-
Notifications
You must be signed in to change notification settings - Fork 716
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Merge pull request #175 from ChiYeungLaw/main
release wizardcoder-34B-python-v1.0
- Loading branch information
Showing
5 changed files
with
165 additions
and
27 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,114 @@ | ||
| import argparse | ||
| import pprint | ||
| import sys | ||
| import os | ||
| import re | ||
| from tqdm import tqdm | ||
| import torch | ||
| from transformers import LlamaTokenizer, AutoModelForCausalLM, GenerationConfig, BitsAndBytesConfig | ||
| from human_eval.data import write_jsonl, read_problems, stream_jsonl | ||
|
|
||
| from vllm import LLM | ||
| from vllm import SamplingParams | ||
|
|
||
| if torch.cuda.is_available(): | ||
| device = "cuda" | ||
| else: | ||
| device = "cpu" | ||
|
|
||
| try: | ||
| if torch.backends.mps.is_available(): | ||
| device = "mps" | ||
| except: | ||
| pass | ||
|
|
||
| def generate_prompt(input): | ||
| INSTRUCTION = f"""Below is an instruction that describes a task. Write a response that appropriately completes the request. | ||
| ### Instruction: | ||
| Create a Python script for this problem: | ||
| {input} | ||
| ### Response:""" | ||
| return INSTRUCTION | ||
|
|
||
|
|
||
| def main(): | ||
| parser = argparse.ArgumentParser() | ||
|
|
||
| parser.add_argument('--model', type=str, default='bigcode/starcoder', help="") | ||
| parser.add_argument('--lora', type=str, default='bigcode/starcoder', help="") | ||
| parser.add_argument('--output_path', type=str, help="") | ||
| parser.add_argument('--start_index', type=int, default=0, help="") | ||
| parser.add_argument('--end_index', type=int, default=164, help="") | ||
| parser.add_argument('--temperature', type=float, default=0.8, help="") | ||
| parser.add_argument('--N', type=int, default=200, help="") | ||
| parser.add_argument('--max_len', type=int, default=512, help="") | ||
| parser.add_argument('--num_gpus', type=int, default=4, help="") | ||
| parser.add_argument('--decoding_style', type=str, default='sampling', help="") | ||
| parser.add_argument('--num_seqs_per_iter', type=int, default=50, help='') | ||
| parser.add_argument('--overwrite', action='store_true', help='') | ||
|
|
||
| args = parser.parse_args() | ||
|
|
||
| argsdict = vars(args) | ||
| print(pprint.pformat(argsdict)) | ||
|
|
||
| problems = read_problems() | ||
|
|
||
| task_ids = sorted(problems.keys())[args.start_index: args.end_index] | ||
| prompts = [problems[task_id]['prompt'] for task_id in task_ids] | ||
| num_samples = len(prompts) | ||
| print("Number of samples: {}".format(num_samples)) | ||
|
|
||
| llm = LLM(base_model, tensor_parallel_size=args.num_gpus) | ||
| sampling_params = SamplingParams(temperature=args.temperature, top_p=1, max_tokens=args.max_len) | ||
|
|
||
| print(f"Loaded {args.model}.") | ||
| for i in tqdm(range(num_samples), ncols=0, total=num_samples): | ||
| output_file = args.output_path + '/{}.jsonl'.format(args.start_index + i) | ||
|
|
||
| if os.path.exists(output_file) and not args.overwrite: | ||
| print(f'Skip {output_file} as it already exists') | ||
| continue | ||
|
|
||
| prompt = prompts[i].replace(' ', '\t') | ||
| prompt_batch = [generate_prompt(prompt)] | ||
|
|
||
| ids_batch = [task_ids[i]] | ||
| completion_seqs = [] | ||
|
|
||
| if args.decoding_style == 'sampling': | ||
| loops = int(args.N / args.num_seqs_per_iter) | ||
| else: | ||
| loops = 1 | ||
|
|
||
| for _ in tqdm(range(loops), total=loops, leave=False, ncols=0): | ||
|
|
||
| with torch.no_grad(): | ||
| completions = llm.generate(prompt_batch, sampling_params) | ||
| gen_seqs = [completions[0].outputs[0].text] | ||
|
|
||
| if gen_seqs is not None: | ||
| assert len(ids_batch) == 1 | ||
| task_id = ids_batch[0] | ||
|
|
||
| for seq_idx, gen_seq in enumerate(gen_seqs): | ||
| completion_seq = gen_seq.split("### Response:")[-1] | ||
| completion_seq = completion_seq.replace('\t', ' ') | ||
| all_code = gen_seq.replace('\t', ' ') | ||
|
|
||
| completion_seqs.append( | ||
| {'task_id': task_id, | ||
| 'completion': completion_seq, | ||
| 'all_code': all_code, | ||
| } | ||
| ) | ||
|
|
||
| print("Saving results to {}".format(output_file)) | ||
| write_jsonl(output_file, completion_seqs) | ||
|
|
||
|
|
||
| if __name__ == '__main__': | ||
| main() |