Merge tag 'v0.10.2rc1' into v0.10.2rc1-ori

examples/offline_inference/spec_decode.py

@@ -5,6 +5,7 @@ from transformers import AutoTokenizer
 
 from vllm import LLM, SamplingParams
 from vllm.benchmarks.datasets import add_dataset_parser, get_samples
+from vllm.inputs import TokensPrompt
 from vllm.v1.metrics.reader import Counter, Vector
 
 try:
@@ -137,7 +138,8 @@ def main():
     sampling_params = SamplingParams(temperature=args.temp, max_tokens=args.output_len)
     if not args.custom_mm_prompts:
         outputs = llm.generate(
-            prompt_token_ids=prompt_ids, sampling_params=sampling_params
+            [TokensPrompt(prompt_token_ids=x) for x in prompt_ids],
+            sampling_params=sampling_params,
         )
     else:
         outputs = llm.chat(prompts, sampling_params=sampling_params)

examples/offline_inference/structured_outputs.py

@@ -85,7 +85,7 @@ def format_output(title: str, output: str):
 
 
 def generate_output(prompt: str, sampling_params: SamplingParams, llm: LLM):
-    outputs = llm.generate(prompts=prompt, sampling_params=sampling_params)
+    outputs = llm.generate(prompt, sampling_params=sampling_params)
     return outputs[0].outputs[0].text
 
 

examples/offline_inference/vision_language.py

@@ -173,6 +173,37 @@ def run_deepseek_vl2(questions: list[str], modality: str) -> ModelRequestData:
     )
 
 
+# Ernie4.5-VL
+def run_ernie45_vl(questions: list[str], modality: str) -> ModelRequestData:
+    model_name = "baidu/ERNIE-4.5-VL-28B-A3B-PT"
+
+    engine_args = EngineArgs(
+        model=model_name,
+        max_model_len=4096,
+        max_num_seqs=5,
+        limit_mm_per_prompt={modality: 1},
+        trust_remote_code=True,
+    )
+
+    if modality == "image":
+        placeholder = "Picture 1:<|IMAGE_START|><|image@placeholder|><|IMAGE_END|>"
+    elif modality == "video":
+        placeholder = "Video 1:<|VIDEO_START|><|video@placeholder|><|VIDEO_END|>"
+
+    prompts = [
+        (
+            f"<|begin_of_sentence|>User: {question}{placeholder}\n"
+            "Assistant: <think></think>"
+        )
+        for question in questions
+    ]
+
+    return ModelRequestData(
+        engine_args=engine_args,
+        prompts=prompts,
+    )
+
+
 # Florence2
 def run_florence2(questions: list[str], modality: str) -> ModelRequestData:
     assert modality == "image"
@@ -283,8 +314,10 @@ def run_glm4v(questions: list[str], modality: str) -> ModelRequestData:
     )
 
     prompts = [
-        f"<|user|>\n<|begin_of_image|><|endoftext|><|end_of_image|>\
-        {question}<|assistant|>"
+        (
+            "<|user|>\n<|begin_of_image|><|endoftext|><|end_of_image|>"
+            f"{question}<|assistant|>"
+        )
         for question in questions
     ]
 
@@ -333,6 +366,80 @@ def run_glm4_1v(questions: list[str], modality: str) -> ModelRequestData:
     )
 
 
+# GLM-4.5V
+def run_glm4_5v(questions: list[str], modality: str) -> ModelRequestData:
+    model_name = "zai-org/GLM-4.5V"
+
+    engine_args = EngineArgs(
+        model=model_name,
+        max_model_len=4096,
+        max_num_seqs=2,
+        mm_processor_kwargs={
+            "size": {"shortest_edge": 12544, "longest_edge": 47040000},
+            "fps": 1,
+        },
+        limit_mm_per_prompt={modality: 1},
+        enforce_eager=True,
+        tensor_parallel_size=4,
+    )
+
+    if modality == "image":
+        placeholder = "<|begin_of_image|><|image|><|end_of_image|>"
+    elif modality == "video":
+        placeholder = "<|begin_of_video|><|video|><|end_of_video|>"
+
+    prompts = [
+        (
+            "[gMASK]<sop><|system|>\nYou are a helpful assistant.<|user|>\n"
+            f"{placeholder}"
+            f"{question}<|assistant|>assistant\n"
+        )
+        for question in questions
+    ]
+
+    return ModelRequestData(
+        engine_args=engine_args,
+        prompts=prompts,
+    )
+
+
+# GLM-4.5V-FP8
+def run_glm4_5v_fp8(questions: list[str], modality: str) -> ModelRequestData:
+    model_name = "zai-org/GLM-4.5V-FP8"
+
+    engine_args = EngineArgs(
+        model=model_name,
+        max_model_len=4096,
+        max_num_seqs=2,
+        mm_processor_kwargs={
+            "size": {"shortest_edge": 12544, "longest_edge": 47040000},
+            "fps": 1,
+        },
+        limit_mm_per_prompt={modality: 1},
+        enforce_eager=True,
+        tensor_parallel_size=4,
+    )
+
+    if modality == "image":
+        placeholder = "<|begin_of_image|><|image|><|end_of_image|>"
+    elif modality == "video":
+        placeholder = "<|begin_of_video|><|video|><|end_of_video|>"
+
+    prompts = [
+        (
+            "[gMASK]<sop><|system|>\nYou are a helpful assistant.<|user|>\n"
+            f"{placeholder}"
+            f"{question}<|assistant|>assistant\n"
+        )
+        for question in questions
+    ]
+
+    return ModelRequestData(
+        engine_args=engine_args,
+        prompts=prompts,
+    )
+
+
 # H2OVL-Mississippi
 def run_h2ovl(questions: list[str], modality: str) -> ModelRequestData:
     assert modality == "image"
@@ -693,15 +800,13 @@ def run_llava_next_video(questions: list[str], modality: str) -> ModelRequestDat
 def run_llava_onevision(questions: list[str], modality: str) -> ModelRequestData:
     if modality == "video":
         prompts = [
-            f"<|im_start|>user <video>\n{question}<|im_end|> \
-        <|im_start|>assistant\n"
+            f"<|im_start|>user <video>\n{question}<|im_end|><|im_start|>assistant\n"
             for question in questions
         ]
 
     elif modality == "image":
         prompts = [
-            f"<|im_start|>user <image>\n{question}<|im_end|> \
-        <|im_start|>assistant\n"
+            f"<|im_start|>user <image>\n{question}<|im_end|><|im_start|>assistant\n"
             for question in questions
         ]
 
