End of training

README.md

@@ -3,6 +3,9 @@ library_name: transformers
 license: apache-2.0
 base_model: facebook/wav2vec2-base
 tags:
+- automatic-speech-recognition
+- nyagen
+- mms
 - generated_from_trainer
 metrics:
 - wer
@@ -16,9 +19,9 @@ should probably proofread and complete it, then remove this comment. -->
 
 # mms-300m-nyagen-balanced-model
 
-This model is a fine-tuned version of [facebook/wav2vec2-base](https://huggingface.co/facebook/wav2vec2-base) on an unknown dataset.
+This model is a fine-tuned version of [facebook/wav2vec2-base](https://huggingface.co/facebook/wav2vec2-base) on the NYAGEN - NYA dataset.
 It achieves the following results on the evaluation set:
-- Loss: 2.8538
+- Loss: 2.7147
 - Wer: 1.0
 
 ## Model description

adapter.nya.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:729f10b5204e4e30fb5732742122381810a5b4cde638da05f0e212d8bbdcac48
-size 95540
+oid sha256:ee414c648d270b015c1e907e469229e695783b80c06e457790e76cdfa6e51e45
+size 2695340

all_results.json

@@ -1,15 +1,15 @@
 {
     "epoch": 1.0,
-    "eval_loss": 3.070291519165039,
-    "eval_runtime": 11.644,
+    "eval_loss": 2.714689016342163,
+    "eval_runtime": 11.5744,
     "eval_samples": 169,
-    "eval_samples_per_second": 14.514,
-    "eval_steps_per_second": 3.693,
+    "eval_samples_per_second": 14.601,
+    "eval_steps_per_second": 3.715,
     "eval_wer": 1.0,
-    "total_flos": 1.3202067631616544e+17,
-    "train_loss": 8.08548327704161,
-    "train_runtime": 138.4297,
+    "total_flos": 1.3292347177359955e+17,
+    "train_loss": 5.12707603833952,
+    "train_runtime": 148.2598,
     "train_samples": 1445,
-    "train_samples_per_second": 10.439,
-    "train_steps_per_second": 1.308
+    "train_samples_per_second": 9.746,
+    "train_steps_per_second": 1.221
 }

eval_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 1.0,
-    "eval_loss": 3.070291519165039,
-    "eval_runtime": 11.644,
+    "eval_loss": 2.714689016342163,
+    "eval_runtime": 11.5744,
     "eval_samples": 169,
-    "eval_samples_per_second": 14.514,
-    "eval_steps_per_second": 3.693,
+    "eval_samples_per_second": 14.601,
+    "eval_steps_per_second": 3.715,
     "eval_wer": 1.0
 }

train_results.json

@@ -1,9 +1,9 @@
 {
     "epoch": 1.0,
-    "total_flos": 1.3202067631616544e+17,
-    "train_loss": 8.08548327704161,
-    "train_runtime": 138.4297,
+    "total_flos": 1.3292347177359955e+17,
+    "train_loss": 5.12707603833952,
+    "train_runtime": 148.2598,
     "train_samples": 1445,
-    "train_samples_per_second": 10.439,
-    "train_steps_per_second": 1.308
+    "train_samples_per_second": 9.746,
+    "train_steps_per_second": 1.221
 }

trainer_state.json

@@ -1,5 +1,5 @@
 {
-  "best_metric": 5.505459308624268,
+  "best_metric": 2.8538482189178467,
   "best_model_checkpoint": "/scratch/skscla001/speech/results/mms-300m-nyagen-balanced-model/checkpoint-100",
   "epoch": 1.0,
   "eval_steps": 100,
@@ -10,28 +10,28 @@
   "log_history": [
     {
       "epoch": 0.5524861878453039,
-      "grad_norm": 14.355249404907227,
+      "grad_norm": 1.579022765159607,
       "learning_rate": 0.000294,
-      "loss": 11.4413,
+      "loss": 7.0001,
       "step": 100
     },
     {
       "epoch": 0.5524861878453039,
-      "eval_loss": 5.505459308624268,
-      "eval_runtime": 11.2792,
-      "eval_samples_per_second": 14.983,
-      "eval_steps_per_second": 3.812,
+      "eval_loss": 2.8538482189178467,
+      "eval_runtime": 12.0623,
+      "eval_samples_per_second": 14.011,
+      "eval_steps_per_second": 3.565,
       "eval_wer": 1.0,
       "step": 100
     },
     {
       "epoch": 1.0,
       "step": 181,
-      "total_flos": 1.3202067631616544e+17,
-      "train_loss": 8.08548327704161,
-      "train_runtime": 138.4297,
-      "train_samples_per_second": 10.439,
-      "train_steps_per_second": 1.308
+      "total_flos": 1.3292347177359955e+17,
+      "train_loss": 5.12707603833952,
+      "train_runtime": 148.2598,
+      "train_samples_per_second": 9.746,
+      "train_steps_per_second": 1.221
     }
   ],
   "logging_steps": 100,
@@ -51,7 +51,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 1.3202067631616544e+17,
+  "total_flos": 1.3292347177359955e+17,
   "train_batch_size": 4,
   "trial_name": null,
   "trial_params": null