support lm_eval vllm backend

Signed-off-by: xinhe3 <xinhe3@habana.ai>

Author: xinhe3

auto_round/__main__.py

@@ -21,6 +21,12 @@ def run_eval():
     elif "--lmms" in sys.argv:
         sys.argv.remove("--lmms")
         run_lmms()
+    elif "--vllm" in sys.argv:
+        sys.argv.remove("--vllm")
+        from auto_round.script.llm import eval_with_vllm, setup_eval_parser
+
+        args = setup_eval_parser()
+        eval_with_vllm(args)
     else:
         from auto_round.script.llm import setup_eval_parser
 

auto_round/script/llm.py

@@ -42,6 +42,8 @@
     set_cuda_visible_devices,
 )
 
+os.environ["TOKENIZERS_PARALLELISM"] = "false"
+
 
 class BasicArgumentParser(argparse.ArgumentParser):
 
@@ -322,6 +324,28 @@ def __init__(self, *args, **kwargs):
             help="Limit the number of examples per task. "
             "If <1, limit is a percentage of the total number of examples.",
         )
+        # vllm related arguments
+        self.add_argument("--revision", default=None, type=str, help="model revision for vllm")
+        self.add_argument("--tokenizer", default=None, type=str, help="tokenizer to use with vllm")
+        self.add_argument(
+            "--tokenizer_mode", default="auto", type=str, help="tokenizer mode for vllm (e.g. auto/fast/slow)"
+        )
+        self.add_argument("--tokenizer_revision", default=None, type=str, help="tokenizer revision for vllm")
+        self.add_argument("--add_bos_token", action="store_true", help="add BOS token when using vllm")
+        self.add_argument("--prefix_token_id", default=None, type=int, help="prefix token id for vllm")
+        self.add_argument("--tensor_parallel_size", default=1, type=int, help="tensor parallel size for vllm")
+        self.add_argument("--data_parallel_size", default=1, type=int, help="data parallel size for vllm")
+        self.add_argument("--quantization", default=None, type=str, help="quantization setting for vllm")
+        self.add_argument("--max_gen_toks", default=256, type=int, help="max generation tokens for vllm")
+        self.add_argument("--swap_space", default=4, type=float, help="swap space (GB) for vllm")
+        self.add_argument("--max_batch_size", default=None, type=int, help="max batch size for vllm")
+        self.add_argument("--max_length", default=None, type=int, help="max generation length for vllm")
+        self.add_argument("--max_model_len", default=None, type=int, help="maximum model sequence length for vllm")
+        self.add_argument("--seed", default=1234, type=int, help="random seed")
+        self.add_argument(
+            "--gpu_memory_utilization", default=0.9, type=float, help="target GPU memory utilization for vllm"
+        )
+        self.add_argument("--lora_local_path", default=None, type=str, help="local LoRA path for vllm")
 
 
 def setup_parser():
@@ -786,15 +810,16 @@ def eval(args):
     if (batch_size := args.eval_bs) is None:
         batch_size = "auto:8"
     is_gguf_file = False
-    if os.path.isfile(args.model) and args.model.endswith(".gguf"):
-        is_gguf_file = True
-        gguf_file = os.path.basename(args.model)
-        model = os.path.dirname(args.model)
-    else:
-        for file in os.listdir(args.model):
-            if file.endswith(".gguf"):
-                is_gguf_file = True
-                gguf_file = file
+    if os.path.exists(args.model):
+        if os.path.isfile(args.model) and args.model.endswith(".gguf"):
+            is_gguf_file = True
+            gguf_file = os.path.basename(args.model)
+            model = os.path.dirname(args.model)
+        else:
+            for file in os.listdir(args.model):
+                if file.endswith(".gguf"):
+                    is_gguf_file = True
+                    gguf_file = file
     eval_model_dtype = get_model_dtype(args.eval_model_dtype)
     if is_gguf_file:
         import torch
@@ -949,3 +974,51 @@ def eval_task_by_task(
         print(make_table(res_all))
 
     print("total eval time:", time.time() - st)
+
+
+def eval_with_vllm(args):
+    import time
+
+    st = time.time()
+
+    from lm_eval import evaluator  # pylint: disable=E0401
+    from lm_eval.models.vllm_causallms import VLLM  # pylint: disable=E0401
+    from lm_eval.utils import make_table  # pylint: disable=E0401
+
+    device_str, _ = get_device_and_parallelism(args.device)
+    eval_model_dtype = get_model_dtype(args.eval_model_dtype, "auto")
+    if (batch_size := args.eval_bs) is None:
+        batch_size = "auto:8"
+
+    vllm_lm = VLLM(
+        pretrained=args.model,
+        dtype=eval_model_dtype,
+        revision=args.revision,
+        trust_remote_code=not args.disable_trust_remote_code,
+        tokenizer=args.tokenizer,
+        tokenizer_mode=args.tokenizer_mode,
+        tokenizer_revision=args.tokenizer_revision,
+        add_bos_token=args.add_bos_token,
+        prefix_token_id=args.prefix_token_id,
+        tensor_parallel_size=args.tensor_parallel_size,
+        quantization=args.quantization,
+        max_gen_toks=args.max_gen_toks,
+        swap_space=args.swap_space,
+        batch_size=batch_size,
+        max_batch_size=args.max_batch_size,
+        max_length=args.max_length,
+        max_model_len=args.max_model_len,
+        seed=args.seed,
+        gpu_memory_utilization=args.gpu_memory_utilization,
+        device=device_str,
+        data_parallel_size=args.data_parallel_size,
+        lora_local_path=args.lora_local_path,
+    )
+    res = evaluator.simple_evaluate(
+        model=vllm_lm,
+        tasks=args.tasks,
+        limit=args.limit,
+    )
+
+    print(make_table(res))
+    print("evaluation running time=%ds" % (time.time() - st))

docs/step_by_step.md

@@ -617,6 +617,8 @@ If not explicitly specify '--task', the default value will be used (typically co
   ~~~
   The last format will be used in evaluation if multiple formats have been exported.
 
+Note: To use the vllm backend, please add `--vllm` into the upper command.
+
 ###  Eval the Quantized model
 
 - AutoRound format

test/test_cuda/test_vllm.py

@@ -7,6 +7,10 @@
 Run `pytest test/test_cuda/test_vllm.py`.
 """
 
+import os
+import shutil
+import subprocess
+
 import pytest
 from vllm import LLM, SamplingParams
 from vllm.platforms import current_platform
@@ -43,3 +47,30 @@ def test_auto_round(model):
         generated_text = output.outputs[0].text
         if "France" in prompt:
             assert "Paris" in generated_text
+
+
+@pytest.mark.parametrize("model", MODELS)
+def test_vllm_lm_eval(model):
+    if shutil.which("auto-round") is None:
+        pytest.skip("auto-round CLI not available")
+
+    env = os.environ.copy()
+    env["VLLM_SKIP_WARMUP"] = "true"
+    env["VLLM_WORKER_MULTIPROC_METHOD"] = "spawn"
+
+    cmd = [
+        "auto-round",
+        "--model",
+        model,
+        "--eval",
+        "--tasks",
+        "lambada_openai",
+        "--eval_bs",
+        "8",
+        "--limit",
+        "10",
+        "--vllm",
+    ]
+
+    proc = subprocess.run(cmd, env=env, stdout=subprocess.PIPE, stderr=subprocess.STDOUT, text=True)
+    assert proc.returncode == 0, f"auto-round failed (rc={proc.returncode}):\n{proc.stdout}"