add files

.gitignore

@@ -0,0 +1,2 @@
+checkpoint-*
+*.txt

README.md

@@ -0,0 +1 @@
+FYTM_fb_all

all_results.json

@@ -0,0 +1,23 @@
+{
+    "epoch": 40.0,
+    "eval_loss": 0.508992612361908,
+    "eval_mem_cpu_alloc_delta": -233472,
+    "eval_mem_cpu_peaked_delta": 233472,
+    "eval_mem_gpu_alloc_delta": 0,
+    "eval_mem_gpu_peaked_delta": 807446528,
+    "eval_runtime": 33.0419,
+    "eval_samples": 405,
+    "eval_samples_per_second": 12.257,
+    "eval_wer": 0.36945244956772333,
+    "init_mem_cpu_alloc_delta": 0,
+    "init_mem_cpu_peaked_delta": 0,
+    "init_mem_gpu_alloc_delta": 0,
+    "init_mem_gpu_peaked_delta": 0,
+    "train_mem_cpu_alloc_delta": 428699648,
+    "train_mem_cpu_peaked_delta": 12730368,
+    "train_mem_gpu_alloc_delta": 3823077376,
+    "train_mem_gpu_peaked_delta": 1210148352,
+    "train_runtime": 7790.1813,
+    "train_samples": 1618,
+    "train_samples_per_second": 1.037
+}

config.json

@@ -0,0 +1,84 @@
+{
+  "_name_or_path": "facebook/wav2vec2-large-xlsr-53",
+  "activation_dropout": 0.1,
+  "apply_spec_augment": true,
+  "architectures": [
+    "Wav2Vec2ForCTC"
+  ],
+  "attention_dropout": 0.1,
+  "bos_token_id": 1,
+  "codevector_dim": 768,
+  "contrastive_logits_temperature": 0.1,
+  "conv_bias": true,
+  "conv_dim": [
+    512,
+    512,
+    512,
+    512,
+    512,
+    512,
+    512
+  ],
+  "conv_kernel": [
+    10,
+    3,
+    3,
+    3,
+    3,
+    2,
+    2
+  ],
+  "conv_stride": [
+    5,
+    2,
+    2,
+    2,
+    2,
+    2,
+    2
+  ],
+  "ctc_loss_reduction": "mean",
+  "ctc_zero_infinity": true,
+  "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,
+  "gradient_checkpointing": true,
+  "hidden_act": "gelu",
+  "hidden_dropout": 0.1,
+  "hidden_size": 1024,
+  "initializer_range": 0.02,
+  "intermediate_size": 4096,
+  "layer_norm_eps": 1e-05,
+  "layerdrop": 0.0,
+  "mask_channel_length": 10,
+  "mask_channel_min_space": 1,
+  "mask_channel_other": 0.0,
+  "mask_channel_prob": 0.0,
+  "mask_channel_selection": "static",
+  "mask_feature_length": 10,
+  "mask_feature_prob": 0.0,
+  "mask_time_length": 10,
+  "mask_time_min_space": 1,
+  "mask_time_other": 0.0,
+  "mask_time_prob": 0.05,
+  "mask_time_selection": "static",
+  "model_type": "wav2vec2",
+  "num_attention_heads": 16,
+  "num_codevector_groups": 2,
+  "num_codevectors_per_group": 320,
+  "num_conv_pos_embedding_groups": 16,
+  "num_conv_pos_embeddings": 128,
+  "num_feat_extract_layers": 7,
+  "num_hidden_layers": 24,
+  "num_negatives": 100,
+  "pad_token_id": 28,
+  "proj_codevector_dim": 768,
+  "transformers_version": "4.6.0",
+  "vocab_size": 29
+}

eval_results.json

@@ -0,0 +1,12 @@
+{
+    "epoch": 40.0,
+    "eval_loss": 0.508992612361908,
+    "eval_mem_cpu_alloc_delta": -233472,
+    "eval_mem_cpu_peaked_delta": 233472,
+    "eval_mem_gpu_alloc_delta": 0,
+    "eval_mem_gpu_peaked_delta": 807446528,
+    "eval_runtime": 33.0419,
+    "eval_samples": 405,
+    "eval_samples_per_second": 12.257,
+    "eval_wer": 0.36945244956772333
+}

preprocessor_config.json

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

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5ba282a4a9b784922cb19602a73f163fa9c589e6a192fd2d308397ae0dfea634
+size 1262042545

