diff --git a/README.md b/README.md new file mode 100644 index 0000000000000000000000000000000000000000..3ffb4736d69e8224479166feda6245a6c3750cbb --- /dev/null +++ b/README.md @@ -0,0 +1,74 @@ +--- +base_model: openai/whisper-small +datasets: +- mozilla-foundation/common_voice_11_0 +language: +- hi +library_name: peft +license: apache-2.0 +tags: +- generated_from_trainer +model-index: +- name: Whisper Small Diny + results: [] +--- + + + +# Whisper Small Diny + +This model is a fine-tuned version of [openai/whisper-small](https://huggingface.co/openai/whisper-small) on the Common Voice 11.0 dataset. +It achieves the following results on the evaluation set: +- Loss: 1.3108 + +## Model description + +More information needed + +## Intended uses & limitations + +More information needed + +## Training and evaluation data + +More information needed + +## Training procedure + +### Training hyperparameters + +The following hyperparameters were used during training: +- learning_rate: 0.0001 +- train_batch_size: 8 +- eval_batch_size: 8 +- seed: 42 +- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 +- lr_scheduler_type: linear +- lr_scheduler_warmup_steps: 50 +- num_epochs: 10 +- mixed_precision_training: Native AMP + +### Training results + +| Training Loss | Epoch | Step | Validation Loss | +|:-------------:|:-----:|:----:|:---------------:| +| 3.5747 | 1.0 | 4 | 3.8772 | +| 3.5154 | 2.0 | 8 | 3.7508 | +| 3.3537 | 3.0 | 12 | 3.5159 | +| 3.1336 | 4.0 | 16 | 3.2003 | +| 2.9481 | 5.0 | 20 | 2.8626 | +| 2.5179 | 6.0 | 24 | 2.5561 | +| 2.3247 | 7.0 | 28 | 2.2656 | +| 2.1071 | 8.0 | 32 | 1.9826 | +| 1.7406 | 9.0 | 36 | 1.6799 | +| 1.5765 | 10.0 | 40 | 1.3108 | + + +### Framework versions + +- PEFT 0.12.0 +- Transformers 4.42.4 +- Pytorch 2.3.1+cu121 +- Datasets 2.20.0 +- Tokenizers 0.19.1 \ No newline at end of file diff --git a/adapter_config.json b/adapter_config.json new file mode 100644 index 0000000000000000000000000000000000000000..4f999b279fa898e3d177198482a62b7dc315ff97 --- /dev/null +++ b/adapter_config.json @@ -0,0 +1,32 @@ +{ + "alpha_pattern": {}, + "auto_mapping": { + "base_model_class": "WhisperForConditionalGeneration", + "parent_library": "transformers.models.whisper.modeling_whisper" + }, + "base_model_name_or_path": "openai/whisper-small", + "bias": "none", + "fan_in_fan_out": false, + "inference_mode": true, + "init_lora_weights": true, + "layer_replication": null, + "layers_pattern": null, + "layers_to_transform": null, + "loftq_config": {}, + "lora_alpha": 64, + "lora_dropout": 0.05, + "megatron_config": null, + "megatron_core": "megatron.core", + "modules_to_save": null, + "peft_type": "LORA", + "r": 32, + "rank_pattern": {}, + "revision": null, + "target_modules": [ + "v_proj", + "q_proj" + ], + "task_type": null, + "use_dora": false, + "use_rslora": false +} \ No newline at end of file diff --git a/adapter_model.safetensors b/adapter_model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..fd44f65c9f1a72e7c990459d27bc06edbb5f45e3 --- /dev/null +++ b/adapter_model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:829fac29b6a03dcfa42cf9b96f1bb7faab8a608e50d70a593b042b6a306ec782 +size 14176064 diff --git a/audio/1.mp3 b/audio/1.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..eb8f988c95a23194739363d3274f2509a70f3e37 Binary files /dev/null and b/audio/1.mp3 differ diff --git a/audio/10.mp3 b/audio/10.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..5b66723f5db2c9d5a242011f191a017889f97d46 Binary files /dev/null and b/audio/10.mp3 differ diff --git a/audio/11.mp3 b/audio/11.