biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

Update transformers to v4.55 (#21931)

Browse Source 
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
Signed-off-by: DarkLight1337 <tlleungac@connect.ust.hk>
Signed-off-by: Isotr0py <2037008807@qq.com>
Signed-off-by: isotr0py <2037008807@qq.com>
Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Co-authored-by: DarkLight1337 <tlleungac@connect.ust.hk>
Co-authored-by: Isotr0py <2037008807@qq.com>
Co-authored-by: Isotr0py <mozf@mail2.sysu.edu.cn>
Co-authored-by: Woosuk Kwon <woosuk.kwon@berkeley.edu>
This commit is contained in:
Harry Mellor
2025-08-06 06:56:14 +01:00
committed by GitHub
parent 6e20924350
commit 796bae07c5
13 changed files with 235 additions and 39 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
tests/models/multimodal/generation/test_common.py
View File

@@ -337,6 +337,10 @@ VLM_TEST_SETTINGS = {
vllm_output_post_proc=model_utils.fuyu_vllm_to_hf_output,
num_logprobs=10,
image_size_factors=[(), (0.25,), (0.25, 0.25, 0.25), (0.25, 0.2, 0.15)],
# FIXME(Isotr0py): This model is broken in Transformers v4.54.1, we
# should enable this again after the fix is released:
# https://github.com/huggingface/transformers/pull/39915
marks=[pytest.mark.skip("HF model is broken")],
),
"gemma3": VLMTestInfo(
models=["google/gemma-3-4b-it"],

24
tests/models/registry.py
View File

@@ -179,8 +179,7 @@ _TEXT_GENERATION_EXAMPLE_MODELS = {
min_transformers_version="4.54"),
"Fairseq2LlamaForCausalLM": _HfExamplesInfo("mgleize/fairseq2-dummy-Llama-3.2-1B"), # noqa: E501
"FalconForCausalLM": _HfExamplesInfo("tiiuae/falcon-7b"),
"FalconH1ForCausalLM":_HfExamplesInfo("tiiuae/Falcon-H1-0.5B-Base",
min_transformers_version="4.53"),
"FalconH1ForCausalLM":_HfExamplesInfo("tiiuae/Falcon-H1-0.5B-Base"),
"GemmaForCausalLM": _HfExamplesInfo("google/gemma-1.1-2b-it"),
"Gemma2ForCausalLM": _HfExamplesInfo("google/gemma-2-9b"),
"Gemma3ForCausalLM": _HfExamplesInfo("google/gemma-3-1b-it"),
@@ -223,7 +222,10 @@ _TEXT_GENERATION_EXAMPLE_MODELS = {
trust_remote_code=True),
"JAISLMHeadModel": _HfExamplesInfo("inceptionai/jais-13b-chat"),
"JambaForCausalLM": _HfExamplesInfo("ai21labs/AI21-Jamba-1.5-Mini",
extras={"tiny": "ai21labs/Jamba-tiny-dev"}), # noqa: E501
extras={
"tiny": "ai21labs/Jamba-tiny-dev",
"random": "ai21labs/Jamba-tiny-random", # noqa: E501
}),
"LlamaForCausalLM": _HfExamplesInfo("meta-llama/Llama-3.2-1B-Instruct",
extras={"guard": "meta-llama/Llama-Guard-3-1B", # noqa: E501
"hermes": "NousResearch/Hermes-3-Llama-3.1-8B", # noqa: E501
@@ -239,8 +241,7 @@ _TEXT_GENERATION_EXAMPLE_MODELS = {
trust_remote_code=True),
"MiniCPM3ForCausalLM": _HfExamplesInfo("openbmb/MiniCPM3-4B",
trust_remote_code=True),
"MiniMaxForCausalLM": _HfExamplesInfo("MiniMaxAI/MiniMax-Text-01-hf",
min_transformers_version="4.53"),
"MiniMaxForCausalLM": _HfExamplesInfo("MiniMaxAI/MiniMax-Text-01-hf"),
"MiniMaxText01ForCausalLM": _HfExamplesInfo("MiniMaxAI/MiniMax-Text-01",
trust_remote_code=True,
revision="a59aa9cbc53b9fb8742ca4e9e1531b9802b6fdc3"), # noqa: E501
@@ -272,6 +273,8 @@ _TEXT_GENERATION_EXAMPLE_MODELS = {
"PhiMoEForCausalLM": _HfExamplesInfo("microsoft/Phi-3.5-MoE-instruct",
trust_remote_code=True),
"Plamo2ForCausalLM": _HfExamplesInfo("pfnet/plamo-2-1b",
max_transformers_version="4.53",
transformers_version_reason="vLLM impl inherits PreTrainedModel and clashes with get_input_embeddings", # noqa: E501
trust_remote_code=True),
"QWenLMHeadModel": _HfExamplesInfo("Qwen/Qwen-7B-Chat",
trust_remote_code=True),
@@ -299,8 +302,7 @@ _TEXT_GENERATION_EXAMPLE_MODELS = {
"Zamba2ForCausalLM": _HfExamplesInfo("Zyphra/Zamba2-7B-instruct"),
"MiMoForCausalLM": _HfExamplesInfo("XiaomiMiMo/MiMo-7B-RL",
trust_remote_code=True),
"Dots1ForCausalLM": _HfExamplesInfo("rednote-hilab/dots.llm1.inst",
min_transformers_version="4.53"),
"Dots1ForCausalLM": _HfExamplesInfo("rednote-hilab/dots.llm1.inst"),
# [Encoder-decoder]
"BartModel": _HfExamplesInfo("facebook/bart-base"),
"BartForConditionalGeneration": _HfExamplesInfo("facebook/bart-large-cnn"),
@@ -326,8 +328,12 @@ _EMBEDDING_EXAMPLE_MODELS = {
"NomicBertModel": _HfExamplesInfo("nomic-ai/nomic-embed-text-v2-moe",
trust_remote_code=True, v0_only=True), # noqa: E501
"Qwen2Model": _HfExamplesInfo("ssmits/Qwen2-7B-Instruct-embed-base"),
"Qwen2ForRewardModel": _HfExamplesInfo("Qwen/Qwen2.5-Math-RM-72B"),
"Qwen2ForProcessRewardModel": _HfExamplesInfo("Qwen/Qwen2.5-Math-PRM-7B"),
"Qwen2ForRewardModel": _HfExamplesInfo("Qwen/Qwen2.5-Math-RM-72B",
max_transformers_version="4.53",
transformers_version_reason="HF model uses remote code that is not compatible with latest Transformers"), # noqa: E501
"Qwen2ForProcessRewardModel": _HfExamplesInfo("Qwen/Qwen2.5-Math-PRM-7B",
max_transformers_version="4.53",
transformers_version_reason="HF model uses remote code that is not compatible with latest Transformers"), # noqa: E501
"RobertaModel": _HfExamplesInfo("sentence-transformers/stsb-roberta-base-v2", v0_only=True), # noqa: E501
"RobertaForMaskedLM": _HfExamplesInfo("sentence-transformers/all-roberta-large-v1", v0_only=True), # noqa: E501
"XLMRobertaModel": _HfExamplesInfo("intfloat/multilingual-e5-small", v0_only=True), # noqa: E501