special_tokens_map.json

@@ -0,0 +1 @@
+{"bos_token": "<s>", "eos_token": "</s>", "unk_token": "[UNK]", "pad_token": "[PAD]"}

tokenizer_config.json

@@ -0,0 +1 @@
+{"unk_token": "[UNK]", "bos_token": "<s>", "eos_token": "</s>", "pad_token": "[PAD]", "do_lower_case": false, "word_delimiter_token": " "}

train_results.json

@@ -0,0 +1,14 @@
+{
+    "epoch": 40.0,
+    "init_mem_cpu_alloc_delta": 0,
+    "init_mem_cpu_peaked_delta": 0,
+    "init_mem_gpu_alloc_delta": 0,
+    "init_mem_gpu_peaked_delta": 0,
+    "train_mem_cpu_alloc_delta": 428699648,
+    "train_mem_cpu_peaked_delta": 12730368,
+    "train_mem_gpu_alloc_delta": 3823077376,
+    "train_mem_gpu_peaked_delta": 1210148352,
+    "train_runtime": 7790.1813,
+    "train_samples": 1618,
+    "train_samples_per_second": 1.037
+}

trainer_state.json

@@ -0,0 +1,247 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 39.99753086419753,
+  "global_step": 8080,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 2.47,
+      "learning_rate": 4.9500000000000004e-05,
+      "loss": 6.895,
+      "step": 500
+    },
+    {
+      "epoch": 2.47,
+      "eval_loss": 3.3130884170532227,
+      "eval_runtime": 33.0838,
+      "eval_samples_per_second": 12.242,
+      "eval_wer": 0.9988472622478386,
+      "step": 500
+    },
+    {
+      "epoch": 4.95,
+      "learning_rate": 9.95e-05,
+      "loss": 3.0254,
+      "step": 1000
+    },
+    {
+      "epoch": 4.95,
+      "eval_loss": 2.1005313396453857,
+      "eval_runtime": 32.9364,
+      "eval_samples_per_second": 12.296,
+      "eval_wer": 0.9988472622478386,
+      "step": 1000
+    },
+    {
+      "epoch": 7.42,
+      "learning_rate": 9.300847457627119e-05,
+      "loss": 1.471,
+      "step": 1500
+    },
+    {
+      "epoch": 7.42,
+      "eval_loss": 0.7423771023750305,
+      "eval_runtime": 32.8882,
+      "eval_samples_per_second": 12.314,
+      "eval_wer": 0.5682997118155619,
+      "step": 1500
+    },
+    {
+      "epoch": 9.9,
+      "learning_rate": 8.594632768361582e-05,
+      "loss": 0.7797,
+      "step": 2000
+    },
+    {
+      "epoch": 9.9,
+      "eval_loss": 0.6373059749603271,
+      "eval_runtime": 32.981,
+      "eval_samples_per_second": 12.28,
+      "eval_wer": 0.4697406340057637,
+      "step": 2000
+    },
+    {
+      "epoch": 12.38,
+      "learning_rate": 7.888418079096046e-05,
+      "loss": 0.5119,
+      "step": 2500
+    },
+    {
+      "epoch": 12.38,
+      "eval_loss": 0.5740929245948792,
+      "eval_runtime": 32.948,
+      "eval_samples_per_second": 12.292,
+      "eval_wer": 0.4489913544668588,
+      "step": 2500
+    },
+    {
+      "epoch": 14.85,
+      "learning_rate": 7.182203389830508e-05,
+      "loss": 0.3907,
+      "step": 3000
+    },
+    {
+      "epoch": 14.85,
+      "eval_loss": 0.5512897968292236,
+      "eval_runtime": 33.0089,
+      "eval_samples_per_second": 12.269,
+      "eval_wer": 0.4265129682997118,
+      "step": 3000
+    },
+    {
+      "epoch": 17.33,
+      "learning_rate": 6.475988700564972e-05,
+      "loss": 0.3139,
+      "step": 3500
+    },
+    {
+      "epoch": 17.33,
+      "eval_loss": 0.5537411570549011,
+      "eval_runtime": 32.9414,
+      "eval_samples_per_second": 12.295,
+      "eval_wer": 0.4092219020172911,
+      "step": 3500
+    },
+    {
+      "epoch": 19.8,
+      "learning_rate": 5.7697740112994355e-05,
+      "loss": 0.2782,
+      "step": 4000
+    },
+    {
+      "epoch": 19.8,
+      "eval_loss": 0.5088971257209778,
+      "eval_runtime": 32.8851,