@@ -815,6 +920,39 @@ def run_minicpmv(questions: list[str], modality: str) -> ModelRequestData:
     return run_minicpmv_base(questions, modality, "openbmb/MiniCPM-V-2_6")
 
 
+def run_minimax_vl_01(questions: list[str], modality: str) -> ModelRequestData:
+    assert modality == "image"
+
+    model_name = "MiniMaxAI/MiniMax-VL-01"
+
+    engine_args = EngineArgs(
+        model=model_name,
+        max_num_seqs=2,
+        limit_mm_per_prompt={modality: 1},
+        trust_remote_code=True,
+        tensor_parallel_size=8,
+    )
+
+    tokenizer = AutoTokenizer.from_pretrained(model_name)
+    messages = [
+        [
+            {
+                "role": "user",
+                "content": [{"type": "image"}, {"type": "text", "text": question}],
+            }
+        ]
+        for question in questions
+    ]
+    prompts = tokenizer.apply_chat_template(
+        messages, add_generation_prompt=True, tokenize=False
+    )
+
+    return ModelRequestData(
+        engine_args=engine_args,
+        prompts=prompts,
+    )
+
+
 # Mistral-3 HF-format
 def run_mistral3(questions: list[str], modality: str) -> ModelRequestData:
     assert modality == "image"
@@ -891,8 +1029,7 @@ def run_molmo(questions: list[str], modality: str) -> ModelRequestData:
     )
 
     prompts = [
-        f"<|im_start|>user <image>\n{question}<|im_end|> \
-        <|im_start|>assistant\n"
+        f"<|im_start|>user <image>\n{question}<|im_end|><|im_start|>assistant\n"
         for question in questions
     ]
 
@@ -998,6 +1135,38 @@ def run_ovis(questions: list[str], modality: str) -> ModelRequestData:
     )
 
 
+# Ovis2_5
+def run_ovis2_5(questions: list[str], modality: str) -> ModelRequestData:
+    model_name = "AIDC-AI/Ovis2.5-2B"
+
+    engine_args = EngineArgs(
+        model=model_name,
+        max_model_len=4096,
+        max_num_seqs=2,
+        trust_remote_code=True,
+        dtype="half",
+        limit_mm_per_prompt={modality: 1},
+    )
+    if modality == "image":
+        placeholder = "<image>"
+    elif modality == "video":
+        placeholder = "<video>"
+
+    tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
+    messages = [
+        [{"role": "user", "content": f"{placeholder}\n{question}"}]
+        for question in questions
+    ]
+    prompts = tokenizer.apply_chat_template(
+        messages, tokenize=False, add_generation_prompt=True
+    )
+
+    return ModelRequestData(
+        engine_args=engine_args,
+        prompts=prompts,
+    )
+
+
 # PaliGemma
 def run_paligemma(questions: list[str], modality: str) -> ModelRequestData:
     assert modality == "image"
@@ -1297,6 +1466,28 @@ def run_qwen2_5_omni(questions: list[str], modality: str):
     )
 
 
+# R-4B
+def run_r_vl(questions: list[str], modality: str) -> ModelRequestData:
+    assert modality == "image"
+    model_name = "YannQi/R-4B"
+
+    prompts = [
+        f"<|im_start|>user <image>\n{question}<|im_end|><|im_start|>assistant\n"
+        for question in questions
+    ]
+
+    engine_args = EngineArgs(
+        model=model_name,
+        max_model_len=16384,
+        limit_mm_per_prompt={modality: 1},
+    )
+
+    return ModelRequestData(
+        engine_args=engine_args,
+        prompts=prompts,
+    )
+
+
 # SkyworkR1V
 def run_skyworkr1v(questions: list[str], modality: str) -> ModelRequestData:
     assert modality == "image"
@@ -1442,12 +1633,15 @@ model_example_map = {
     "chameleon": run_chameleon,
     "command_a_vision": run_command_a_vision,
     "deepseek_vl_v2": run_deepseek_vl2,
+    "ernie45_vl": run_ernie45_vl,
     "florence2": run_florence2,
     "fuyu": run_fuyu,
     "gemma3": run_gemma3,
     "gemma3n": run_gemma3n,
     "glm4v": run_glm4v,
     "glm4_1v": run_glm4_1v,
+    "glm4_5v": run_glm4_5v,
+    "glm4_5v_fp8": run_glm4_5v_fp8,
     "h2ovl_chat": run_h2ovl,
     "hyperclovax_seed_vision": run_hyperclovax_seed_vision,
     "idefics3": run_idefics3,
@@ -1463,12 +1657,14 @@ model_example_map = {
     "mantis": run_mantis,
     "minicpmo": run_minicpmo,
     "minicpmv": run_minicpmv,
+    "minimax_vl_01": run_minimax_vl_01,
     "mistral3": run_mistral3,
     "mllama": run_mllama,
     "molmo": run_molmo,
     "nemotron_vl": run_nemotron_vl,
     "NVLM_D": run_nvlm_d,
     "ovis": run_ovis,
+    "ovis2_5": run_ovis2_5,
     "paligemma": run_paligemma,
     "paligemma2": run_paligemma2,
     "phi3_v": run_phi3v,
@@ -1479,6 +1675,7 @@ model_example_map = {
     "qwen2_vl": run_qwen2_vl,
     "qwen2_5_vl": run_qwen2_5_vl,
     "qwen2_5_omni": run_qwen2_5_omni,
+    "rvl": run_r_vl,
     "skywork_chat": run_skyworkr1v,
     "smolvlm": run_smolvlm,
     "step3": run_step3,

examples/offline_inference/vision_language_multi_image.py

@@ -680,6 +680,36 @@ def load_ovis(question: str, image_urls: list[str]) -> ModelRequestData:
     )
 