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..eb0cdd4742baf8460de36d092d7cdbfbaf6cf397 Binary files /dev/null and b/audio/11.mp3 differ diff --git a/audio/12.mp3 b/audio/12.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..90731fdfc6b5031c8db49d6d5a5c29df7bd940f8 Binary files /dev/null and b/audio/12.mp3 differ diff --git a/audio/13.mp3 b/audio/13.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..aa19e7ae14b327a0b21f1cde64722149322495aa Binary files /dev/null and b/audio/13.mp3 differ diff --git a/audio/14.mp3 b/audio/14.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..7bb4facc36b7839ebe1368b90a925b7686ccb81b Binary files /dev/null and b/audio/14.mp3 differ diff --git a/audio/15.mp3 b/audio/15.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..dab9a681c45156d2959abf6d9aaf0a8ec091c678 Binary files /dev/null and b/audio/15.mp3 differ diff --git a/audio/16.mp3 b/audio/16.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..8da5f6d4c7660d03fed307864a4073b2e9e3070c Binary files /dev/null and b/audio/16.mp3 differ diff --git a/audio/17.mp3 b/audio/17.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..af7ad32ea73307f4011788001fe75a096469e453 Binary files /dev/null and b/audio/17.mp3 differ diff --git a/audio/18.mp3 b/audio/18.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..fd871211e0b2ae1a87f6ee5ce81b60280d9cde2d Binary files /dev/null and b/audio/18.mp3 differ diff --git a/audio/19.mp3 b/audio/19.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..f937093687d59ab19f8b4c18d72ad996d4172b64 Binary files /dev/null and b/audio/19.mp3 differ diff --git a/audio/2.mp3 b/audio/2.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..523e628a87976499956fffdaa29fb4e4a14bebcb Binary files /dev/null and b/audio/2.mp3 differ diff --git a/audio/20.mp3 b/audio/20.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..df1d8477ea6834ccd4a99b0045d80341a90f68c6 Binary files /dev/null and b/audio/20.mp3 differ diff --git a/audio/21.mp3 b/audio/21.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..f7fa90fc66e42c58dd78bc6eac145849f08fe180 Binary files /dev/null and b/audio/21.mp3 differ diff --git a/audio/22.mp3 b/audio/22.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..754d82853501bc48f1ef7564ef1e1dd5cb0211f8 Binary files /dev/null and b/audio/22.mp3 differ diff --git a/audio/23.mp3 b/audio/23.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..b6ee741df7abb2b95652ab93b594a5db8de8a3ec Binary files /dev/null and b/audio/23.mp3 differ diff --git a/audio/24.mp3 b/audio/24.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..2ac7322f131b7c12db023ab09d2526384ad6556c Binary files /dev/null and b/audio/24.mp3 differ diff --git a/audio/25.mp3 b/audio/25.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..6072a95836c50d768c73288869f6e2083c6e83c1 Binary files /dev/null and b/audio/25.mp3 differ diff --git a/audio/26.mp3 b/audio/26.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..02dc772f4cd166b1edfe52039a31608b915b1032 Binary files /dev/null and b/audio/26.mp3 differ diff --git a/audio/27.mp3 b/audio/27.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..2a02b798d1a0e64b68de66c253b7db6fe7dcf03b Binary files /dev/null and b/audio/27.mp3 differ diff --git a/audio/28.mp3 b/audio/28.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..4337f343fdd0a7d153417b728640726d61a60541 Binary files /dev/null and b/audio/28.mp3 differ diff --git a/audio/29.mp3 b/audio/29.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..f9621867ee5e12587f131dee6fa1c3549a6c7c89 Binary files /dev/null and b/audio/29.mp3 differ diff --git a/audio/3.mp3 b/audio/3.