any other parameters
  model_generate_parameters = {
      "top_p": 0.9,
      "temperature": 0.2,
      "max_new_tokens": 512,
      "do_sample": True,
      "pad_token_id": tokenizer.eos_token_id,
  }
  
  inputs.to(model.device)
  model.eval()
  with torch.no_grad():
      tokens = model.generate(**inputs, **model_generate_parameters)
      tokens = tokens[:, inputs.input_ids.shape[1] :]
      raw_outputs = tokenizer.batch_decode(tokens, skip_special_tokens=True)
      outputs = [_postprocess_output_cypher(output) for output in raw_outputs]
      
  print(outputs)
  > ["MATCH (a:Actor {Name: 'Tom Hanks'})-[:ActedIn]->(m:Movie) RETURN m"]