 
+# ovis2_5
+def load_ovis2_5(question: str, image_urls: list[str]) -> ModelRequestData:
+    model_name = "AIDC-AI/Ovis2.5-2B"
+
+    engine_args = EngineArgs(
+        model=model_name,
+        max_model_len=8192,
+        max_num_seqs=2,
+        trust_remote_code=True,
+        dtype="half",
+        limit_mm_per_prompt={"image": len(image_urls)},
+    )
+
+    placeholders = "\n".join(
+        f"Image-{i}: <image>\n" for i, _ in enumerate(image_urls, start=1)
+    )
+    messages = [{"role": "user", "content": f"{placeholders}\n{question}"}]
+
+    tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
+    prompt = tokenizer.apply_chat_template(
+        messages, tokenize=False, add_generation_prompt=True
+    )
+
+    return ModelRequestData(
+        engine_args=engine_args,
+        prompt=prompt,
+        image_data=[fetch_image(url) for url in image_urls],
+    )
+
+
 def load_pixtral_hf(question: str, image_urls: list[str]) -> ModelRequestData:
     model_name = "mistral-community/pixtral-12b"
 
@@ -962,6 +992,39 @@ def load_qwen2_5_vl(question: str, image_urls: list[str]) -> ModelRequestData:
     )
 
 
+def load_r_vl(question: str, image_urls: list[str]) -> ModelRequestData:
+    model_name = "YannQi/R-4B"
+    engine_args = EngineArgs(
+        model=model_name,
+        max_model_len=16384,
+        max_num_seqs=16,
+        limit_mm_per_prompt={"image": len(image_urls)},
+    )
+
+    placeholders = [{"type": "image", "image": url} for url in image_urls]
+    messages = [
+        {
+            "role": "user",
+            "content": [
+                *placeholders,
+                {"type": "text", "text": question},
+            ],
+        }
+    ]
+
+    processor = AutoProcessor.from_pretrained(model_name, trust_remote_code=True)
+
+    prompt = processor.apply_chat_template(
+        messages, tokenize=False, add_generation_prompt=True
+    )
+
+    return ModelRequestData(
+        engine_args=engine_args,
+        prompt=prompt,
+        image_data=[fetch_image(url) for url in image_urls],
+    )
+
+
 def load_smolvlm(question: str, image_urls: list[str]) -> ModelRequestData:
     model_name = "HuggingFaceTB/SmolVLM2-2.2B-Instruct"
 
@@ -1064,6 +1127,76 @@ def load_tarsier2(question: str, image_urls: list[str]) -> ModelRequestData:
     )
 
 
+# GLM-4.5V
+def load_glm4_5v(question: str, image_urls: list[str]) -> ModelRequestData:
+    model_name = "zai-org/GLM-4.5V"
+
+    engine_args = EngineArgs(
+        model=model_name,
+        max_model_len=32768,
+        max_num_seqs=2,
+        limit_mm_per_prompt={"image": len(image_urls)},
+        enforce_eager=True,
+        tensor_parallel_size=4,
+    )
+    placeholders = [{"type": "image", "image": url} for url in image_urls]
+    messages = [
+        {
+            "role": "user",
+            "content": [
+                *placeholders,
+                {"type": "text", "text": question},
+            ],
+        }
+    ]
+    processor = AutoProcessor.from_pretrained(model_name)
+    prompt = processor.apply_chat_template(
+        messages, tokenize=False, add_generation_prompt=True
+    )
+    image_data = [fetch_image(url) for url in image_urls]
+
+    return ModelRequestData(
+        engine_args=engine_args,
+        prompt=prompt,
+        image_data=image_data,
+    )
+
+
+# GLM-4.5V-FP8
+def load_glm4_5v_fp8(question: str, image_urls: list[str]) -> ModelRequestData:
+    model_name = "zai-org/GLM-4.5V-FP8"
+
+    engine_args = EngineArgs(
+        model=model_name,
+        max_model_len=32768,
+        max_num_seqs=2,
+        limit_mm_per_prompt={"image": len(image_urls)},
+        enforce_eager=True,
+        tensor_parallel_size=4,
+    )
+    placeholders = [{"type": "image", "image": url} for url in image_urls]
+    messages = [
+        {
+            "role": "user",
+            "content": [
+                *placeholders,
+                {"type": "text", "text": question},
+            ],
+        }
+    ]
+    processor = AutoProcessor.from_pretrained(model_name)
+    prompt = processor.apply_chat_template(
+        messages, tokenize=False, add_generation_prompt=True
+    )
+    image_data = [fetch_image(url) for url in image_urls]
+
+    return ModelRequestData(
+        engine_args=engine_args,
+        prompt=prompt,
+        image_data=image_data,
+    )
+
+
 model_example_map = {
     "aria": load_aria,
     "aya_vision": load_aya_vision,
@@ -1085,6 +1218,7 @@ model_example_map = {
     "mllama": load_mllama,
     "NVLM_D": load_nvlm_d,
     "ovis": load_ovis,
+    "ovis2_5": load_ovis2_5,
     "phi3_v": load_phi3v,
     "phi4_mm": load_phi4mm,
     "phi4_multimodal": load_phi4_multimodal,
@@ -1092,10 +1226,13 @@ model_example_map = {
     "qwen_vl_chat": load_qwen_vl_chat,
     "qwen2_vl": load_qwen2_vl,
     "qwen2_5_vl": load_qwen2_5_vl,
+    "rvl": load_r_vl,
     "smolvlm": load_smolvlm,
     "step3": load_step3,
     "tarsier": load_tarsier,
     "tarsier2": load_tarsier2,
+    "glm4_5v": load_glm4_5v,
+    "glm4_5v_fp8": load_glm4_5v_fp8,
 }
 
