Training in progress, step 3800, checkpoint

checkpoint-3800/config.json

@@ -0,0 +1,109 @@
+{
+  "_name_or_path": "batoula187/wav2vec2-xlsr-arabic2",
+  "activation_dropout": 0.0,
+  "adapter_attn_dim": null,
+  "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": 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": 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,
+  "gradient_checkpointing": false,
+  "hidden_act": "gelu",
+  "hidden_dropout": 0.0,
+  "hidden_size": 1024,
+  "initializer_range": 0.02,
+  "intermediate_size": 4096,
+  "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": 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,
+  "output_hidden_size": 1024,
+  "pad_token_id": 88,
+  "proj_codevector_dim": 768,
+  "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": "4.41.2",
+  "use_weighted_layer_sum": false,
+  "vocab_size": 91,
+  "xvector_output_dim": 512
+}

checkpoint-3800/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b61d9d1e0e4901dd1f827f63e718d8803cc364afbf382961ca7280639bfda1c0
+size 1262180580

checkpoint-3800/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:32394ab43cfc75954352aa889f323062cce7ab9e1fd23049ebf1facb5e616520
+size 2490905910

checkpoint-3800/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
+}

checkpoint-3800/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:deabe82adc6067262996a05bd9fb2349563d7056447221697be37e0ecda65477
+size 14244

checkpoint-3800/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:684c3e2167c5a4e88fb6c84465f9d1b861ad3145efebcf9adf0596790d98dbaa
+size 1064