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..d6707083f2a6fb2af6b41450cd8c5756be79a13d Binary files /dev/null and b/audio/3.mp3 differ diff --git a/audio/30.mp3 b/audio/30.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..aa16df5b48d7e915b6f1b180e6ece2640fa79e52 Binary files /dev/null and b/audio/30.mp3 differ diff --git a/audio/31.mp3 b/audio/31.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..c9acdd5b16431f2e10fb96512026034fa4611850 Binary files /dev/null and b/audio/31.mp3 differ diff --git a/audio/32.mp3 b/audio/32.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..11dc2cbd18b0f64789d51f4ccaf8e90c0e00f131 Binary files /dev/null and b/audio/32.mp3 differ diff --git a/audio/33.mp3 b/audio/33.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..23a85dc22193664311f38ada3fa4958cd281637d Binary files /dev/null and b/audio/33.mp3 differ diff --git a/audio/34.mp3 b/audio/34.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..2dacd50dd77e222233633a6e1c58d161f365ef1a Binary files /dev/null and b/audio/34.mp3 differ diff --git a/audio/35.mp3 b/audio/35.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..aa3f07febec2e9157303c60addb5aabbb3302b55 Binary files /dev/null and b/audio/35.mp3 differ diff --git a/audio/36.mp3 b/audio/36.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..6fea81ca55668ae94be2e633f333eb893b5329ae Binary files /dev/null and b/audio/36.mp3 differ diff --git a/audio/37.mp3 b/audio/37.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..45de31c59dd863499c57539df12b26c877ba7ee0 Binary files /dev/null and b/audio/37.mp3 differ diff --git a/audio/38.mp3 b/audio/38.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..23d569317d22c56838cac6fd72363fd2461906d8 Binary files /dev/null and b/audio/38.mp3 differ diff --git a/audio/39.mp3 b/audio/39.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..8b826ab6eec332d70bffc234d46885fd1e21f21e Binary files /dev/null and b/audio/39.mp3 differ diff --git a/audio/4.mp3 b/audio/4.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..eefa3356c2fc895ce02b15ec48e549a3f6d8403c Binary files /dev/null and b/audio/4.mp3 differ diff --git a/audio/40.mp3 b/audio/40.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..201057e62ff213972e079798f57787c179337f3a Binary files /dev/null and b/audio/40.mp3 differ diff --git a/audio/41.mp3 b/audio/41.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..be5c40a0d801441c96da88d3b82cda7ce1711d56 Binary files /dev/null and b/audio/41.mp3 differ diff --git a/audio/42.mp3 b/audio/42.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..7adafabcc291adfea9bf8e21e4453bcfa38a7e1f Binary files /dev/null and b/audio/42.mp3 differ diff --git a/audio/43.mp3 b/audio/43.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..9e5b7ffcb22384dbc42fe290e7ce3811d520c9af Binary files /dev/null and b/audio/43.mp3 differ diff --git a/audio/44.mp3 b/audio/44.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..451cb42b60cc9a8be93f91c201727b9dfd64481d Binary files /dev/null and b/audio/44.mp3 differ diff --git a/audio/45.mp3 b/audio/45.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..da817c1ec927b9b3e27b7038f33c558bc37ccb40 Binary files /dev/null and b/audio/45.mp3 differ diff --git a/audio/46.mp3 b/audio/46.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..68c5795ad7fc4d5333b5cabb404b0e8dec3647ff Binary files /dev/null and b/audio/46.mp3 differ diff --git a/audio/47.mp3 b/audio/47.