 

examples/online_serving/kv_events_subscriber.py

@@ -27,10 +27,12 @@ class BlockStored(KVCacheEvent):
     token_ids: list[int]
     block_size: int
     lora_id: Optional[int]
+    medium: Optional[str]
 
 
 class BlockRemoved(KVCacheEvent):
     block_hashes: list[int]
+    medium: Optional[str]
 
 
 class AllBlocksCleared(KVCacheEvent):

examples/online_serving/prithvi_geospatial_mae.py

+# SPDX-License-Identifier: Apache-2.0
+# SPDX-FileCopyrightText: Copyright contributors to the vLLM project
+
+import base64
+import os
+
+import requests
+
+# This example shows how to perform an online inference that generates
+# multimodal data. In this specific case this example will take a geotiff
+# image as input, process it using the multimodal data processor, and
+# perform inference.
+# Reuirements :
+# - install plugin at:
+#   https://github.com/christian-pinto/prithvi_io_processor_plugin
+# - start vllm in serving mode with the below args
+#   --model='christian-pinto/Prithvi-EO-2.0-300M-TL-VLLM'
+#   --task embed --trust-remote-code
+#   --skip-tokenizer-init --enforce-eager
+#   --io-processor-plugin prithvi_to_tiff_india
+
+
+def main():
+    image_url = "https://huggingface.co/christian-pinto/Prithvi-EO-2.0-300M-TL-VLLM/resolve/main/India_900498_S2Hand.tif"  # noqa: E501
+    server_endpoint = "http://localhost:8000/pooling"
+
+    request_payload_url = {
+        "data": {
+            "data": image_url,
+            "data_format": "url",
+            "image_format": "tiff",
+            "out_data_format": "b64_json",
+        },
+        "priority": 0,
+        "model": "christian-pinto/Prithvi-EO-2.0-300M-TL-VLLM",
+    }
+
+    ret = requests.post(server_endpoint, json=request_payload_url)
+
+    print(f"response.status_code: {ret.status_code}")
+    print(f"response.reason:{ret.reason}")
+
+    response = ret.json()
+
+    decoded_image = base64.b64decode(response["data"]["data"])
+
+    out_path = os.path.join(os.getcwd(), "online_prediction.tiff")
+
+    with open(out_path, "wb") as f:
+        f.write(decoded_image)
+
+
+if __name__ == "__main__":
+    main()

examples/tool_chat_template_deepseekv31.jinja

+{% if not add_generation_prompt is defined %}
+  {% set add_generation_prompt = false %}
+{% endif %}
+{% if not thinking is defined %}
+  {% set thinking = false %}
+{% endif %}
+{% set ns = namespace(is_first=false, is_tool=false, system_prompt='', is_first_sp=true, is_last_user=false) %}
+{%- for message in messages %}
+  {%- if message['role'] == 'system' %}
+    {%- if ns.is_first_sp %}
+      {% set ns.system_prompt = ns.system_prompt + message['content'] %}
+      {% set ns.is_first_sp = false %}
+    {%- else %}
+      {% set ns.system_prompt = ns.system_prompt + '\n\n' + message['content'] %}
+    {%- endif %}
+  {%- endif %}
+{%- endfor %}
+
+{% if tools is defined and tools is not none %}
+  {% set tool_ns = namespace(text='## Tools\nYou have access to the following tools:\n') %}
+  {% for tool in tools %}
+    {% set tool_ns.text = tool_ns.text + '\n### ' + tool.function.name + '\nDescription: ' + tool.function.description + '\n\nParameters: ' + (tool.function.parameters | tojson) + '\n' %}
+  {% endfor %}
+  {% set tool_ns.text = tool_ns.text + "\nIMPORTANT: ALWAYS adhere to this exact format for tool use:\n<｜tool▁calls▁begin｜><｜tool▁call▁begin｜>tool_call_name<｜tool▁sep｜>tool_call_arguments<｜tool▁call▁end｜>{{additional_tool_calls}}<｜tool▁calls▁end｜>\n\nWhere:\n\n- `tool_call_name` must be an exact match to one of the available tools\n- `tool_call_arguments` must be valid JSON that strictly follows the tool's Parameters Schema\n- For multiple tool calls, chain them directly without separators or spaces\n" %}
+  {% set ns.system_prompt = ns.system_prompt + '\n\n' + tool_ns.text %}
+{% endif %}
+
+{{ bos_token }}{{ ns.system_prompt }}
+{%- for message in messages %}
+  {%- if message['role'] == 'user' %}
+    {%- set ns.is_tool = false -%}
+    {%- set ns.is_first = false -%}
+    {%- set ns.is_last_user = true -%}
+    {{'<｜User｜>' + message['content']}}
+  {%- endif %}
+  {%- if message['role'] == 'assistant' and message['tool_calls'] is defined and message['tool_calls'] is not none %}
+    {%- if ns.is_last_user %}
+      {{'<｜Assistant｜></think>'}}
+    {%- endif %}
+    {%- set ns.is_last_user = false -%}
+    {%- set ns.is_first = false %}
+    {%- set ns.is_tool = false -%}
+    {%- for tool in message['tool_calls'] %}
+      {%- if not ns.is_first %}
+        {%- if message['content'] is none %}
+          {{'<｜tool▁calls▁begin｜><｜tool▁call▁begin｜>'+ tool['function']['name'] + '<｜tool▁sep｜>' + tool['function']['arguments']|tojson + '<｜tool▁call▁end｜>'}}
+        {%- else %}
+          {{message['content'] + '<｜tool▁calls▁begin｜><｜tool▁call▁begin｜>' + tool['function']['name'] + '<｜tool▁sep｜>' + tool['function']['arguments']|tojson + '<｜tool▁call▁end｜>'}}
+        {%- endif %}
+        {%- set ns.is_first = true -%}
+      {%- else %}
+        {{'<｜tool▁call▁begin｜>'+ tool['function']['name'] + '<｜tool▁sep｜>' + tool['function']['arguments']|tojson + '<｜tool▁call▁end｜>'}}
+      {%- endif %}
+    {%- endfor %}
+    {{'<｜tool▁calls▁end｜><｜end▁of▁sentence｜>'}}
+  {%- endif %}
+  {%- if message['role'] == 'assistant' and (message['tool_calls'] is not defined or message['tool_calls'] is none) %}
+    {%- if ns.is_last_user %}
+      {{'<｜Assistant｜>'}}
+      {%- if message['prefix'] is defined and message['prefix'] and thinking %}
+        {{'<think>'}}  
+      {%- else %}
+        {{'</think>'}}
+      {%- endif %}
+    {%- endif %}
+    {%- set ns.is_last_user = false -%}
+    {%- if ns.is_tool %}
+      {{message['content'] + '<｜end▁of▁sentence｜>'}}
+      {%- set ns.is_tool = false -%}
+    {%- else %}
+      {%- set content = message['content'] -%}
+      {%- if '</think>' in content %}
+        {%- set content = content.split('</think>', 1)[1] -%}
+      {%- endif %}
+      {{content + '<｜end▁of▁sentence｜>'}}
+    {%- endif %}
+  {%- endif %}
+  {%- if message['role'] == 'tool' %}
+    {%- set ns.is_last_user = false -%}
+    {%- set ns.is_tool = true -%}
+    {{'<｜tool▁output▁begin｜>' + message['content'] + '<｜tool▁output▁end｜>'}}
+  {%- endif %}
+{%- endfor -%}
+{%- if add_generation_prompt and ns.is_last_user and not ns.is_tool %}
+  {{'<｜Assistant｜>'}}
+  {%- if not thinking %}
+    {{'</think>'}}
+  {%- else %}
+    {{'<think>'}}
+  {%- endif %}
+{% endif %}

