Clean dummy training.

added_tokens.json

@@ -1 +0,0 @@
-{"<s>": 51, "</s>": 52}

all_results.json

@@ -1,14 +0,0 @@
-{
-    "epoch": 0.03,
-    "eval_loss": 155.7938690185547,
-    "eval_runtime": 55.6971,
-    "eval_samples": 301,
-    "eval_samples_per_second": 5.404,
-    "eval_steps_per_second": 0.682,
-    "eval_wer": 1.084206116464181,
-    "train_loss": 93.28157958984374,
-    "train_runtime": 23.9574,
-    "train_samples": 704,
-    "train_samples_per_second": 0.835,
-    "train_steps_per_second": 0.417
-}

checkpoint-10/config.json

@@ -1,96 +0,0 @@
-{
-  "_name_or_path": "hf-test/xls-r-dummy",
-  "activation_dropout": 0.0,
-  "adapter_kernel_size": 3,
-  "adapter_stride": 2,
-  "add_adapter": false,
-  "apply_spec_augment": true,
-  "architectures": [
-    "Wav2Vec2ForCTC"
-  ],
-  "attention_dropout": 0.0,
-  "bos_token_id": 1,
-  "classifier_proj_size": 256,
-  "codevector_dim": 256,
-  "contrastive_logits_temperature": 0.1,
-  "conv_bias": false,
-  "conv_dim": [
-    32,
-    32,
-    32
-  ],
-  "conv_kernel": [
-    8,
-    8,
-    8
-  ],
-  "conv_stride": [
-    4,
-    4,
-    4
-  ],
-  "ctc_loss_reduction": "mean",
-  "ctc_zero_infinity": false,
-  "diversity_loss_weight": 0.1,
-  "do_stable_layer_norm": true,
-  "eos_token_id": 2,
-  "feat_extract_activation": "gelu",
-  "feat_extract_dropout": 0.0,
-  "feat_extract_norm": "layer",
-  "feat_proj_dropout": 0.0,
-  "feat_quantizer_dropout": 0.0,
-  "final_dropout": 0.0,
-  "hidden_act": "gelu",
-  "hidden_dropout": 0.0,
-  "hidden_dropout_prob": 0.1,
-  "hidden_size": 16,
-  "initializer_range": 0.02,
-  "intermediate_size": 20,
-  "layer_norm_eps": 1e-05,
-  "layerdrop": 0.0,
-  "mask_feature_length": 10,
-  "mask_feature_min_masks": 0,
-  "mask_feature_prob": 0.0,
-  "mask_time_length": 10,
-  "mask_time_min_masks": 2,
-  "mask_time_prob": 0.05,
-  "model_type": "wav2vec2",
-  "num_adapter_layers": 3,
-  "num_attention_heads": 2,
-  "num_codevector_groups": 2,
-  "num_codevectors_per_group": 320,
-  "num_conv_pos_embedding_groups": 2,
-  "num_conv_pos_embeddings": 16,
-  "num_feat_extract_layers": 3,
-  "num_hidden_layers": 4,
-  "num_negatives": 10,
-  "output_hidden_size": 16,
-  "pad_token_id": 50,
-  "proj_codevector_dim": 256,
-  "tdnn_dilation": [
-    1,
-    2,
-    3,
-    1,
-    1
-  ],
-  "tdnn_dim": [
-    512,
-    512,
-    512,
-    512,
-    1500
-  ],
-  "tdnn_kernel": [
-    5,
-    3,
-    3,
-    1,
-    1
-  ],
-  "torch_dtype": "float32",
-  "transformers_version": "4.16.0.dev0",
-  "use_weighted_layer_sum": false,
-  "vocab_size": 53,
-  "xvector_output_dim": 512
-}

checkpoint-10/optimizer.pt

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:95e07815b6a602c91d38abaafb686592e1aeb8e52ba77b67b766cd507a56b076
-size 128273

