fix flashlight version to use before it is reorganization

.circleci/config.yml

@@ -60,7 +60,7 @@ jobs:
             apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends gcc-5 g++-5 && \
             export CC=/usr/bin/gcc-5 && export CXX=/usr/bin/g++-5 && \
             git clone --recursive https://github.com/facebookresearch/flashlight.git && \
-            cd flashlight && mkdir -p build && cd build && \
+            cd flashlight && git checkout v0.2 && mkdir -p build && cd build && \
             cmake .. -DCMAKE_BUILD_TYPE=Debug -DFLASHLIGHT_BACKEND=CUDA -DFL_BUILD_CONTRIB=ON -DFL_BUILD_TESTS=OFF -DFL_BUILD_EXAMPLES=OFF && \
             make install -j$(proc) && \
             export MKLROOT=/opt/intel/mkl && export KENLM_ROOT_DIR=/root/kenlm && \
@@ -98,7 +98,7 @@ jobs:
             apt-get update && DEBIAN_FRONTEND=noninteractive apt-get install -y --no-install-recommends gcc-5 g++-5 && \
             export MKLROOT=/opt/intel/mkl && export CC=/usr/bin/gcc-5 && export CXX=/usr/bin/g++-5 && \
             cd /root && git clone --recursive https://github.com/facebookresearch/flashlight.git && \
-            cd flashlight && mkdir -p build && cd build && \
+            cd flashlight && git checkout v0.2 && mkdir -p build && cd build && \
             cmake .. -DCMAKE_BUILD_TYPE=Debug -DFLASHLIGHT_BACKEND=CPU -DFL_BUILD_CONTRIB=ON -DFL_BUILD_TESTS=OFF -DFL_BUILD_EXAMPLES=OFF && \
             make install
             export KENLM_ROOT_DIR=/root/kenlm && \

Dockerfile-CPU

@@ -14,7 +14,7 @@ COPY . /root/wav2letter
 # ------------------------------------------------------------------
 RUN export MKLROOT=/opt/intel/mkl && \
     cd /root && git clone --recursive https://github.com/facebookresearch/flashlight.git && \
-    cd flashlight && mkdir -p build && cd build && \
+    cd flashlight && git checkout v0.2 && mkdir -p build && cd build && \
     cmake .. -DCMAKE_BUILD_TYPE=Release -DFLASHLIGHT_BACKEND=CPU && \
     make -j$(nproc) && make install && \
 # ==================================================================

Dockerfile-CUDA

@@ -13,7 +13,7 @@ COPY . /root/wav2letter
 # flashlight https://github.com/facebookresearch/flashlight.git
 # ------------------------------------------------------------------
 RUN cd /root && git clone --recursive https://github.com/facebookresearch/flashlight.git && \
-    cd /root/flashlight && mkdir -p build && \
+    cd /root/flashlight && git checkout v0.2 && mkdir -p build && \
     cd build && cmake .. -DCMAKE_BUILD_TYPE=Release -DFLASHLIGHT_BACKEND=CUDA && \
     make -j$(nproc) && make install && \
 # ==================================================================

README.md

@@ -4,7 +4,7 @@
 [![](https://github.com/facebookresearch/wav2letter/workflows/Publish%20Docker%20images/badge.svg)](https://hub.docker.com/r/wav2letter/wav2letter/tags)
 [![Join the chat at https://gitter.im/wav2letter/community](https://badges.gitter.im/wav2letter/community.svg)](https://gitter.im/wav2letter/community?utm_source=badge&utm_medium=badge&utm_campaign=pr-badge&utm_content=badge)
 
-wav2letter++ is a fast, open source speech processing toolkit from the Speech team at Facebook AI Research built to facilitate research in end-to-end models for speech recognition. It is written entirely in C++ and uses the [ArrayFire](https://github.com/arrayfire/arrayfire) tensor library and the [flashlight](https://github.com/facebookresearch/flashlight) machine learning library for maximum efficiency. Our approach is detailed in this [arXiv paper](https://arxiv.org/abs/1812.07625).
+wav2letter++ is a fast, open source speech processing toolkit from the Speech team at Facebook AI Research built to facilitate research in end-to-end models for speech recognition. It is written entirely in C++ and uses the [ArrayFire](https://github.com/arrayfire/arrayfire) tensor library and the [flashlight (use its branch v0.2)](https://github.com/facebookresearch/flashlight) machine learning library for maximum efficiency. Our approach is detailed in this [arXiv paper](https://arxiv.org/abs/1812.07625).
 
 This repository also contains **pre-trained** models and implementations for various ASR results including:
 - [NEW] [Pratap et al. (2020): Scaling Online Speech Recognition Using ConvNets](recipes/models/streaming_convnets/)