checkpoint-3800/trainer_state.json

@@ -0,0 +1,337 @@
+{
+  "best_metric": 1.18520188331604,
+  "best_model_checkpoint": "batoula187/wav2vec2-xlsr-arabic2/checkpoint-1800",
+  "epoch": 8.033826638477802,
+  "eval_steps": 200,
+  "global_step": 3800,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.42283298097251587,
+      "grad_norm": 0.43856075406074524,
+      "learning_rate": 0.00011999999999999999,
+      "loss": 0.1174,
+      "step": 200
+    },
+    {
+      "epoch": 0.42283298097251587,
+      "eval_loss": 1.2291227579116821,
+      "eval_runtime": 64.2106,
+      "eval_samples_per_second": 16.321,
+      "eval_steps_per_second": 2.04,
+      "eval_wer": 0.6153393487642213,
+      "step": 200
+    },
+    {
+      "epoch": 0.8456659619450317,
+      "grad_norm": 5.825935363769531,
+      "learning_rate": 0.00023999999999999998,
+      "loss": 0.098,
+      "step": 400
+    },
+    {
+      "epoch": 0.8456659619450317,
+      "eval_loss": 1.232483148574829,
+      "eval_runtime": 64.9125,
+      "eval_samples_per_second": 16.145,
+      "eval_steps_per_second": 2.018,
+      "eval_wer": 0.6275009807767752,
+      "step": 400
+    },
+    {
+      "epoch": 1.2684989429175475,
+      "grad_norm": 0.9385707378387451,
+      "learning_rate": 0.00029780861943024103,
+      "loss": 0.1301,
+      "step": 600
+    },
+    {
+      "epoch": 1.2684989429175475,
+      "eval_loss": 1.1968672275543213,
+      "eval_runtime": 63.9557,
+      "eval_samples_per_second": 16.386,
+      "eval_steps_per_second": 2.048,
+      "eval_wer": 0.6127893291486858,
+      "step": 600
+    },
+    {
+      "epoch": 1.6913319238900635,
+      "grad_norm": 0.5655494332313538,
+      "learning_rate": 0.00029346968590211834,
+      "loss": 0.1514,
+      "step": 800
+    },
+    {
+      "epoch": 1.6913319238900635,
+      "eval_loss": 1.2292981147766113,
+      "eval_runtime": 64.6544,
+      "eval_samples_per_second": 16.209,
+      "eval_steps_per_second": 2.026,
+      "eval_wer": 0.6488819144762652,
+      "step": 800
+    },
+    {
+      "epoch": 2.1141649048625792,
+      "grad_norm": 1.0750213861465454,
+      "learning_rate": 0.0002890869247626004,
+      "loss": 0.1494,
+      "step": 1000
+    },
+    {
+      "epoch": 2.1141649048625792,
+      "eval_loss": 1.3062251806259155,
+      "eval_runtime": 63.5313,
+      "eval_samples_per_second": 16.496,
+      "eval_steps_per_second": 2.062,
+      "eval_wer": 0.670066692820714,
+      "step": 1000
+    },
+    {
+      "epoch": 2.536997885835095,
+      "grad_norm": 0.8127875328063965,
+      "learning_rate": 0.0002847041636230825,
+      "loss": 0.1382,
+      "step": 1200
+    },
+    {
+      "epoch": 2.536997885835095,
+      "eval_loss": 1.2222816944122314,
+      "eval_runtime": 63.8482,
+      "eval_samples_per_second": 16.414,
+      "eval_steps_per_second": 2.052,
+      "eval_wer": 0.6261278932914869,
+      "step": 1200
+    },
+    {
+      "epoch": 2.9598308668076108,
+      "grad_norm": 0.8811041116714478,
+      "learning_rate": 0.00028032140248356464,
+      "loss": 0.1382,
+      "step": 1400
+    },
+    {
+      "epoch": 2.9598308668076108,
+      "eval_loss": 1.3115968704223633,
+      "eval_runtime": 65.0526,
+      "eval_samples_per_second": 16.11,
+      "eval_steps_per_second": 2.014,
+      "eval_wer": 0.6506473126716359,
+      "step": 1400
+    },
+    {
+      "epoch": 3.382663847780127,
+      "grad_norm": 0.7833051085472107,
+      "learning_rate": 0.0002759386413440467,
+      "loss": 0.1239,
+      "step": 1600
+    },
+    {
+      "epoch": 3.382663847780127,
+      "eval_loss": 1.1977170705795288,
+      "eval_runtime": 64.8745,
+      "eval_samples_per_second": 16.154,
+      "eval_steps_per_second": 2.019,
+      "eval_wer": 0.6188701451549627,
+      "step": 1600
+    },
+    {
+      "epoch": 3.8054968287526427,
+      "grad_norm": 0.4267895817756653,
+      "learning_rate": 0.0002715558802045288,
+      "loss": 0.1228,
+      "step": 1800
+    },
+    {
+      "epoch": 3.8054968287526427,
+      "eval_loss": 1.18520188331604,
+      "eval_runtime": 64.2457,
+      "eval_samples_per_second": 16.312,
+      "eval_steps_per_second": 2.039,
+      "eval_wer": 0.6280894468418988,
+      "step": 1800
+    },
+    {
+      "epoch": 4.2283298097251585,
+      "grad_norm": 0.7231135964393616,
+      "learning_rate": 0.00026717311906501094,
+      "loss": 0.1117,
+      "step": 2000
+    },
+    {
+      "epoch": 4.2283298097251585,
+      "eval_loss": 1.3370016813278198,
