feat(vllm): bump to 0.2.2 (#695)

* feat(vllm): bump to 0.2.2

Signed-off-by: Aaron <29749331+aarnphm@users.noreply.github.com>

* chore: update changelog

Signed-off-by: Aaron <29749331+aarnphm@users.noreply.github.com>

* chore: move up to CUDA 12.1

Signed-off-by: Aaron <29749331+aarnphm@users.noreply.github.com>

* fix: remove auto-gptq installation

since the builder image doesn't have access to GPU

Signed-off-by: Aaron <29749331+aarnphm@users.noreply.github.com>

* fix: update containerization warning

Signed-off-by: Aaron <29749331+aarnphm@users.noreply.github.com>

---------

Signed-off-by: Aaron <29749331+aarnphm@users.noreply.github.com>

openllm-python/src/openllm/_llm.py

@@ -199,7 +199,7 @@ class LLM(t.Generic[M, T], ReprMixin):
       system_message=system_message,
       LLM__model_attrs=model_attrs,
       LLM__tokenizer_attrs=tokenizer_attrs,
-      llm_dtype__=torch_dtype.lower(),
+      llm_dtype__=dtype.lower(),
       llm_backend__=backend,
       llm_config__=llm_config,
       llm_trust_remote_code__=trust_remote_code,

openllm-python/src/openllm/bundle/_package.py

@@ -65,15 +65,7 @@ def construct_python_options(llm, llm_fs, extra_dependencies=None, adapter_map=N
   built_wheels = [build_editable(llm_fs.getsyspath('/'), p) for p in ('openllm_core', 'openllm_client', 'openllm')]
   if all(i for i in built_wheels):
     wheels = [llm_fs.getsyspath(f"/{i.split('/')[-1]}") for i in built_wheels]
-  return PythonOptions(
-    packages=packages,
-    wheels=wheels,
-    lock_packages=True,
-    extra_index_url=[
-      'https://download.pytorch.org/whl/cu118',
-      'https://huggingface.github.io/autogptq-index/whl/cu118/',
-    ],
-  )
+  return PythonOptions(packages=packages, wheels=wheels, lock_packages=True)
 
 
 def construct_docker_options(

openllm-python/src/openllm/bundle/oci/Dockerfile

@@ -10,13 +10,13 @@ ENV PATH /opt/conda/bin:$PATH
 ENV DEBIAN_FRONTEND=noninteractive
 
 RUN apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends \
-        build-essential \
-        ca-certificates \
-        ccache \
-        curl \
-        libssl-dev ca-certificates make \
-        git && \
-        rm -rf /var/lib/apt/lists/*
+  build-essential \
+  ca-certificates \
+  ccache \
+  curl \
+  libssl-dev ca-certificates make \
+  git && \
+  rm -rf /var/lib/apt/lists/*
 
 
 # Install required dependencies
@@ -29,8 +29,8 @@ COPY hatch.toml README.md CHANGELOG.md openllm-python/pyproject.toml ./
 # pip install autoawq --no-cache-dir && \
 RUN --mount=type=cache,target=/root/.cache/pip \
   pip install --extra-index-url "https://huggingface.github.io/autogptq-index/whl/cu118/" \
-              -v --no-cache-dir \
-              "ray==2.6.0" "einops" "vllm==0.2.1.post1" "auto-gptq[triton]" "torch==2.0.1" xformers && \
+  -v --no-cache-dir \
+  "ray==2.6.0" "vllm==0.2.2" xformers && \
   pip install --no-cache-dir -e .
 
 FROM base-container

openllm-python/src/openllm/bundle/oci/__init__.py

@@ -50,12 +50,13 @@ class RefResolver:
     else:
       raise ValueError(f'Unknown strategy: {strategy_or_version}')
 
+  # fmt: off
   @property
-  def tag(self):
-    return 'latest' if self.strategy in {'latest', 'nightly'} else repr(self.version)
-
+  def tag(self):return 'latest' if self.strategy in {'latest','nightly'} else repr(self.version)
   @staticmethod
-  def construct_base_image(reg, strategy=None):
+  def construct_base_image(reg,strategy=None):
+    if reg == 'gh': logger.warning("Setting base registry to 'gh' will affect cold start performance on GCP/AWS.")
+    elif reg == 'docker': logger.warning('docker is base image is yet to be supported. Falling back to "ecr".'); reg = 'ecr'
     return f'{_CONTAINER_REGISTRY[reg]}:{RefResolver.from_strategy(strategy).tag}'