Remove grok model arch from vllm (#46706)
Signed-off-by: Xianbao QIAN <xianbao.qian@gmail.com>
This commit is contained in:
@@ -1,43 +0,0 @@
|
||||
# SPDX-License-Identifier: Apache-2.0
|
||||
# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
|
||||
import pytest
|
||||
|
||||
from ...utils import dummy_hf_overrides
|
||||
|
||||
MODELS = ["xai-org/grok-2"]
|
||||
|
||||
|
||||
def _grok2_dummy_overrides(hf_config):
|
||||
hf_config = dummy_hf_overrides(hf_config, model_arch="Grok1ForCausalLM")
|
||||
text_config = hf_config.get_text_config()
|
||||
text_config.update(
|
||||
{
|
||||
"hidden_size": 256,
|
||||
"intermediate_size": 512,
|
||||
"moe_intermediate_size": 256,
|
||||
"num_attention_heads": 4,
|
||||
"num_key_value_heads": 2,
|
||||
"head_dim": 64,
|
||||
}
|
||||
)
|
||||
return hf_config
|
||||
|
||||
|
||||
@pytest.mark.parametrize("model", MODELS)
|
||||
def test_dummy_generate(vllm_runner, monkeypatch, model: str) -> None:
|
||||
with monkeypatch.context() as m:
|
||||
m.setenv("VLLM_ALLOW_INSECURE_SERIALIZATION", "1")
|
||||
with vllm_runner(
|
||||
model,
|
||||
load_format="dummy",
|
||||
max_model_len=128,
|
||||
hf_overrides=_grok2_dummy_overrides,
|
||||
enforce_eager=True,
|
||||
) as llm:
|
||||
prompt = "Hello from Grok-2"
|
||||
tokenizer = llm.get_llm().get_tokenizer()
|
||||
prompt_len = len(tokenizer.encode(prompt))
|
||||
outputs = llm.generate_greedy([prompt], max_tokens=1)
|
||||
output_ids, output_str = outputs[0]
|
||||
assert len(output_ids) > prompt_len
|
||||
assert output_str is not None
|
||||
@@ -319,10 +319,6 @@ _TEXT_GENERATION_EXAMPLE_MODELS = {
|
||||
"GraniteMoeSharedForCausalLM": _HfExamplesInfo(
|
||||
"ibm-research/moe-7b-1b-active-shared-experts"
|
||||
),
|
||||
"Grok1ModelForCausalLM": _HfExamplesInfo(
|
||||
"hpcai-tech/grok-1", trust_remote_code=True
|
||||
),
|
||||
"Grok1ForCausalLM": _HfExamplesInfo("xai-org/grok-2", trust_remote_code=True),
|
||||
"HrmTextForCausalLM": _HfExamplesInfo(
|
||||
"sapientinc/HRM-Text-1B",
|
||||
min_transformers_version="5.9.0",
|
||||
|
||||
@@ -9,7 +9,6 @@ from transformers import (
|
||||
)
|
||||
|
||||
from vllm.tokenizers import TokenizerLike, get_tokenizer
|
||||
from vllm.tokenizers.grok2 import Grok2Tokenizer
|
||||
from vllm.tokenizers.hf import HfTokenizer
|
||||
from vllm.tokenizers.mistral import MistralTokenizer
|
||||
|
||||
@@ -35,10 +34,6 @@ def test_tokenizer_like_protocol():
|
||||
assert isinstance(tokenizer, MistralTokenizer)
|
||||
_assert_tokenizer_like(tokenizer)
|
||||
|
||||
tokenizer = get_tokenizer("xai-org/grok-2", tokenizer_mode="grok2")
|
||||
assert isinstance(tokenizer, Grok2Tokenizer)
|
||||
_assert_tokenizer_like(tokenizer)
|
||||
|
||||
tokenizer = get_tokenizer("deepseek-ai/DeepSeek-V3", tokenizer_mode="deepseek_v32")
|
||||
assert isinstance(tokenizer, HfTokenizer)
|
||||
|
||||
|
||||
Reference in New Issue
Block a user