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..3058fe464f01ad65a1d84eebed42208ea665eeec Binary files /dev/null and b/audio/47.mp3 differ diff --git a/audio/48.mp3 b/audio/48.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..031daa90fd6040685e181d4725168ed95b12b1ee Binary files /dev/null and b/audio/48.mp3 differ diff --git a/audio/49.mp3 b/audio/49.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..6df256135a593f9650bc0d9bfcb7cac966fa1dcb Binary files /dev/null and b/audio/49.mp3 differ diff --git a/audio/5.mp3 b/audio/5.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..1638565ab2f1e213a464bcc7d86fca860d6ebc8f Binary files /dev/null and b/audio/5.mp3 differ diff --git a/audio/50.mp3 b/audio/50.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..fa42a86849cb920c7eb0e2aa24ab1bf4ecdfa48e Binary files /dev/null and b/audio/50.mp3 differ diff --git a/audio/6.mp3 b/audio/6.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..e4e9db2a0f2947b40682c74ea395d1d53bd85d55 Binary files /dev/null and b/audio/6.mp3 differ diff --git a/audio/7.mp3 b/audio/7.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..49fb96dc57507a46f9e191438fd78539670bfee2 Binary files /dev/null and b/audio/7.mp3 differ diff --git a/audio/8.mp3 b/audio/8.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..0fcb12a4304bc04c5f9b5b9fb600b45b3a3fb4c9 Binary files /dev/null and b/audio/8.mp3 differ diff --git a/audio/9.mp3 b/audio/9.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..b56f044c2a836f933a4d81664b13bedbb8ec94d0 Binary files /dev/null and b/audio/9.mp3 differ diff --git a/converted_audio/1.mp3 b/converted_audio/1.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..be02c867910483da3ef2872618edc4529218ae23 Binary files /dev/null and b/converted_audio/1.mp3 differ diff --git a/converted_audio/2.mp3 b/converted_audio/2.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..e42d81550d809787f27af1c925d52493b33f2779 Binary files /dev/null and b/converted_audio/2.mp3 differ diff --git a/converted_audio/3.mp3 b/converted_audio/3.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..c6645fc3d332d1c4f7cb3a3cf94d0544efd74858 Binary files /dev/null and b/converted_audio/3.mp3 differ diff --git a/converted_audio/4.mp3 b/converted_audio/4.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..22dbb2e788e0627d6f9ab1ccf2780178b9f2668b Binary files /dev/null and b/converted_audio/4.mp3 differ diff --git a/converted_audio/5.mp3 b/converted_audio/5.mp3 new file mode 100644 index 0000000000000000000000000000000000000000..0b4d9fc96c9937bccacb03aa315bb648d3c3de61 Binary files /dev/null and b/converted_audio/5.mp3 differ diff --git a/preprocessor_config.json b/preprocessor_config.json new file mode 100644 index 0000000000000000000000000000000000000000..91876762a536a746d268353c5cba57286e76b058 --- /dev/null +++ b/preprocessor_config.json @@ -0,0 +1,14 @@ +{ + "chunk_length": 30, + "feature_extractor_type": "WhisperFeatureExtractor", + "feature_size": 80, + "hop_length": 160, + "n_fft": 400, + "n_samples": 480000, + "nb_max_frames": 3000, + "padding_side": "right", + "padding_value": 0.0, + "processor_class": "WhisperProcessor", + "return_attention_mask": false, + "sampling_rate": 16000 +} diff --git a/runs/Aug12_12-38-38_3e8ad4f624c2/events.out.tfevents.1723466376.3e8ad4f624c2.653.0 b/runs/Aug12_12-38-38_3e8ad4f624c2/events.out.tfevents.1723466376.3e8ad4f624c2.653.0 new file mode 100644 index 0000000000000000000000000000000000000000..fd3b76655fdb23e4a6cbfda0d59e6f346f5afdbf --- /dev/null +++ b/runs/Aug12_12-38-38_3e8ad4f624c2/events.out.tfevents.1723466376.3e8ad4f624c2.653.0 