examples/tool_chat_template_gemma3_pythonic.jinja

+{#- Begin-of-sequence token to start the model prompt -#}
+{{ bos_token }}
+{#- Extracts the system message. Gemma does not support system messages so it will be prepended to first user message. -#}
+{%- if messages[0]['role'] == 'system' -%}
+    {%- if messages[0]['content'] is string -%}
+        {%- set first_user_prefix = messages[0]['content'] + '\n\n' -%}
+    {%- else -%}
+        {%- set first_user_prefix = messages[0]['content'][0]['text'] + '\n\n' -%}
+    {%- endif -%}
+    {%- set loop_messages = messages[1:] -%}
+{%- else -%}
+    {%- set first_user_prefix = "" -%}
+    {%- set loop_messages = messages -%}
+{%- endif -%}
+{#- Set tools to none if not defined for this ChatCompletion request (helps avoid errors later) -#}
+{%- if not tools is defined %}
+    {%- set tools = none %}
+{%- endif %}
+{#- Validate alternating user/assistant messages (excluding 'tool' messages and ones with tool_calls) -#}
+{%- for message in loop_messages | rejectattr("role", "equalto", "tool") | selectattr("tool_calls", "undefined") -%}
+    {%- if (message['role'] == 'user') != (loop.index0 % 2 == 0) %}
+        {{ raise_exception("Conversation roles must alternate user/assistant/user/assistant/...") }}
+    {%- endif -%}
+{%- endfor -%}
+
+{#- Main loop over all messages in the conversation history -#}
+{%- for message in loop_messages -%}
+    {#- Normalize roles for model prompt formatting -#}
+    {%- if (message['role'] == 'assistant') -%}
+        {%- set role = "model" -%}
+    {%- elif (message['role'] == 'tool') -%}
+        {%- set role = "user" -%}
+    {%- else -%}
+        {%- set role = message['role'] -%}
+    {%- endif -%}
+    {#- Mark the start of a message block with the appropriate role -#}
+    {{ '<start_of_turn>' + role + '\n' -}}
+
+    {#- Insert system message content (if present) at the beginning of the first message. -#}
+    {%- if loop.first -%}
+        {{ first_user_prefix }}
+        {#- Append system message with tool information if using tools in message request. -#}
+        {%- if tools is not none -%}
+            {{- "Tools (functions) are available. If you decide to invoke one or more of the tools, you must respond with a python list of the function calls.\n" -}}
+            {{- "Example Format: [func_name1(params_name1=params_value1, params_name2=params_value2...), func_name2(params)] \n" -}}
+            {{- "Do not use variables. DO NOT USE MARKDOWN SYNTAX. You SHOULD NOT include any other text in the response if you call a function. If none of the functions can be used, point it out. If you lack the parameters required by the function, also point it out.\n" -}}
+            {{- "Here is a list of functions in JSON format that you can invoke.\n" -}}
+            {{- tools | tojson(indent=4) -}}
+            {{- "\n\n" -}}
+        {%- endif -%}
+    {%- endif -%}
+
+    {#- Format model tool calls (turns where model indicates they want to call a tool) -#}
+    {%- if 'tool_calls' in message -%}
+        {#- Opening bracket for tool call list. -#}
+        {{- '[' -}}
+        {#- For each tool call -#}
+        {%- for tool_call in message.tool_calls -%}
+            {#- Get tool call function. -#}
+            {%- if tool_call.function is defined -%}
+                {%- set tool_call = tool_call.function -%}
+            {%- endif -%}
+            {#- Function name & opening parenthesis. -#}
+            {{- tool_call.name + '(' -}}
+
+            {#-- Handle arguments as list (positional) or dict (named) --#}
+            {#-- Named arguments (dict) --#}
+            {%- if tool_call.arguments is iterable and tool_call.arguments is mapping -%}
+                {%- set first = true -%}
+                {%- for key, val in tool_call.arguments.items() -%}
+                    {%- if not first %}, {% endif -%}
+                    {{ key }}={{ val | tojson }}
+                    {%- set first = false -%}
+                {%- endfor -%}
+            {#-- Positional arguments (list) --#}
+            {%- elif tool_call.arguments is iterable -%}
+                {{- tool_call.arguments | map('tojson') | join(', ') -}}
+            {#-- Fallback: single positional value --#}
+            {%- else -%}
+                {{- tool_call.arguments | tojson -}}
+            {#-- Closing parenthesis. --#}
+            {%- endif -%}
+                {{- ')' -}}
+            {#-- If more than one tool call, place comma and move to formatting next tool call --#}
+            {%- if not loop.last -%}, {% endif -%}
+        {%- endfor -%}
+        {#- Closing bracket for tool call list. -#}
+        {{- ']' -}}
+    {%- endif -%}
+    
+    {#- Tool response start tag (for messages from a tool) -#}
+    {%- if (message['role'] == 'tool') -%}
+        {{ '<tool_response>\n' -}}
+    {%- endif -%}
+
+    {#- Render the message content: handle plain string or multimodal content like image/text -#}
+    {%- if message['content'] is string -%}
+        {{ message['content'] | trim }}
+    {%- elif message['content'] is iterable -%}
+        {%- for item in message['content'] -%}
+            {%- if item['type'] == 'image' -%}
+                {{ '<start_of_image>' }}
+            {%- elif item['type'] == 'text' -%}
+                {{ item['text'] | trim }}
+            {%- endif -%}
+        {%- endfor -%}
+    {%- else -%}
+        {{ raise_exception("Invalid content type") }}
+    {%- endif -%}
+
+    {#- Tool response end tag -#}
+    {%- if (message['role'] == 'tool') -%}
+        {{ '</tool_response>' -}}
+    {%- endif -%}
+
+    {#- Mark end of a single turn -#}
+    {{ '<end_of_turn>\n' }}
+{%- endfor -%}
+
+{#- If generation is to be triggered, add model prompt prefix -#}
+{%- if add_generation_prompt -%}
+    {{'<start_of_turn>model\n'}}
+{%- endif -%}
\ No newline at end of file

