File size: 1,890 Bytes
cf3b6c5 3dc6935 cf3b6c5 32e9007 cf3b6c5 32e9007 cf3b6c5 32e9007 cf3b6c5 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 |
import streamlit as st
import datasets
humaneval_v_data = datasets.load_from_disk("humaneval_v_test_hf")
st.set_page_config(layout="wide", page_title="HumanEval-V Viewer")
st.markdown("> <i>This is a viewer for the HumanEval-V benchmark, which consists of 108 coding tasks. Use the navigation buttons or enter an index to browse the tasks. If you encounter any issues, we encourage you to start a discussion [here](https://huggingface.co/datasets/HumanEval-V/HumanEval-V-Benchmark/discussions)</i>.", unsafe_allow_html=True)
st.markdown("---")
max_index = 108
# Initialize session state for index if not present
if 'index' not in st.session_state:
st.session_state.index = 1
buttons = st.columns([2, 1.1, 5.9])
with buttons[0]:
st.markdown("# HumanEval-V Viewer")
with buttons[1]:
# Number input for navigation
index_input = st.number_input(
f"Go to index (1-{max_index}):",
min_value=1,
max_value=108,
value=st.session_state.index,
key="index_input",
help="Enter an index and jump to that index.",
step=1
)
coding_task = humaneval_v_data[index_input-1]
qid = coding_task["qid"]
image = coding_task["image"]
function_signature = coding_task["function_signature"]
ground_truth = coding_task["ground_truth_solution"]
test_script = coding_task["test_script"]
upper_columns = st.columns([2, 7])
with upper_columns[0]:
st.markdown(f"### Question ID: {qid}")
st.image(image, use_column_width=True)
st.markdown("---")
with upper_columns[1]:
st.markdown(f"### Function Signature:")
st.markdown(f"")
st.markdown(f"""```python
{function_signature}
```""")
st.markdown(f"### Test Script:")
st.markdown(f"")
st.markdown(f"""```python
{test_script}
```""")
st.markdown(f"### Ground Truth Solution:")
st.markdown(f"")
st.markdown(f"""```python
{ground_truth}
```""") |