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:da7d06a221063fc89119b06a4561884da9865de220e6a8f656fc6a5db56b3903 +size 8320 diff --git a/runs/Aug12_12-50-26_3e8ad4f624c2/events.out.tfevents.1723467032.3e8ad4f624c2.653.1 b/runs/Aug12_12-50-26_3e8ad4f624c2/events.out.tfevents.1723467032.3e8ad4f624c2.653.1 new file mode 100644 index 0000000000000000000000000000000000000000..59536ade73439a049a80a0efb5b4abc9c60c3826 --- /dev/null +++ b/runs/Aug12_12-50-26_3e8ad4f624c2/events.out.tfevents.1723467032.3e8ad4f624c2.653.1 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d905603b4a1c1919fd0a58f9fc506d59ee99cf33b610a6532c3c46ae848fb82c +size 32601 diff --git a/runs/Aug12_12-50-26_3e8ad4f624c2/events.out.tfevents.1723467499.3e8ad4f624c2.653.2 b/runs/Aug12_12-50-26_3e8ad4f624c2/events.out.tfevents.1723467499.3e8ad4f624c2.653.2 new file mode 100644 index 0000000000000000000000000000000000000000..36b8f4e329a0730fa3020f7d4e0af2d2af7f45b6 --- /dev/null +++ b/runs/Aug12_12-50-26_3e8ad4f624c2/events.out.tfevents.1723467499.3e8ad4f624c2.653.2 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:72cbeda15d27c684cb6cc4ab74aaa89191c22dbe663421c5ac242308598549d2 +size 6872 diff --git a/runs/Aug12_12-58-56_3e8ad4f624c2/events.out.tfevents.1723467544.3e8ad4f624c2.653.3 b/runs/Aug12_12-58-56_3e8ad4f624c2/events.out.tfevents.1723467544.3e8ad4f624c2.653.3 new file mode 100644 index 0000000000000000000000000000000000000000..e6ed1c7b5bdf67281973d93294aa83512b0cb55b --- /dev/null +++ b/runs/Aug12_12-58-56_3e8ad4f624c2/events.out.tfevents.1723467544.3e8ad4f624c2.653.3 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c89133ece4a3efdb4d9e08091205f549970db99cda96e2a865bad490b23d9d96 +size 16155 diff --git a/runs/Aug12_13-07-21_3e8ad4f624c2/events.out.tfevents.1723468049.3e8ad4f624c2.653.4 b/runs/Aug12_13-07-21_3e8ad4f624c2/events.out.tfevents.1723468049.3e8ad4f624c2.653.4 new file mode 100644 index 0000000000000000000000000000000000000000..f6e6b58979fb2daccc608cbcd0b1ffde426be7ea --- /dev/null +++ b/runs/Aug12_13-07-21_3e8ad4f624c2/events.out.tfevents.1723468049.3e8ad4f624c2.653.4 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:40f8a7a157beb5e573a99eea4ad26258ec2ab21e3382d94a26dcb58da70d3e82 +size 7078 diff --git a/runs/Aug12_13-18-41_3e8ad4f624c2/events.out.tfevents.1723468728.3e8ad4f624c2.14956.0 b/runs/Aug12_13-18-41_3e8ad4f624c2/events.out.tfevents.1723468728.3e8ad4f624c2.14956.0 new file mode 100644 index 0000000000000000000000000000000000000000..65686b9ef3d4458986f7079544371730923f6196 --- /dev/null +++ b/runs/Aug12_13-18-41_3e8ad4f624c2/events.out.tfevents.1723468728.3e8ad4f624c2.14956.0 @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6deacbf9bcc0be294fe5de30180c67f221876e8b277c0995e3a24f3e7f636621 +size 11418 diff --git a/test_whisper.xlsx b/test_whisper.xlsx new file mode 100644 index 0000000000000000000000000000000000000000..dd2be974ad93638e563ad72105be1ef32dcd1c10 Binary files /dev/null and b/test_whisper.xlsx differ diff --git a/training_args.bin b/training_args.bin new file mode 100644 index 0000000000000000000000000000000000000000..0d5e5990c1f85bdbe1cad8fa473794497afefb3e --- /dev/null +++ b/training_args.bin @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:41fee3748b5e4d38b71c95ed50576544a333b45e2a1e738257e33a3775255c9d +size 5304 diff --git a/whisper_Diny.xlsx b/whisper_Diny.xlsx new file mode 100644 index 0000000000000000000000000000000000000000..1f60bfd6f704b5a68e88eb27d78e5c0c5b4ee06f Binary files /dev/null and b/whisper_Diny.xlsx differ