+      "eval_runtime": 64.7812,
+      "eval_samples_per_second": 16.178,
+      "eval_steps_per_second": 2.022,
+      "eval_wer": 0.6494703805413887,
+      "step": 2000
+    },
+    {
+      "epoch": 4.651162790697675,
+      "grad_norm": 0.40635955333709717,
+      "learning_rate": 0.000262790357925493,
+      "loss": 0.1118,
+      "step": 2200
+    },
+    {
+      "epoch": 4.651162790697675,
+      "eval_loss": 1.3265222311019897,
+      "eval_runtime": 63.5949,
+      "eval_samples_per_second": 16.479,
+      "eval_steps_per_second": 2.06,
+      "eval_wer": 0.6431934091800706,
+      "step": 2200
+    },
+    {
+      "epoch": 5.07399577167019,
+      "grad_norm": 0.9417168498039246,
+      "learning_rate": 0.0002584075967859751,
+      "loss": 0.1101,
+      "step": 2400
+    },
+    {
+      "epoch": 5.07399577167019,
+      "eval_loss": 1.345849633216858,
+      "eval_runtime": 64.4165,
+      "eval_samples_per_second": 16.269,
+      "eval_steps_per_second": 2.034,
+      "eval_wer": 0.6310317771675167,
+      "step": 2400
+    },
+    {
+      "epoch": 5.496828752642706,
+      "grad_norm": 0.7866037487983704,
+      "learning_rate": 0.00025402483564645724,
+      "loss": 0.1328,
+      "step": 2600
+    },
+    {
+      "epoch": 5.496828752642706,
+      "eval_loss": 1.2545326948165894,
+      "eval_runtime": 64.0398,
+      "eval_samples_per_second": 16.365,
+      "eval_steps_per_second": 2.046,
+      "eval_wer": 0.6341702628481758,
+      "step": 2600
+    },
+    {
+      "epoch": 5.9196617336152215,
+      "grad_norm": 0.47020798921585083,
+      "learning_rate": 0.00024964207450693936,
+      "loss": 0.1384,
+      "step": 2800
+    },
+    {
+      "epoch": 5.9196617336152215,
+      "eval_loss": 1.2806007862091064,
+      "eval_runtime": 64.3065,
+      "eval_samples_per_second": 16.297,
+      "eval_steps_per_second": 2.037,
+      "eval_wer": 0.6265202040015693,
+      "step": 2800
+    },
+    {
+      "epoch": 6.342494714587738,
+      "grad_norm": 0.605645477771759,
+      "learning_rate": 0.0002452593133674214,
+      "loss": 0.1334,
+      "step": 3000
+    },
+    {
+      "epoch": 6.342494714587738,
+      "eval_loss": 1.2484089136123657,
+      "eval_runtime": 65.2815,
+      "eval_samples_per_second": 16.054,
+      "eval_steps_per_second": 2.007,
+      "eval_wer": 0.6369164378187524,
+      "step": 3000
+    },
+    {
+      "epoch": 6.765327695560254,
+      "grad_norm": 0.6708455085754395,
+      "learning_rate": 0.00024087655222790357,
+      "loss": 0.1383,
+      "step": 3200
+    },
+    {
+      "epoch": 6.765327695560254,
+      "eval_loss": 1.2701318264007568,
+      "eval_runtime": 63.6121,
+      "eval_samples_per_second": 16.475,
+      "eval_steps_per_second": 2.059,
+      "eval_wer": 0.6479011377010593,
+      "step": 3200
+    },
+    {
+      "epoch": 7.188160676532769,
+      "grad_norm": 0.3204992711544037,
+      "learning_rate": 0.00023649379108838566,
+      "loss": 0.1281,
+      "step": 3400
+    },
+    {
+      "epoch": 7.188160676532769,
+      "eval_loss": 1.1926395893096924,
+      "eval_runtime": 65.1953,
+      "eval_samples_per_second": 16.075,
+      "eval_steps_per_second": 2.009,
+      "eval_wer": 0.6314240878775991,
+      "step": 3400
+    },
+    {
+      "epoch": 7.6109936575052854,
+      "grad_norm": 2.2028682231903076,
+      "learning_rate": 0.00023211102994886775,
+      "loss": 0.1232,
+      "step": 3600
+    },
+    {
+      "epoch": 7.6109936575052854,
+      "eval_loss": 1.2255371809005737,
+      "eval_runtime": 65.3961,
+      "eval_samples_per_second": 16.025,
+      "eval_steps_per_second": 2.003,
+      "eval_wer": 0.6186739897999215,
+      "step": 3600
+    },
+    {
+      "epoch": 8.033826638477802,
+      "grad_norm": 0.33132538199424744,
+      "learning_rate": 2e-05,
+      "loss": 0.0727,
+      "step": 3800
+    },
+    {
+      "epoch": 8.033826638477802,
+      "eval_loss": 1.23982572555542,
+      "eval_runtime": 66.6836,
+      "eval_samples_per_second": 15.716,
+      "eval_steps_per_second": 1.575,
+      "eval_wer": 0.6014123185562966,
+      "step": 3800
+    }
+  ],
+  "logging_steps": 200,
+  "max_steps": 11825,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 25,
+  "save_steps": 200,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1.2164480425230754e+19,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-3800/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a1cdb4afb19ad76c1497ce9c323bae80a0da0c233e1e1189c8502c0632a60f4e
+size 5112