examples/tool_chat_template_phi4_mini.jinja

+{%- if messages and messages[0]['role'] == 'system' %}
+    {%- set system_message = messages[0]['content']|trim %}
+    {%- set messages = messages[1:] %}
+{%- else %}
+    {%- set system_message = "You are a helpful assistant." %}
+{%- endif %}
+
 {%- if messages %}
-    {%- if system_message or tools %}
 <|system|>
-
-{%- if system_message %}
 {{ system_message }}
-{%- endif %}
+{%- if tools %}
 In addition to plain text responses, you can chose to call one or more of the provided functions.
 
 Use the following rule to decide when to call a function:
@@ -19,13 +23,11 @@ If you decide to call functions:
   * make sure you pick the right functions that match the user intent
 
 
-{%- if tools %}
         {%- for t in tools %}
             {{- t | tojson(indent=4) }}
             {{- "\n\n" }}
         {%- endfor %}
 {%- endif %}<|end|>
-    {%- endif %}
 
     {%- for message in messages %}
         {%- if message.role != "system" %}

examples/tool_chat_template_qwen3coder.jinja

+{% macro render_extra_keys(json_dict, handled_keys) %}
+    {%- if json_dict is mapping %}
+        {%- for json_key in json_dict if json_key not in handled_keys %}
+            {%- if json_dict[json_key] is mapping or (json_dict[json_key] is sequence and json_dict[json_key] is not string) %}
+                {{- '\n<' ~ json_key ~ '>' ~ (json_dict[json_key] | tojson | safe) ~ '</' ~ json_key ~ '>' }}
+            {%- else %}
+                {{-'\n<' ~ json_key ~ '>' ~ (json_dict[json_key] | string) ~ '</' ~ json_key ~ '>' }}
+            {%- endif %}
+        {%- endfor %}
+    {%- endif %}
+{% endmacro %}
+
+{%- if messages[0]["role"] == "system" %}
+    {%- set system_message = messages[0]["content"] %}
+    {%- set loop_messages = messages[1:] %}
+{%- else %}
+    {%- set loop_messages = messages %}
+{%- endif %}
+
+{%- if not tools is defined %}
+    {%- set tools = [] %}
+{%- endif %}
+
+{%- if system_message is defined %}
+    {{- "<|im_start|>system\n" + system_message }}
+{%- else %}
+    {%- if tools is iterable and tools | length > 0 %}
+        {{- "<|im_start|>system\nYou are Qwen, a helpful AI assistant that can interact with a computer to solve tasks." }}
+    {%- endif %}
+{%- endif %}
+{%- if tools is iterable and tools | length > 0 %}
+    {{- "\n\n# Tools\n\nYou have access to the following functions:\n\n" }}
+    {{- "<tools>" }}
+    {%- for tool in tools %}
+        {%- if tool.function is defined %}
+            {%- set tool = tool.function %}
+        {%- endif %}
+        {{- "\n<function>\n<name>" ~ tool.name ~ "</name>" }}
+        {%- if tool.description is defined %}
+            {{- '\n<description>' ~ (tool.description | trim) ~ '</description>' }}
+        {%- endif %}
+        {{- '\n<parameters>' }}
+        {%- if tool.parameters is defined and tool.parameters is mapping and tool.parameters.properties is defined and tool.parameters.properties is mapping %}
+            {%- for param_name, param_fields in tool.parameters.properties|items %}
+                {{- '\n<parameter>' }}
+                {{- '\n<name>' ~ param_name ~ '</name>' }}
+                {%- if param_fields.type is defined %}
+                    {{- '\n<type>' ~ (param_fields.type | string) ~ '</type>' }}
+                {%- endif %}
+                {%- if param_fields.description is defined %}
+                    {{- '\n<description>' ~ (param_fields.description | trim) ~ '</description>' }}
+                {%- endif %}
+                {%- set handled_keys = ['name', 'type', 'description'] %}
+                {{- render_extra_keys(param_fields, handled_keys) }}
+                {{- '\n</parameter>' }}
+            {%- endfor %}
+        {%- endif %}
+        {% set handled_keys = ['type', 'properties'] %}
+        {{- render_extra_keys(tool.parameters, handled_keys) }}
+        {{- '\n</parameters>' }}
+        {%- set handled_keys = ['type', 'name', 'description', 'parameters'] %}
+        {{- render_extra_keys(tool, handled_keys) }}
+        {{- '\n</function>' }}
+    {%- endfor %}
+    {{- "\n</tools>" }}
+    {{- '\n\nIf you choose to call a function ONLY reply in the following format with NO suffix:\n\n<tool_call>\n<function=example_function_name>\n<parameter=example_parameter_1>\nvalue_1\n</parameter>\n<parameter=example_parameter_2>\nThis is the value for the second parameter\nthat can span\nmultiple lines\n</parameter>\n</function>\n</tool_call>\n\n<IMPORTANT>\nReminder:\n- Function calls MUST follow the specified format: an inner <function=...></function> block must be nested within <tool_call></tool_call> XML tags\n- Required parameters MUST be specified\n- You may provide optional reasoning for your function call in natural language BEFORE the function call, but NOT after\n- If there is no function call available, answer the question like normal with your current knowledge and do not tell the user about function calls\n</IMPORTANT>' }}
+{%- endif %}
+{%- if system_message is defined %}
+    {{- '<|im_end|>\n' }}
+{%- else %}
+    {%- if tools is iterable and tools | length > 0 %}
+        {{- '<|im_end|>\n' }}
+    {%- endif %}
+{%- endif %}
+{%- for message in loop_messages %}
+    {%- if message.role == "assistant" and message.tool_calls is defined and message.tool_calls is iterable and message.tool_calls | length > 0 %}
+        {{- '<|im_start|>' + message.role }}
+        {%- if message.content is defined and message.content is string and message.content | trim | length > 0 %}
+            {{- '\n' + message.content | trim + '\n' }}
+        {%- endif %}
+        {%- for tool_call in message.tool_calls %}
+            {%- if tool_call.function is defined %}
+                {%- set tool_call = tool_call.function %}
+            {%- endif %}
+            {{- '\n<tool_call>\n<function=' + tool_call.name + '>\n' }}
+            {%- if tool_call.arguments is defined %}
+                {%- for args_name, args_value in tool_call.arguments|items %}
+                    {{- '<parameter=' + args_name + '>\n' }}
+                    {%- set args_value = args_value | tojson | safe if args_value is mapping or (args_value is sequence and args_value is not string) else args_value | string %}
+                    {{- args_value }}
+                    {{- '\n</parameter>\n' }}
+                {%- endfor %}
+            {%- endif %}
+            {{- '</function>\n</tool_call>' }}
+        {%- endfor %}
+        {{- '<|im_end|>\n' }}
+    {%- elif message.role == "user" or message.role == "system" or message.role == "assistant" %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
+    {%- elif message.role == "tool" %}
+        {%- if loop.previtem and loop.previtem.role != "tool" %}
+            {{- '<|im_start|>user\n' }}
+        {%- endif %}
+        {{- '<tool_response>\n' }}
+        {{- message.content }}
+        {{- '\n</tool_response>\n' }}
+        {%- if not loop.last and loop.nextitem.role != "tool" %}
+            {{- '<|im_end|>\n' }}
+        {%- elif loop.last %}
+            {{- '<|im_end|>\n' }}
+        {%- endif %}
+    {%- else %}
+        {{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>\n' }}
+    {%- endif %}
+{%- endfor %}
+{%- if add_generation_prompt %}
+    {{- '<|im_start|>assistant\n' }}
+{%- endif %}