checkpoint-10/preprocessor_config.json

@@ -1,9 +0,0 @@
-{
-  "do_normalize": true,
-  "feature_extractor_type": "Wav2Vec2FeatureExtractor",
-  "feature_size": 1,
-  "padding_side": "right",
-  "padding_value": 0.0,
-  "return_attention_mask": false,
-  "sampling_rate": 16000
-}

checkpoint-10/pytorch_model.bin

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:eb8eef8a9fa7d277d5cf7f07d589b29066da209acdca616f16a416b99f7829cf
-size 142694

checkpoint-10/rng_state.pth

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:1fb7f4cc8611df92247513a82d0d9f686595c80e63eeeaec2b5122289dfd478f
-size 13611

checkpoint-10/scheduler.pt

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:bd0948b7f2428191a1f974dc186da72bec015e92c49a1d4aa9e79f447a92a62b
-size 623

checkpoint-10/trainer_state.json

@@ -1,15 +0,0 @@
-{
-  "best_metric": null,
-  "best_model_checkpoint": null,
-  "epoch": 0.028409090909090908,
-  "global_step": 10,
-  "is_hyper_param_search": false,
-  "is_local_process_zero": true,
-  "is_world_process_zero": true,
-  "log_history": [],
-  "max_steps": 10,
-  "num_train_epochs": 1,
-  "total_flos": 508182128640.0,
-  "trial_name": null,
-  "trial_params": null
-}

checkpoint-10/training_args.bin

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:0d984518be407c10d2f161021e8f4824522fcbc805d42098d3c9a603e88255e3
-size 2991

config.json

@@ -1,96 +0,0 @@
-{
-  "_name_or_path": "hf-test/xls-r-dummy",
-  "activation_dropout": 0.0,
-  "adapter_kernel_size": 3,
-  "adapter_stride": 2,
-  "add_adapter": false,
-  "apply_spec_augment": true,
-  "architectures": [
-    "Wav2Vec2ForCTC"
-  ],
-  "attention_dropout": 0.0,
-  "bos_token_id": 1,
-  "classifier_proj_size": 256,
-  "codevector_dim": 256,
-  "contrastive_logits_temperature": 0.1,
-  "conv_bias": false,
-  "conv_dim": [
-    32,
-    32,
-    32
-  ],
-  "conv_kernel": [
-    8,
-    8,
-    8
-  ],
-  "conv_stride": [
-    4,
-    4,
-    4
-  ],
-  "ctc_loss_reduction": "mean",
-  "ctc_zero_infinity": false,
-  "diversity_loss_weight": 0.1,
-  "do_stable_layer_norm": true,
-  "eos_token_id": 2,
-  "feat_extract_activation": "gelu",
-  "feat_extract_dropout": 0.0,
-  "feat_extract_norm": "layer",
-  "feat_proj_dropout": 0.0,
-  "feat_quantizer_dropout": 0.0,
-  "final_dropout": 0.0,
-  "hidden_act": "gelu",
-  "hidden_dropout": 0.0,
-  "hidden_dropout_prob": 0.1,
-  "hidden_size": 16,
-  "initializer_range": 0.02,
-  "intermediate_size": 20,
-  "layer_norm_eps": 1e-05,
-  "layerdrop": 0.0,
-  "mask_feature_length": 10,
-  "mask_feature_min_masks": 0,
-  "mask_feature_prob": 0.0,
-  "mask_time_length": 10,
-  "mask_time_min_masks": 2,
-  "mask_time_prob": 0.05,
-  "model_type": "wav2vec2",
-  "num_adapter_layers": 3,
-  "num_attention_heads": 2,
-  "num_codevector_groups": 2,
-  "num_codevectors_per_group": 320,
-  "num_conv_pos_embedding_groups": 2,
-  "num_conv_pos_embeddings": 16,
-  "num_feat_extract_layers": 3,
-  "num_hidden_layers": 4,
-  "num_negatives": 10,
-  "output_hidden_size": 16,
-  "pad_token_id": 50,
-  "proj_codevector_dim": 256,
-  "tdnn_dilation": [
-    1,
-    2,
-    3,
-    1,
-    1
-  ],
-  "tdnn_dim": [
-    512,
-    512,
-    512,
-    512,
-    1500
-  ],
-  "tdnn_kernel": [
-    5,
-    3,
-    3,
-    1,
-    1
-  ],
-  "torch_dtype": "float32",
-  "transformers_version": "4.16.0.dev0",
-  "use_weighted_layer_sum": false,
-  "vocab_size": 53,
-  "xvector_output_dim": 512
-}

