|
|
import argparse |
|
|
import time |
|
|
|
|
|
import datasets |
|
|
import torch |
|
|
from transformers import AutoModelForCausalLM, AutoTokenizer |
|
|
from transformers.generation import GenerationConfig |
|
|
|
|
|
|
|
|
MODEL_ID = "Qwen/Qwen3-4B-Instruct-2507" |
|
|
|
|
|
|
|
|
def main(): |
|
|
parser = argparse.ArgumentParser() |
|
|
parser.add_argument("--samples", type=int, default=100, help="Number of prompts to run") |
|
|
parser.add_argument("--batch-size", "-bs", type=int, default=32, help="Static batch size") |
|
|
parser.add_argument("--max-new-tokens", type=int, default=512, help="Max new tokens per request") |
|
|
parser.add_argument("--warmup", type=int, default=1, help="Warmup batches (excluded from timing)") |
|
|
args = parser.parse_args() |
|
|
|
|
|
|
|
|
model = AutoModelForCausalLM.from_pretrained( |
|
|
MODEL_ID, |
|
|
attn_implementation="sdpa", |
|
|
torch_dtype=torch.bfloat16, |
|
|
).cuda().eval() |
|
|
|
|
|
|
|
|
tokenizer = AutoTokenizer.from_pretrained(MODEL_ID, padding_side="left") |
|
|
if tokenizer.pad_token_id is None: |
|
|
tokenizer.pad_token = tokenizer.eos_token |
|
|
|
|
|
|
|
|
dataset = datasets.load_dataset("openai/gsm8k", "socratic", split="test") |
|
|
dataset = dataset.select(range(args.samples)) |
|
|
|
|
|
|
|
|
encoded = tokenizer(list(dataset["question"]), padding=False, truncation=False) |
|
|
inputs = [{"input_ids": ids, "attention_mask": attn} |
|
|
for ids, attn in zip(encoded["input_ids"], encoded["attention_mask"])] |
|
|
|
|
|
|
|
|
gen_cfg = GenerationConfig( |
|
|
do_sample=False, |
|
|
max_new_tokens=args.max_new_tokens, |
|
|
eos_token_id=tokenizer.eos_token_id, |
|
|
pad_token_id=tokenizer.pad_token_id, |
|
|
use_cuda_graph=False, |
|
|
) |
|
|
|
|
|
|
|
|
def make_batch(items): |
|
|
batch = tokenizer.pad(items, padding=True, return_tensors="pt") |
|
|
return {k: v.cuda(non_blocking=True) for k, v in batch.items()} |
|
|
|
|
|
|
|
|
model_inputs = [] |
|
|
if args.warmup > 0: |
|
|
warm = make_batch(inputs[: min(len(inputs), args.batch_size * args.warmup)]) |
|
|
with torch.no_grad(): |
|
|
_ = model.generate(**warm, generation_config=gen_cfg) |
|
|
|
|
|
|
|
|
token_count = 0 |
|
|
bs = args.batch_size |
|
|
start = time.time() |
|
|
with torch.no_grad(): |
|
|
for i in range(0, len(inputs), bs): |
|
|
batch_items = inputs[i : i + bs] |
|
|
batch = make_batch(batch_items) |
|
|
|
|
|
|
|
|
outputs = model.generate(**batch, generation_config=gen_cfg) |
|
|
|
|
|
|
|
|
|
|
|
pad_id = tokenizer.pad_token_id |
|
|
input_lens = batch["attention_mask"].sum(dim=1).tolist() |
|
|
for row, in_len in zip(outputs, input_lens): |
|
|
seq = row.tolist() |
|
|
gen_part = seq[int(in_len):] |
|
|
token_count += sum(1 for t in gen_part if t != pad_id) |
|
|
|
|
|
end = time.time() |
|
|
elapsed = end - start |
|
|
tps = token_count / elapsed if elapsed > 0 else 0.0 |
|
|
|
|
|
print("-" * 20) |
|
|
print("--- Finished Static Batching Benchmark ---\n") |
|
|
print(f"Model: {MODEL_ID}") |
|
|
print(f"Attention: sdpa | Batch size: {args.batch_size} | Samples: {args.samples} | Max new tokens: {args.max_new_tokens}") |
|
|
print(f"Generation time (no warmup): {elapsed:.2f} s for {token_count} generated tokens -> {tps:.2f} tok/s") |
|
|
|
|
|
|
|
|
if __name__ == "__main__": |
|
|
main() |
|
|
|
|
|
|
|
|
|