from local

egs/librispeech/ASR/run_v3.sh

@@ -0,0 +1,34 @@
+git pull
+
+#export CUDA_VISIBLE_DEVICES="0,1,2,3,4,5,6,7"
+#export CUDA_VISIBLE_DEVICES="4,5,6,7"
+
+./pruned_transducer_stateless_d2v/train_wandb.py \
+	--wandb true \
+	--input-strategy AudioSamples \
+	--enable-spec-aug False \
+	--enable-musan True \
+	--multi-optim True \
+	--world-size 8 \
+	--num-epochs 30 \
+	--start-epoch 1 \
+	--full-libri 0 \
+	--exp-dir ./pruned_transducer_stateless_d2v/$1 \
+	--max-duration 200 \
+	--freeze-finetune-updates 2000 \
+	--use-fp16 1 \
+	--peak-enc-lr 0.0001 \
+	--peak-dec-lr 0.003 \
+	--accum-grads 1 \
+	--encoder-type d2v \
+	--additional-block True \
+	--encoder-dim 768 \
+	--decoder-dim 768 \
+	--joiner-dim 768 \
+	--prune-range 20 \
+	--context-size 2 \
+
+#./pruned_transducer_stateless_d2v/train.py \
+#--initial-lr 0.0003 \
+#--decoding-method greedy_search
+