mkdocs.yaml

@@ -129,15 +129,16 @@ markdown_extensions:
   - toc:
       permalink: true
   # For math rendering
-  - mdx_math:
-      enable_dollar_delimiter: true
+  - pymdownx.arithmatex:
+      generic: true
 
 extra_css:
   - mkdocs/stylesheets/extra.css
 
 extra_javascript:
   - mkdocs/javascript/run_llm_widget.js
-  - https://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS_HTML
+  - mkdocs/javascript/mathjax.js
+  - https://unpkg.com/mathjax@3.2.2/es5/tex-mml-chtml.js
   - mkdocs/javascript/edit_and_feedback.js
   - mkdocs/javascript/slack_and_forum.js
 

pyproject.toml

@@ -6,7 +6,7 @@ requires = [
     "packaging>=24.2",
     "setuptools>=77.0.3,<80.0.0",
     "setuptools-scm>=8.0",
-    "torch == 2.7.1",
+    "torch == 2.8.0",
     "wheel",
     "jinja2",
 ]
@@ -24,13 +24,14 @@ classifiers = [
     "Programming Language :: Python :: 3.10",
     "Programming Language :: Python :: 3.11",
     "Programming Language :: Python :: 3.12",
+    "Programming Language :: Python :: 3.13",
     "Intended Audience :: Developers",
     "Intended Audience :: Information Technology",
     "Intended Audience :: Science/Research",
     "Topic :: Scientific/Engineering :: Artificial Intelligence",
     "Topic :: Scientific/Engineering :: Information Analysis",
 ]
-requires-python = ">=3.9,<3.13"
+requires-python = ">=3.9,<3.14"
 dynamic = [ "version", "dependencies", "optional-dependencies"]
 
 [project.urls]

requirements/build.txt

@@ -4,7 +4,8 @@ ninja
 packaging>=24.2
 setuptools>=77.0.3,<80.0.0
 setuptools-scm>=8
-torch==2.7.1
+torch==2.8.0
 wheel
 jinja2>=3.1.6
 regex
+build

requirements/common.txt

@@ -7,20 +7,21 @@ requests >= 2.26.0
 tqdm
 blake3
 py-cpuinfo
-transformers >= 4.55.0
+transformers >= 4.55.2
 tokenizers >= 0.21.1  # Required for fast incremental detokenization.
 protobuf # Required by LlamaTokenizer.
 fastapi[standard] >= 0.115.0 # Required by FastAPI's form models in the OpenAI API server's audio transcriptions endpoint.
 aiohttp
 openai >= 1.99.1  # For Responses API with reasoning content
-pydantic >= 2.10
+pydantic >= 2.11.7
 prometheus_client >= 0.18.0
 pillow  # Required for image processing
 prometheus-fastapi-instrumentator >= 7.0.0
 tiktoken >= 0.6.0  # Required for DBRX tokenizer
-lm-format-enforcer >= 0.10.11, < 0.11
+lm-format-enforcer == 0.11.3
 llguidance >= 0.7.11, < 0.8.0; platform_machine == "x86_64" or platform_machine == "arm64" or platform_machine == "aarch64"
-outlines_core == 0.2.10
+outlines_core == 0.2.10 ; platform_machine != "s390x"
+outlines == 0.1.11 ; platform_machine == "s390x"
 # required for outlines backend disk cache
 diskcache == 5.6.3
 lark == 1.2.2
@@ -38,7 +39,7 @@ pyyaml
 six>=1.16.0; python_version > '3.11' # transitive dependency of pandas that needs to be the latest version for python 3.12
 setuptools>=77.0.3,<80; python_version > '3.11' # Setuptools is used by triton, we need to ensure a modern version is installed for 3.12+ so that it does not try to import distutils, which was removed in 3.12
 einops # Required for Qwen2-VL.
-compressed-tensors == 0.10.2 # required for compressed-tensors
+compressed-tensors == 0.11.0 # required for compressed-tensors
 depyf==0.19.0 # required for profiling and debugging with compilation config
 cloudpickle # allows pickling lambda functions in model_executor/models/registry.py
 watchfiles # required for http server to monitor the updates of TLS files

requirements/cpu.txt

 # Common dependencies
 -r common.txt
 
-numba == 0.60.0; python_version == '3.9' # v0.61 doesn't support Python 3.9. Required for N-gram speculative decoding
-numba == 0.61.2; python_version > '3.9'
+numba == 0.60.0; python_version == '3.9' and platform_machine != "s390x" # v0.61 doesn't support Python 3.9. Required for N-gram speculative decoding
+numba == 0.61.2; python_version > '3.9' and platform_machine != "s390x"
 
