Upgrade python version from 3.10 to 3.12 (#782)

updating the Qeff python version to 3.12 still keeping support for 3.10
3.11.

Signed-off-by: Rishin Raj <rishinr@qti.qualcomm.com>
Co-authored-by: Hem Agnihotri <hemagnih@qti.qualcomm.com>
Signed-off-by: Dipankar Sarkar <dipankar@qti.qualcomm.com>

Author: 2 people

Dockerfile

@@ -7,8 +7,8 @@ FROM docker-registry.qualcomm.com/library/ubuntu:20.04
 RUN apt-get update && apt-get install -y \
     git \
     tmux \
-    python3.10 \
-    python3.10-venv \
+    python3.12 \
+    python3.12-venv \
     python3-pip
 
 # pip recognizes this variable
@@ -24,7 +24,7 @@ RUN mkdir -p /app/qefficient-library
 COPY . /app/qefficient-library
 
 # Create Virtual Env for the docker image
-RUN python3.10 -m venv /app/llm_env
+RUN python3.12 -m venv /app/llm_env
 RUN . /app/llm_env/bin/activate
 WORKDIR /app/qefficient-library
 
@@ -33,7 +33,7 @@ WORKDIR /app/qefficient-library
 RUN pip install torch==2.0.0+cpu --extra-index-url https://download.pytorch.org/whl/cpu --no-deps
 RUN pip install datasets==2.17.0 fsspec==2023.10.0 multidict==6.0.5 sentencepiece --no-deps
 
-RUN python3.10 -m pip install .
+RUN python3.12 -m pip install .
 WORKDIR /app/qefficient-library
 
 # Set the environment variable for the model card name and token ID
@@ -45,7 +45,7 @@ ENV TOKEN_ID = ""
 # Print a success message
 CMD ["echo", "qefficient-transformers repository cloned and setup installed inside Docker image."]
 CMD ["echo", "Starting the Model Download and Export to Onnx Stage for QEff."]
-CMD python3.10 -m QEfficient.cloud.export --model-name "$MODEL_NAME"
+CMD python3.12 -m QEfficient.cloud.export --model-name "$MODEL_NAME"
 
 # Example usage:
 # docker build -t qefficient-library .
@@ -55,4 +55,4 @@ CMD python3.10 -m QEfficient.cloud.export --model-name "$MODEL_NAME"
 # 2. For smaller models, 32GiB RAM is sufficient, but larger LLMs we require good CPU/RAM (Context 7B model would require atleast 64GiB).
 # 3. The exact minimum system configuration are tough to decide, since its all function of model parameters.
 
-# docker run -e MODEL_NAME=gpt2 -e TOKEN_ID=<your-token-id> qefficient-library
+# docker run -e MODEL_NAME=gpt2 -e TOKEN_ID=<your-token-id> qefficient-library

README.md

@@ -95,9 +95,9 @@ For other models, there is comprehensive documentation to inspire upon the chang
 ## Quick Installation
 ```bash
 
-# Create Python virtual env and activate it. (Recommended Python 3.10)
-sudo apt install python3.10-venv
-python3.10 -m venv qeff_env
+# Create Python virtual env and activate it. (Recommended Python 3.12)
+sudo apt install python3.12-venv
+python3.12 -m venv qeff_env
 source qeff_env/bin/activate
 pip install -U pip
 
@@ -136,4 +136,4 @@ Thanks to:
 If you run into any problems with the code, please file Github issues directly to this repo.
 
 ## Contributing
-This project welcomes contributions and suggestions. Please check the License. Integration with a CLA Bot is underway. 
+This project welcomes contributions and suggestions. Please check the License. Integration with a CLA Bot is underway.

docs/source/finetune.md

@@ -11,7 +11,7 @@ For QEfficient Library : https://github.com/quic/efficient-transformers
 
 For torch_qaic, assuming QEfficient is already installed,
 ```bash
-pip install /opt/qti-aic/integrations/torch_qaic/py310/torch_qaic-0.1.0-cp310-cp310-linux_x86_64.whl
+pip install /opt/qti-aic/integrations/torch_qaic/py312/torch_qaic-0.1.0-cp312-cp312-linux_x86_64.whl
 ```
 If qeff-env inside docker is used then torch_qaic and accelerate packages are already installed.
 

docs/source/installation.md

@@ -48,7 +48,7 @@ Efficient Transformers have been validated to work with the same compatible SDK.
 ```bash
 # Create Python virtual env and activate it. (Required Python 3.10)
 
-python3.10 -m venv qeff_env
+python3.12 -m venv qeff_env
 source qeff_env/bin/activate
 pip install -U pip
 

examples/performance/on_device_sampling.py

@@ -114,7 +114,7 @@ def main(args, **kwargs):
     """
     Example usage:
     1. For continuous batching:
-        python3.10 examples/on_device_sampling.py \
+        python examples/on_device_sampling.py \
             --model-name 'meta-llama/Llama-3.1-8B' \
             --prompt-len 128 \
             --ctx-len 256 \
@@ -134,7 +134,7 @@ def main(args, **kwargs):
             --random-number 26
 
