End of training

README.md

@@ -3,6 +3,9 @@ library_name: transformers
 license: cc-by-nc-4.0
 base_model: facebook/mms-1b-all
 tags:
+- automatic-speech-recognition
+- lozgen
+- mms
 - generated_from_trainer
 metrics:
 - wer
@@ -16,10 +19,10 @@ should probably proofread and complete it, then remove this comment. -->
 
 # mms-1b-lozgen-female-model
 
-This model is a fine-tuned version of [facebook/mms-1b-all](https://huggingface.co/facebook/mms-1b-all) on an unknown dataset.
+This model is a fine-tuned version of [facebook/mms-1b-all](https://huggingface.co/facebook/mms-1b-all) on the LOZGEN - LOZ dataset.
 It achieves the following results on the evaluation set:
 - Loss: 0.6218
-- Wer: 0.3962
+- Wer: 0.3955
 
 ## Model description
 

adapter.loz.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7cf3b5bf926d6637226f52d9ef3d2c50c6c6558133127ebb564c1f56b4bd8423
+size 8793408

all_results.json

@@ -0,0 +1,15 @@
+{
+    "epoch": 15.652173913043478,
+    "eval_loss": 0.621826171875,
+    "eval_runtime": 19.2044,
+    "eval_samples": 249,
+    "eval_samples_per_second": 12.966,
+    "eval_steps_per_second": 3.281,
+    "eval_wer": 0.3954524144397562,
+    "total_flos": 5.624155339355839e+18,
+    "train_loss": 1.1473070081075032,
+    "train_runtime": 1548.2979,
+    "train_samples": 458,
+    "train_samples_per_second": 8.874,
+    "train_steps_per_second": 2.228
+}

eval_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 15.652173913043478,
+    "eval_loss": 0.621826171875,
+    "eval_runtime": 19.2044,
+    "eval_samples": 249,
+    "eval_samples_per_second": 12.966,
+    "eval_steps_per_second": 3.281,
+    "eval_wer": 0.3954524144397562
+}

train_results.json

@@ -0,0 +1,9 @@
+{
+    "epoch": 15.652173913043478,
+    "total_flos": 5.624155339355839e+18,
+    "train_loss": 1.1473070081075032,
+    "train_runtime": 1548.2979,
+    "train_samples": 458,
+    "train_samples_per_second": 8.874,
+    "train_steps_per_second": 2.228
+}