 # Dependencies for CPUs
 packaging>=24.2
 setuptools>=77.0.3,<80.0.0
 --extra-index-url https://download.pytorch.org/whl/cpu
 torch==2.6.0+cpu; platform_machine == "x86_64" # torch>2.6.0+cpu has performance regression on x86 platform, see https://github.com/pytorch/pytorch/pull/151218
-torch==2.7.0; platform_system == "Darwin"
-torch==2.7.0; platform_machine == "ppc64le"
-torch==2.6.0; platform_machine == "aarch64" # for arm64 CPUs, torch 2.7.0 has a issue: https://github.com/vllm-project/vllm/issues/17960
+torch==2.8.0; platform_system == "Darwin"
+torch==2.8.0; platform_machine == "ppc64le" or platform_machine == "aarch64"
 
 # required for the image processor of minicpm-o-2_6, this must be updated alongside torch
 torchaudio; platform_machine != "ppc64le" and platform_machine != "s390x"
-torchaudio==2.7.0; platform_machine == "ppc64le"
+torchaudio==2.8.0; platform_machine == "ppc64le"
 
 # required for the image processor of phi3v, this must be updated alongside torch
 torchvision; platform_machine != "ppc64le" and platform_machine != "s390x"
-torchvision==0.22.0; platform_machine == "ppc64le"
+torchvision==0.23.0; platform_machine == "ppc64le"
 datasets # for benchmark scripts
 
 # Intel Extension for PyTorch, only for x86_64 CPUs

requirements/cuda.txt

@@ -6,9 +6,9 @@ numba == 0.61.2; python_version > '3.9'
 
 # Dependencies for NVIDIA GPUs
 ray[cgraph]>=2.48.0 # Ray Compiled Graph, required for pipeline parallelism in V1.
-torch==2.7.1
-torchaudio==2.7.1
+torch==2.8.0
+torchaudio==2.8.0
 # These must be updated alongside torch
-torchvision==0.22.1 # Required for phi3v processor. See https://github.com/pytorch/vision?tab=readme-ov-file#installation for corresponding version
-# https://github.com/facebookresearch/xformers/releases/tag/v0.0.31
-xformers==0.0.31; platform_system == 'Linux' and platform_machine == 'x86_64'  # Requires PyTorch >= 2.7
\ No newline at end of file
+torchvision==0.23.0 # Required for phi3v processor. See https://github.com/pytorch/vision?tab=readme-ov-file#installation for corresponding version
+# https://github.com/facebookresearch/xformers/releases/tag/v0.0.32.post1
+xformers==0.0.32.post1; platform_system == 'Linux' and platform_machine == 'x86_64'  # Requires PyTorch >= 2.8

requirements/docs.txt

@@ -7,27 +7,12 @@ mkdocs-awesome-nav
 mkdocs-glightbox
 mkdocs-git-revision-date-localized-plugin
 mkdocs-minify-plugin
-python-markdown-math
 regex
 ruff
 
 # Required for argparse hook only
 -f https://download.pytorch.org/whl/cpu
 cachetools
-cbor2
-cloudpickle
-fastapi
 msgspec
-openai
-openai-harmony
-partial-json-parser
-pillow
-psutil
-pybase64
 pydantic
-setproctitle
 torch
-transformers
-zmq
-uvloop
-prometheus-client

requirements/nightly_torch_test.txt

@@ -27,7 +27,7 @@ mistral_common[image,audio] >= 1.8.2 # required for voxtral test
 num2words # required for smolvlm test
 opencv-python-headless >= 4.11.0 # required for video test
 datamodel_code_generator # required for minicpm3 test
-lm-eval[api]==0.4.8 # required for model evaluation test
+lm-eval[api] @ git+https://github.com/EleutherAI/lm-evaluation-harness.git@206b7722158f58c35b7ffcd53b035fdbdda5126d # required for model evaluation test
 mteb>=1.38.11, <2 # required for mteb test
 transformers==4.52.4
 tokenizers==0.21.1

requirements/rocm-build.txt

 # Common dependencies
 -r common.txt
 
---extra-index-url https://download.pytorch.org/whl/rocm6.2.4
-torch==2.7.0
-torchvision==0.22.0
-torchaudio==2.7.0
+--extra-index-url https://download.pytorch.org/whl/rocm6.3
+torch==2.8.0
+torchvision==0.23.0
+torchaudio==2.8.0
 
-triton==3.2
+triton==3.3.0
 cmake>=3.26.1,<4
 packaging>=24.2
 setuptools>=77.0.3,<80.0.0

requirements/test.in

@@ -22,9 +22,9 @@ sentence-transformers # required for embedding tests
 soundfile # required for audio tests
 jiwer # required for audio tests
 timm >=1.0.17 # required for internvl and gemma3n-mm test
-torch==2.7.1
-torchaudio==2.7.1
-torchvision==0.22.1
+torch==2.8.0
+torchaudio==2.8.0
+torchvision==0.23.0
 transformers_stream_generator # required for qwen-vl test
 matplotlib # required for qwen-vl test
 mistral_common[image,audio] >= 1.8.2 # required for voxtral test
@@ -32,9 +32,10 @@ num2words # required for smolvlm test
 open_clip_torch==2.32.0 # Required for nemotron_vl test
 opencv-python-headless >= 4.11.0 # required for video test
 datamodel_code_generator # required for minicpm3 test
-lm-eval[api]==0.4.8 # required for model evaluation test
+# TODO: Use lm-eval[api]==0.4.10 once released
+lm-eval[api] @ git+https://github.com/EleutherAI/lm-evaluation-harness.git@206b7722158f58c35b7ffcd53b035fdbdda5126d # required for model evaluation test
 mteb[bm25s]>=1.38.11, <2 # required for mteb test
-transformers==4.55.0
+transformers==4.55.2
 tokenizers==0.21.1
 schemathesis>=3.39.15 # Required for openai schema test.
 # quantization
@@ -53,3 +54,4 @@ runai-model-streamer-s3==0.11.0
 fastsafetensors>=0.1.10
 pydantic>=2.10 # 2.9 leads to error on python 3.10
 terratorch==1.1rc2 # required for PrithviMAE test
+decord==0.6.0