     2. For non-continuous batching:
-        python3.10 examples/on_device_sampling.py \
+        python examples/on_device_sampling.py \
             --model-name 'meta-llama/Llama-3.1-8B' \
             --prompt-len 128 \
             --ctx-len 256 \
@@ -154,7 +154,7 @@ def main(args, **kwargs):
             --random-number 26
 
     3. With guided decoding:
-        python3.10 examples/on_device_sampling.py \
+        python examples/on_device_sampling.py \
             --model-name 'meta-llama/Llama-3.1-8B' \
             --prompt-len 128 \
             --ctx-len 256 \

pyproject.toml

@@ -14,10 +14,10 @@ classifiers = [
     "Intended Audience :: Developers",
     "Intended Audience :: Education",
     "Operating System :: Linux",
-    "Programming Language :: Python :: 3.10",
+    "Programming Language :: Python :: 3.12",
     "Topic :: Scientific/Engineering :: Artificial Intelligence for Inference Accelerator",
 ]
-requires-python = ">=3.8,<3.11"
+requires-python = ">=3.8,<3.13"
 dependencies = [
     "transformers==4.57.0",
     "diffusers== 0.35.1",
@@ -48,8 +48,12 @@ dependencies = [
     "torch@https://download.pytorch.org/whl/cpu/torch-2.4.1%2Bcpu-cp38-cp38-linux_x86_64.whl ; python_version=='3.8' and platform_machine=='x86_64'",
     "torch@https://download.pytorch.org/whl/cpu/torch-2.7.0%2Bcpu-cp39-cp39-manylinux_2_28_x86_64.whl ; python_version=='3.9' and platform_machine=='x86_64'",
     "torch@https://download.pytorch.org/whl/cpu/torch-2.7.0%2Bcpu-cp310-cp310-manylinux_2_28_x86_64.whl ; python_version=='3.10' and platform_machine=='x86_64'",
+    "torch@https://download.pytorch.org/whl/cpu/torch-2.7.0%2Bcpu-cp311-cp311-manylinux_2_28_x86_64.whl ; python_version=='3.11' and platform_machine=='x86_64'",
+    "torch@https://download.pytorch.org/whl/cpu/torch-2.7.0%2Bcpu-cp312-cp312-manylinux_2_28_x86_64.whl ; python_version=='3.12' and platform_machine=='x86_64'",
     "torchvision@https://download.pytorch.org/whl/cpu/torchvision-0.22.0%2Bcpu-cp39-cp39-manylinux_2_28_x86_64.whl ; python_version=='3.9' and platform_machine=='x86_64'",
     "torchvision@https://download.pytorch.org/whl/cpu/torchvision-0.22.0%2Bcpu-cp310-cp310-manylinux_2_28_x86_64.whl ; python_version=='3.10' and platform_machine=='x86_64'",
+    "torchvision@https://download.pytorch.org/whl/cpu/torchvision-0.22.0%2Bcpu-cp311-cp311-manylinux_2_28_x86_64.whl ; python_version=='3.11' and platform_machine=='x86_64'",
+    "torchvision@https://download.pytorch.org/whl/cpu/torchvision-0.22.0%2Bcpu-cp312-cp312-manylinux_2_28_x86_64.whl ; python_version=='3.12' and platform_machine=='x86_64'",
 ]
 
 [project.optional-dependencies]

scripts/Jenkinsfile

@@ -17,8 +17,8 @@ pipeline {
                    sudo docker exec ${BUILD_TAG} bash -c "
                    cd /efficient-transformers &&
                    apt update &&
-                   apt install -y python3.10-venv &&
-                   python3.10 -m venv preflight_qeff &&
+                   DEBIAN_FRONTEND=noninteractive apt install -y tzdata python3.12-venv python3.12-dev build-essential &&
+                   python3.12 -m venv preflight_qeff &&
                    . preflight_qeff/bin/activate &&
                    pip install --upgrade pip setuptools &&
                    pip install .[test] &&
@@ -202,7 +202,9 @@ pipeline {
                     sudo docker exec ${BUILD_TAG} bash -c "
                     cd /efficient-transformers &&
                     . preflight_qeff/bin/activate &&
-                    pip install /opt/qti-aic/integrations/torch_qaic/py310/torch_qaic-0.1.0-cp310-cp310-linux_x86_64.whl &&
+                    # TODO: Update torch_qaic path to py312 when migrating to Python 3.12
+                    pip install /opt/qti-aic/integrations/torch_qaic/py312/torch_qaic-0.1.0-cp312-cp312-linux_x86_64.whl &&
+                    # pip install /opt/qti-aic/integrations/torch_qaic/py310/torch_qaic-0.1.0-cp310-cp310-linux_x86_64.whl &&
                     pip install torch==2.9.0 torchvision==0.24.0 torchaudio==2.9.0 --index-url https://download.pytorch.org/whl/cpu &&
                     mkdir -p $PWD/cli_qaic_finetuning &&
                     export TOKENIZERS_PARALLELISM=false &&