🚨🚨🚨Deprecate evaluation_strategy to eval_strategy🚨🚨🚨 (#30190)

* Alias

* Note alias

* Tests and src

* Rest

* Clean

* Change typing?

* Fix tests

* Deprecation versions

examples/research_projects/wav2vec2/FINE_TUNE_XLSR_WAV2VEC2.md

@@ -182,7 +182,7 @@ Here we will run the script on the *Turkish* Common Voice dataset for demonstrat
 		--per_device_train_batch_size="16" \
 		--learning_rate="3e-4" \
 		--warmup_steps="500" \
-		--evaluation_strategy="steps" \
+		--eval_strategy="steps" \
 		--save_steps="400" \
 		--eval_steps="400" \
 		--logging_steps="400" \
@@ -209,7 +209,7 @@ Here we will run the script on the *Turkish* Common Voice dataset for demonstrat
 		--per_device_train_batch_size="16" \
 		--learning_rate="3e-4" \
 		--warmup_steps="500" \
-		--evaluation_strategy="steps" \
+		--eval_strategy="steps" \
 		--save_steps="400" \
 		--eval_steps="400" \
 		--logging_steps="400" \

examples/research_projects/wav2vec2/README.md

@@ -18,7 +18,7 @@ python run_asr.py \
 --num_train_epochs="30" \
 --per_device_train_batch_size="20" \
 --per_device_eval_batch_size="20" \
---evaluation_strategy="steps" \
+--eval_strategy="steps" \
 --save_steps="500" \
 --eval_steps="100" \
 --logging_steps="50" \
@@ -73,7 +73,7 @@ python run_asr.py \
 --per_device_train_batch_size="1" \
 --per_device_eval_batch_size="1" \
 --gradient_accumulation_steps="8" \
---evaluation_strategy="steps" \
+--eval_strategy="steps" \
 --save_steps="500" \
 --eval_steps="100" \
 --logging_steps="50" \
@@ -152,7 +152,7 @@ ZeRO-2:
 PYTHONPATH=../../../src deepspeed --num_gpus 2 \
 run_asr.py \
 --output_dir=output_dir --num_train_epochs=2 --per_device_train_batch_size=2 \
---per_device_eval_batch_size=2 --evaluation_strategy=steps --save_steps=500 --eval_steps=100 \
+--per_device_eval_batch_size=2 --eval_strategy=steps --save_steps=500 --eval_steps=100 \
 --logging_steps=5 --learning_rate=5e-4 --warmup_steps=3000 \
 --model_name_or_path=patrickvonplaten/wav2vec2_tiny_random_robust \
 --dataset_name=hf-internal-testing/librispeech_asr_dummy --dataset_config_name=clean \
@@ -176,7 +176,7 @@ ZeRO-3:
 PYTHONPATH=../../../src deepspeed --num_gpus 2 \
 run_asr.py \
 --output_dir=output_dir --num_train_epochs=2 --per_device_train_batch_size=2 \
---per_device_eval_batch_size=2 --evaluation_strategy=steps --save_steps=500 --eval_steps=100 \
+--per_device_eval_batch_size=2 --eval_strategy=steps --save_steps=500 --eval_steps=100 \
 --logging_steps=5 --learning_rate=5e-4 --warmup_steps=3000 \
 --model_name_or_path=patrickvonplaten/wav2vec2_tiny_random_robust \
 --dataset_name=hf-internal-testing/librispeech_asr_dummy --dataset_config_name=clean \

examples/research_projects/wav2vec2/finetune_base_100.sh

@@ -4,7 +4,7 @@ python run_asr.py \
 --num_train_epochs="30" \
 --per_device_train_batch_size="32" \
 --per_device_eval_batch_size="32" \
---evaluation_strategy="steps" \
+--eval_strategy="steps" \
 --save_total_limit="3" \
 --save_steps="500" \
 --eval_steps="100" \

examples/research_projects/wav2vec2/finetune_base_timit_asr.sh

@@ -4,7 +4,7 @@ python run_asr.py \
 --num_train_epochs="30" \
 --per_device_train_batch_size="20" \
 --per_device_eval_batch_size="20" \
---evaluation_strategy="steps" \
+--eval_strategy="steps" \
 --save_steps="500" \
 --eval_steps="100" \
 --logging_steps="50" \

examples/research_projects/wav2vec2/finetune_large_lv60_100.sh

@@ -4,7 +4,7 @@ python run_asr.py \
 --num_train_epochs="30" \
 --per_device_train_batch_size="16" \
 --per_device_eval_batch_size="16" \
---evaluation_strategy="steps" \
+--eval_strategy="steps" \
 --save_total_limit="3" \
 --save_steps="500" \
 --eval_steps="100" \

examples/research_projects/wav2vec2/finetune_large_lv60_timit_asr.sh

@@ -5,7 +5,7 @@ python run_asr.py \
 --per_device_train_batch_size="2" \
 --per_device_eval_batch_size="2" \
 --gradient_accumulation_steps="4" \
---evaluation_strategy="steps" \
+--eval_strategy="steps" \
 --save_steps="500" \
 --eval_steps="100" \
 --logging_steps="50" \

examples/research_projects/wav2vec2/finetune_large_xlsr_53_arabic_speech_corpus.sh

@@ -5,7 +5,7 @@ python run_asr.py \
 --per_device_train_batch_size="1" \
 --per_device_eval_batch_size="1" \
 --gradient_accumulation_steps="8" \
---evaluation_strategy="steps" \
+--eval_strategy="steps" \
 --save_steps="500" \
 --eval_steps="100" \
 --logging_steps="50" \

examples/research_projects/wav2vec2/finetune_wav2vec2_xlsr_turkish.sh

@@ -6,7 +6,7 @@ python run_common_voice.py \
     --overwrite_output_dir \
     --num_train_epochs="5" \
     --per_device_train_batch_size="16" \
-    --evaluation_strategy="steps" \
+    --eval_strategy="steps" \
     --learning_rate="3e-4" \
     --warmup_steps="500" \
     --fp16 \

examples/research_projects/wav2vec2/test_wav2vec2_deepspeed.py

@@ -161,7 +161,7 @@ class TestDeepSpeedWav2Vec2(TestCasePlus):
             --num_train_epochs {str(num_train_epochs)}
             --per_device_train_batch_size 2
             --per_device_eval_batch_size 2
-            --evaluation_strategy steps
+            --eval_strategy steps
             --learning_rate 5e-4
             --warmup_steps 8
             --orthography timit