End of training

README.md

@@ -1,5 +1,6 @@
 ---
 license: apache-2.0
+base_model: openai/whisper-tiny
 tags:
 - generated_from_trainer
 datasets:
@@ -7,22 +8,21 @@ datasets:
 metrics:
 - wer
 model-index:
-- name: whisper-tiny-finetuned-minds14
+- name: whisper-small-dv
   results:
   - task:
       name: Automatic Speech Recognition
       type: automatic-speech-recognition
     dataset:
-      name: minds14
+      name: PolyAI/minds14
       type: PolyAI/minds14
       config: en-US
-      split: train[450:]
+      split: train
       args: en-US
     metrics:
     - name: Wer
       type: wer
-      value: 26.711906
-pipeline_tag: automatic-speech-recognition
+      value: 27.390791027154666
 ---
 
 <!-- This model card has been generated automatically according to the information the Trainer had access to. You
@@ -32,13 +32,9 @@ should probably proofread and complete it, then remove this comment. -->
 
 This model is a fine-tuned version of [openai/whisper-tiny](https://huggingface.co/openai/whisper-tiny) on the PolyAI/minds14 dataset.
 It achieves the following results on the evaluation set:
-- eval_loss: 0.5904
-- eval_wer: 26.7119
-- eval_runtime: 61.9232
-- eval_samples_per_second: 1.825
-- eval_steps_per_second: 0.129
-- epoch: 17.24
-- step: 500
+- Loss: 0.5904
+- Wer Ortho: 26.7119
+- Wer: 27.3908
 
 ## Model description
 
@@ -64,12 +60,19 @@ The following hyperparameters were used during training:
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: constant_with_warmup
 - lr_scheduler_warmup_steps: 50
-- training_steps: 2500
+- training_steps: 500
 - mixed_precision_training: Native AMP
 
+### Training results
+
+| Training Loss | Epoch | Step | Validation Loss | Wer Ortho | Wer     |
+|:-------------:|:-----:|:----:|:---------------:|:---------:|:-------:|
+| 0.0002        | 17.24 | 500  | 0.5904          | 26.7119   | 27.3908 |
+
+
 ### Framework versions
 
 - Transformers 4.35.2
 - Pytorch 2.1.0+cu121
 - Datasets 2.16.1
-- Tokenizers 0.15.1
+- Tokenizers 0.15.1

generation_config.json

@@ -51,11 +51,15 @@
   "forced_decoder_ids": [
     [
       1,
-      null
+      50259
     ],
     [
       2,
       50359
+    ],
+    [
+      3,
+      50363
     ]
   ],
   "is_multilingual": true,
@@ -160,6 +164,7 @@
     "<|yo|>": 50325,
     "<|zh|>": 50260
   },
+  "language": "en",
   "max_initial_timestamp_index": 50,
   "max_length": 448,
   "no_timestamps_token_id": 50363,
@@ -256,6 +261,7 @@
     50361,
     50362
   ],
+  "task": "transcribe",
   "task_to_id": {
     "transcribe": 50359,
     "translate": 50358