eval_results.json

@@ -1,9 +0,0 @@
-{
-    "epoch": 0.03,
-    "eval_loss": 155.7938690185547,
-    "eval_runtime": 55.6971,
-    "eval_samples": 301,
-    "eval_samples_per_second": 5.404,
-    "eval_steps_per_second": 0.682,
-    "eval_wer": 1.084206116464181
-}

preprocessor_config.json

@@ -1,9 +0,0 @@
-{
-  "do_normalize": true,
-  "feature_extractor_type": "Wav2Vec2FeatureExtractor",
-  "feature_size": 1,
-  "padding_side": "right",
-  "padding_value": 0.0,
-  "return_attention_mask": false,
-  "sampling_rate": 16000
-}

pytorch_model.bin

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:eb8eef8a9fa7d277d5cf7f07d589b29066da209acdca616f16a416b99f7829cf
-size 142694

run-cpu.sh

@@ -1,21 +1,33 @@
 python run_speech_recognition_ctc.py \
 	--dataset_name="mozilla-foundation/common_voice_7_0" \
-	--model_name_or_path="hf-test/xls-r-dummy" \
-	--dataset_config_name="ab" \
+	--model_name_or_path="facebook/wav2vec2-xls-r-300m" \
+	--dataset_config_name="sv-SE" \
 	--output_dir="./" \
 	--overwrite_output_dir \
 	--max_steps="10" \
-	--per_device_train_batch_size="2" \
-	--learning_rate="3e-4" \
-	--save_total_limit="1" \
+	--per_device_train_batch_size="8" \
+	--per_device_eval_batch_size="8" \
+	--gradient_accumulation_steps="4" \
+	--learning_rate="7.5e-5" \
+	--warmup_steps="2000" \
+	--length_column_name="input_length" \
 	--evaluation_strategy="steps" \
 	--text_column_name="sentence" \
-	--length_column_name="input_length" \
+	--chars_to_ignore , ? . ! \- \; \: \" “ % ‘ ” � — ’ … – \
 	--save_steps="5" \
+	--eval_steps="5" \
+	--logging_steps="1" \
 	--layerdrop="0.0" \
+	--activation_dropout="0.1" \
+	--save_total_limit="3" \
 	--freeze_feature_encoder \
+	--feat_proj_dropout="0.0" \
+	--mask_time_prob="0.75" \
+	--mask_time_length="10" \
+	--mask_feature_prob="0.25" \
+	--mask_feature_length="64" \
 	--gradient_checkpointing \
-	--group_by_length \
-	--push_to_hub \
 	--use_auth_token \
-	--do_train --do_eval
+	--group_by_length \
+	--do_train --do_eval \
+	--push_to_hub