trainer_state.json

@@ -0,0 +1,339 @@
+{
+  "best_metric": 0.6179416179656982,
+  "best_model_checkpoint": "/scratch/skscla001/speech/results/mms-1b-lozgen-female-model/checkpoint-1500",
+  "epoch": 15.652173913043478,
+  "eval_steps": 100,
+  "global_step": 1800,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.8695652173913043,
+      "grad_norm": 3.3569300174713135,
+      "learning_rate": 0.000285,
+      "loss": 6.7809,
+      "step": 100
+    },
+    {
+      "epoch": 0.8695652173913043,
+      "eval_loss": 3.1491334438323975,
+      "eval_runtime": 19.3955,
+      "eval_samples_per_second": 12.838,
+      "eval_steps_per_second": 3.248,
+      "eval_wer": 0.9758556024378809,
+      "step": 100
+    },
+    {
+      "epoch": 1.7391304347826086,
+      "grad_norm": 3.8020479679107666,
+      "learning_rate": 0.0002914925373134328,
+      "loss": 2.6957,
+      "step": 200
+    },
+    {
+      "epoch": 1.7391304347826086,
+      "eval_loss": 2.001054286956787,
+      "eval_runtime": 19.3287,
+      "eval_samples_per_second": 12.882,
+      "eval_steps_per_second": 3.259,
+      "eval_wer": 0.9278012189404594,
+      "step": 200
+    },
+    {
+      "epoch": 2.608695652173913,
+      "grad_norm": 1.5241467952728271,
+      "learning_rate": 0.0002825373134328358,
+      "loss": 1.1829,
+      "step": 300
+    },
+    {
+      "epoch": 2.608695652173913,
+      "eval_loss": 0.806449294090271,
+      "eval_runtime": 19.2752,
+      "eval_samples_per_second": 12.918,
+      "eval_steps_per_second": 3.268,
+      "eval_wer": 0.6420534458509142,
+      "step": 300
+    },
+    {
+      "epoch": 3.4782608695652173,
+      "grad_norm": 1.655979871749878,
+      "learning_rate": 0.0002735820895522388,
+      "loss": 0.8527,
+      "step": 400
+    },
+    {
+      "epoch": 3.4782608695652173,
+      "eval_loss": 0.7470424175262451,
+      "eval_runtime": 19.1645,
+      "eval_samples_per_second": 12.993,
+      "eval_steps_per_second": 3.287,
+      "eval_wer": 0.53328645100797,
+      "step": 400
+    },
+    {
+      "epoch": 4.3478260869565215,
+      "grad_norm": 5.353499412536621,
+      "learning_rate": 0.00026462686567164175,
+      "loss": 0.8376,
+      "step": 500
+    },
+    {
+      "epoch": 4.3478260869565215,
+      "eval_loss": 0.69962078332901,
+      "eval_runtime": 19.293,
+      "eval_samples_per_second": 12.906,
+      "eval_steps_per_second": 3.265,
+      "eval_wer": 0.49226441631504925,
+      "step": 500
+    },
+    {
+      "epoch": 5.217391304347826,
+      "grad_norm": 2.5344457626342773,
+      "learning_rate": 0.00025567164179104475,
+      "loss": 0.7567,
+      "step": 600
+    },
+    {
+      "epoch": 5.217391304347826,
+      "eval_loss": 0.6791635751724243,
+      "eval_runtime": 19.2801,
+      "eval_samples_per_second": 12.915,
+      "eval_steps_per_second": 3.268,
+      "eval_wer": 0.46858884200656353,
+      "step": 600
+    },
+    {
+      "epoch": 6.086956521739131,
+      "grad_norm": 0.9048321843147278,
+      "learning_rate": 0.00024671641791044774,
+      "loss": 0.7014,
+      "step": 700
+    },
+    {
+      "epoch": 6.086956521739131,
+      "eval_loss": 0.6882011890411377,
+      "eval_runtime": 19.2193,
+      "eval_samples_per_second": 12.956,
+      "eval_steps_per_second": 3.278,
+      "eval_wer": 0.45546179090482886,
+      "step": 700
+    },
+    {
+      "epoch": 6.956521739130435,
+      "grad_norm": 2.4095747470855713,
+      "learning_rate": 0.00023776119402985074,
+      "loss": 0.7242,
+      "step": 800
+    },
+    {
+      "epoch": 6.956521739130435,
+      "eval_loss": 0.6556691527366638,
+      "eval_runtime": 19.2162,
+      "eval_samples_per_second": 12.958,
+      "eval_steps_per_second": 3.278,
+      "eval_wer": 0.4578059071729958,
+      "step": 800
+    },
+    {
+      "epoch": 7.826086956521739,
+      "grad_norm": 1.2900141477584839,
+      "learning_rate": 0.0002288059701492537,
+      "loss": 0.6895,
+      "step": 900
+    },
+    {
+      "epoch": 7.826086956521739,
+      "eval_loss": 0.6531555652618408,
+      "eval_runtime": 19.3114,
+      "eval_samples_per_second": 12.894,
+      "eval_steps_per_second": 3.262,
+      "eval_wer": 0.4409282700421941,
+      "step": 900
+    },
+    {
+      "epoch": 8.695652173913043,
+      "grad_norm": 1.196593999862671,
+      "learning_rate": 0.00021985074626865668,
+      "loss": 0.6543,
+      "step": 1000
+    },
+    {
+      "epoch": 8.695652173913043,
+      "eval_loss": 0.6477147340774536,
