Skip to content

Commit

Permalink
update
Browse files Browse the repository at this point in the history
  • Loading branch information
ChiYeungLaw authored Aug 30, 2023
1 parent 6e86dfa commit 5d475ec
Showing 1 changed file with 3 additions and 0 deletions.
3 changes: 3 additions & 0 deletions WizardCoder/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -257,6 +257,8 @@ evaluate_functional_correctness ${output_path}.jsonl

### How to Reproduce the 73.2 Pass@1 on HumanEval with Greedy Decoding?

❗❗❗**This performance is 100% reproducible!**

- Step 1: Setup the environment
```bash
conda create -n eval python=3.10
Expand Down Expand Up @@ -309,6 +311,7 @@ CUDA_VISIBLE_DEVICES=0,1,2,3 python humaneval_gen_vllm.py --model ${model} \
```

- Step 5: Get the score

use the code `WizardLM/blob/main/WizardCoder/src/process_humaneval.py` to get the score.
```bash
output_path=preds/T0.0_N1
Expand Down

0 comments on commit 5d475ec

Please sign in to comment.