run.sh

@@ -1,22 +1,34 @@
 python run_speech_recognition_ctc.py \
 	--dataset_name="mozilla-foundation/common_voice_7_0" \
-	--model_name_or_path="hf-test/xls-r-dummy" \
-	--dataset_config_name="ab" \
+	--model_name_or_path="facebook/wav2vec2-xls-r-300m" \
+	--dataset_config_name="sv-SE" \
 	--output_dir="./" \
 	--overwrite_output_dir \
-	--max_steps="10" \
-	--per_device_train_batch_size="2" \
-	--learning_rate="3e-4" \
-	--save_total_limit="1" \
+	--num_train_epochs="50" \
+	--per_device_train_batch_size="8" \
+	--per_device_eval_batch_size="8" \
+	--gradient_accumulation_steps="4" \
+	--learning_rate="7.5e-5" \
+	--warmup_steps="2000" \
+	--length_column_name="input_length" \
 	--evaluation_strategy="steps" \
 	--text_column_name="sentence" \
-	--length_column_name="input_length" \
-	--save_steps="5" \
+	--chars_to_ignore , ? . ! \- \; \: \" “ % ‘ ” � — ’ … – \
+	--save_steps="500" \
+	--eval_steps="500" \
+	--logging_steps="100" \
 	--layerdrop="0.0" \
+	--activation_dropout="0.1" \
+	--save_total_limit="3" \
 	--freeze_feature_encoder \
+	--feat_proj_dropout="0.0" \
+	--mask_time_prob="0.75" \
+	--mask_time_length="10" \
+	--mask_feature_prob="0.25" \
+	--mask_feature_length="64" \
 	--gradient_checkpointing \
-	# --fp16 \
-	--group_by_length \
-	--push_to_hub \
 	--use_auth_token \
-	--do_train --do_eval
+	--fp16 \
+	--group_by_length \
+	--do_train --do_eval \
+	--push_to_hub

special_tokens_map.json

@@ -1 +0,0 @@
-{"bos_token": "<s>", "eos_token": "</s>", "unk_token": "[UNK]", "pad_token": "[PAD]", "additional_special_tokens": [{"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}]}

tokenizer_config.json

@@ -1 +0,0 @@
-{"unk_token": "[UNK]", "bos_token": "<s>", "eos_token": "</s>", "pad_token": "[PAD]", "do_lower_case": false, "word_delimiter_token": "|", "special_tokens_map_file": null, "tokenizer_file": null, "name_or_path": "./", "tokenizer_class": "Wav2Vec2CTCTokenizer"}

train_results.json

@@ -1,8 +0,0 @@
-{
-    "epoch": 0.03,
-    "train_loss": 93.28157958984374,
-    "train_runtime": 23.9574,
-    "train_samples": 704,
-    "train_samples_per_second": 0.835,
-    "train_steps_per_second": 0.417
-}

trainer_state.json

@@ -1,25 +0,0 @@
-{
-  "best_metric": null,
-  "best_model_checkpoint": null,
-  "epoch": 0.028409090909090908,
-  "global_step": 10,
-  "is_hyper_param_search": false,
-  "is_local_process_zero": true,
-  "is_world_process_zero": true,
-  "log_history": [
-    {
-      "epoch": 0.03,
-      "step": 10,
-      "total_flos": 508182128640.0,
-      "train_loss": 93.28157958984374,
-      "train_runtime": 23.9574,
-      "train_samples_per_second": 0.835,
-      "train_steps_per_second": 0.417
-    }
-  ],
-  "max_steps": 10,
-  "num_train_epochs": 1,
-  "total_flos": 508182128640.0,
-  "trial_name": null,
-  "trial_params": null
-}

training_args.bin

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:0d984518be407c10d2f161021e8f4824522fcbc805d42098d3c9a603e88255e3
-size 2991

vocab.json

@@ -1 +0,0 @@
-{"!": 1, ",": 2, "-": 3, ".": 4, ":": 5, ";": 6, "?": 7, "а": 8, "б": 9, "в": 10, "г": 11, "д": 12, "е": 13, "ж": 14, "з": 15, "и": 16, "к": 17, "л": 18, "м": 19, "н": 20, "о": 21, "п": 22, "р": 23, "с": 24, "т": 25, "у": 26, "ф": 27, "х": 28, "ц": 29, "ч": 30, "ш": 31, "ы": 32, "ь": 33, "џ": 34, "қ": 35, "ҟ": 36, "ҩ": 37, "ҭ": 38, "ҳ": 39, "ҵ": 40, "ҷ": 41, "ҽ": 42, "ҿ": 43, "ә": 44, "ӡ": 45, "ӷ": 46, "ԥ": 47, "–": 48, "|": 0, "[UNK]": 49, "[PAD]": 50}