+      "eval_runtime": 19.299,
+      "eval_samples_per_second": 12.902,
+      "eval_steps_per_second": 3.264,
+      "eval_wer": 0.42780121894045947,
+      "step": 1000
+    },
+    {
+      "epoch": 9.565217391304348,
+      "grad_norm": 5.2444281578063965,
+      "learning_rate": 0.00021089552238805968,
+      "loss": 0.6386,
+      "step": 1100
+    },
+    {
+      "epoch": 9.565217391304348,
+      "eval_loss": 0.6591968536376953,
+      "eval_runtime": 19.2958,
+      "eval_samples_per_second": 12.904,
+      "eval_steps_per_second": 3.265,
+      "eval_wer": 0.4219409282700422,
+      "step": 1100
+    },
+    {
+      "epoch": 10.434782608695652,
+      "grad_norm": 1.0960686206817627,
+      "learning_rate": 0.00020194029850746268,
+      "loss": 0.6145,
+      "step": 1200
+    },
+    {
+      "epoch": 10.434782608695652,
+      "eval_loss": 0.6434625387191772,
+      "eval_runtime": 19.1713,
+      "eval_samples_per_second": 12.988,
+      "eval_steps_per_second": 3.286,
+      "eval_wer": 0.42569151429910923,
+      "step": 1200
+    },
+    {
+      "epoch": 11.304347826086957,
+      "grad_norm": 1.0903316736221313,
+      "learning_rate": 0.00019298507462686568,
+      "loss": 0.6513,
+      "step": 1300
+    },
+    {
+      "epoch": 11.304347826086957,
+      "eval_loss": 0.6570419073104858,
+      "eval_runtime": 19.3251,
+      "eval_samples_per_second": 12.885,
+      "eval_steps_per_second": 3.26,
+      "eval_wer": 0.4146741678387248,
+      "step": 1300
+    },
+    {
+      "epoch": 12.173913043478262,
+      "grad_norm": 1.123496413230896,
+      "learning_rate": 0.00018402985074626862,
+      "loss": 0.5788,
+      "step": 1400
+    },
+    {
+      "epoch": 12.173913043478262,
+      "eval_loss": 0.6258613467216492,
+      "eval_runtime": 19.2989,
+      "eval_samples_per_second": 12.902,
+      "eval_steps_per_second": 3.264,
+      "eval_wer": 0.4085794655414909,
+      "step": 1400
+    },
+    {
+      "epoch": 13.043478260869565,
+      "grad_norm": 0.7355613708496094,
+      "learning_rate": 0.00017507462686567162,
+      "loss": 0.6061,
+      "step": 1500
+    },
+    {
+      "epoch": 13.043478260869565,
+      "eval_loss": 0.6179416179656982,
+      "eval_runtime": 19.2693,
+      "eval_samples_per_second": 12.922,
+      "eval_steps_per_second": 3.269,
+      "eval_wer": 0.4006094702297234,
+      "step": 1500
+    },
+    {
+      "epoch": 13.91304347826087,
+      "grad_norm": 2.3467090129852295,
+      "learning_rate": 0.00016611940298507462,
+      "loss": 0.5647,
+      "step": 1600
+    },
+    {
+      "epoch": 13.91304347826087,
+      "eval_loss": 0.6186335682868958,
+      "eval_runtime": 19.1768,
+      "eval_samples_per_second": 12.984,
+      "eval_steps_per_second": 3.285,
+      "eval_wer": 0.4006094702297234,
+      "step": 1600
+    },
+    {
+      "epoch": 14.782608695652174,
+      "grad_norm": 1.0731009244918823,
+      "learning_rate": 0.00015716417910447762,
+      "loss": 0.5715,
+      "step": 1700
+    },
+    {
+      "epoch": 14.782608695652174,
+      "eval_loss": 0.627108633518219,
+      "eval_runtime": 19.405,
+      "eval_samples_per_second": 12.832,
+      "eval_steps_per_second": 3.247,
+      "eval_wer": 0.3984997655883732,
+      "step": 1700
+    },
+    {
+      "epoch": 15.652173913043478,
+      "grad_norm": 0.8772910237312317,
+      "learning_rate": 0.0001482089552238806,
+      "loss": 0.5502,
+      "step": 1800
+    },
+    {
+      "epoch": 15.652173913043478,
+      "eval_loss": 0.6218141317367554,
+      "eval_runtime": 19.3309,
+      "eval_samples_per_second": 12.881,
+      "eval_steps_per_second": 3.259,
+      "eval_wer": 0.3961556493202063,
+      "step": 1800
+    },
+    {
+      "epoch": 15.652173913043478,
+      "step": 1800,
+      "total_flos": 5.624155339355839e+18,
+      "train_loss": 1.1473070081075032,
+      "train_runtime": 1548.2979,
+      "train_samples_per_second": 8.874,
+      "train_steps_per_second": 2.228
+    }
+  ],
+  "logging_steps": 100,
+  "max_steps": 3450,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 30,
+  "save_steps": 400,
+  "stateful_callbacks": {
+    "EarlyStoppingCallback": {
+      "args": {
+        "early_stopping_patience": 3,
+        "early_stopping_threshold": 0.0
+      },
+      "attributes": {
+        "early_stopping_patience_counter": 1
+      }
+    },
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 5.624155339355839e+18,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}