+      "eval_samples_per_second": 12.316,
+      "eval_wer": 0.3994236311239193,
+      "step": 4000
+    },
+    {
+      "epoch": 22.28,
+      "learning_rate": 5.0635593220338987e-05,
+      "loss": 0.2385,
+      "step": 4500
+    },
+    {
+      "epoch": 22.28,
+      "eval_loss": 0.48669153451919556,
+      "eval_runtime": 32.8076,
+      "eval_samples_per_second": 12.345,
+      "eval_wer": 0.38904899135446686,
+      "step": 4500
+    },
+    {
+      "epoch": 24.75,
+      "learning_rate": 4.357344632768362e-05,
+      "loss": 0.1983,
+      "step": 5000
+    },
+    {
+      "epoch": 24.75,
+      "eval_loss": 0.5359306335449219,
+      "eval_runtime": 32.9284,
+      "eval_samples_per_second": 12.299,
+      "eval_wer": 0.3930835734870317,
+      "step": 5000
+    },
+    {
+      "epoch": 27.23,
+      "learning_rate": 3.651129943502825e-05,
+      "loss": 0.1622,
+      "step": 5500
+    },
+    {
+      "epoch": 27.23,
+      "eval_loss": 0.5525504350662231,
+      "eval_runtime": 32.7893,
+      "eval_samples_per_second": 12.352,
+      "eval_wer": 0.39481268011527376,
+      "step": 5500
+    },
+    {
+      "epoch": 29.7,
+      "learning_rate": 2.9449152542372883e-05,
+      "loss": 0.1617,
+      "step": 6000
+    },
+    {
+      "epoch": 29.7,
+      "eval_loss": 0.4957633316516876,
+      "eval_runtime": 32.7274,
+      "eval_samples_per_second": 12.375,
+      "eval_wer": 0.3792507204610951,
+      "step": 6000
+    },
+    {
+      "epoch": 32.18,
+      "learning_rate": 2.2387005649717515e-05,
+      "loss": 0.1307,
+      "step": 6500
+    },
+    {
+      "epoch": 32.18,
+      "eval_loss": 0.4741367697715759,
+      "eval_runtime": 32.873,
+      "eval_samples_per_second": 12.32,
+      "eval_wer": 0.37521613832853024,
+      "step": 6500
+    },
+    {
+      "epoch": 34.65,
+      "learning_rate": 1.5324858757062146e-05,
+      "loss": 0.1123,
+      "step": 7000
+    },
+    {
+      "epoch": 34.65,
+      "eval_loss": 0.5188319087028503,
+      "eval_runtime": 32.9211,
+      "eval_samples_per_second": 12.302,
+      "eval_wer": 0.37521613832853024,
+      "step": 7000
+    },
+    {
+      "epoch": 37.13,
+      "learning_rate": 8.27683615819209e-06,
+      "loss": 0.1098,
+      "step": 7500
+    },
+    {
+      "epoch": 37.13,
+      "eval_loss": 0.5156564116477966,
+      "eval_runtime": 32.9215,
+      "eval_samples_per_second": 12.302,
+      "eval_wer": 0.37060518731988473,
+      "step": 7500
+    },
+    {
+      "epoch": 39.6,
+      "learning_rate": 1.2146892655367232e-06,
+      "loss": 0.0982,
+      "step": 8000
+    },
+    {
+      "epoch": 39.6,
+      "eval_loss": 0.5081274509429932,
+      "eval_runtime": 32.9436,
+      "eval_samples_per_second": 12.294,
+      "eval_wer": 0.37002881844380403,
+      "step": 8000
+    },
+    {
+      "epoch": 40.0,
+      "step": 8080,
+      "total_flos": 3.820109864655936e+16,
+      "train_runtime": 7790.1813,
+      "train_samples_per_second": 1.037
+    }
+  ],
+  "max_steps": 8080,
+  "num_train_epochs": 40,
+  "total_flos": 3.820109864655936e+16,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c6a8c2af59db205282b449c11f41bbfb261328c9eeb88e6fabcdce47fc64e9e6
+size 2479

vocab.json

@@ -0,0 +1 @@
+{"f": 0, "r": 1, "q": 2, "o": 3, "g": 4, "b": 5, "t": 6, "u": 7, "n": 8, "v": 9, "e": 10, "p": 11, "d": 12, "a": 13, "j": 14, "m": 15, "z": 16, "w": 17, "k": 18, "s": 19, "l": 20, " ": 21, "y": 22, "c": 23, "h": 24, "x": 25, "i": 26, "[UNK]": 27, "[PAD]": 28}