metadata
library_name: transformers
metrics:
- bleu: 0.67
- chrf: 0.73
Model Card for Model ID
This is the Gemma-2b-IT model fine-tuned for the Python code generation task.
Model Details
Model Description
This is the model card of a 🤗 transformers model that has been pushed on the Hub. This model card has been automatically generated.
- Developed by: Mohammed Ashraf
- Model type: google/gemma-2b
- Finetuned from model [optional]: google/gemma-2b-it
Uses
Direct Use
Use this model to generate Python code."
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM
model_id = "mrSoul7766/gemma-2b-it-python-code-gen-adapter"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id)
text = """<start_of_turn>Convert JSON data to a CSV file<end_of_turn>
<start_of_turn>model"""
#device = "cuda:0"
inputs = tokenizer(text, return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=200)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))
Out-of-Scope Use
This model is trained on very basic Python code, so it might not be able to handle complex code.
How to Get Started with the Model
Use the code below to get started with the model.
[More Information Needed]
Training Details
Training Data
Fine-tuning Data: flytech/python-codes-25k
Training Procedure
Training Hyperparameters
- Training regime: fp16
- learning_rate: 2e-4
Evaluation
Testing Data & Metrics
Testing Data
iamtarun/python_code_instructions_18k_alpaca
Metrics
- chrf: 0.73
- codeblue: 0.67
- codeblue_ngram: 0.53
Results
[More Information Needed]
import json
import pandas as pd
# Load the JSON data
with open('data.json', 'r') as f:
data = json.load(f)
# Create the DataFrame
df = pd.DataFrame(data)
Summary
Environmental Impact
Carbon emissions can be estimated using the Machine Learning Impact calculator presented in Lacoste et al. (2019).
- Hardware Type: H100
- Hours used: 30 minutes
- Cloud Provider: Google-cloud
Technical Specifications [optional]
Model Architecture and Objective
Hardware
- Hardware Type: H100
- Hours used: 30 minutes
- Cloud Provider: Google-cloud
Software
- bitsandbytes==0.42.0
- peft==0.8.2
- trl==0.7.10
- accelerate==0.27.1
- datasets==2.17.0
- transformers==4.38.0