diff --git "a/train_kh.ipynb" "b/train_kh.ipynb" --- "a/train_kh.ipynb" +++ "b/train_kh.ipynb" @@ -3,7 +3,7 @@ { "cell_type": "code", "execution_count": 1, - "id": "0ee7433e", + "id": "ce307102", "metadata": {}, "outputs": [], "source": [ @@ -16,7 +16,7 @@ { "cell_type": "code", "execution_count": null, - "id": "90323ec7", + "id": "2f4fbaa9", "metadata": { "collapsed": true, "jupyter": { @@ -19167,7 +19167,7 @@ }, { "cell_type": "markdown", - "id": "eda834f4", + "id": "a4ca0989", "metadata": {}, "source": [ "### Load KH Data" @@ -19176,7 +19176,7 @@ { "cell_type": "code", "execution_count": 6, - "id": "e8b86dab", + "id": "a6f8b139", "metadata": {}, "outputs": [], "source": [ @@ -19198,30 +19198,112 @@ }, { "cell_type": "code", - "execution_count": 2, - "id": "0b17a0e1", + "execution_count": 3, + "id": "e39ba56b", "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ - "Using custom data configuration default-decaf49f8e8b5be8\n", - "Reusing dataset csv (/workspace/.cache/huggingface/datasets/csv/default-decaf49f8e8b5be8/0.0.0/6b9057d9e23d9d8a2f05b985917a0da84d70c5dae3d22ddd8a3f22fb01c69d9e)\n", - "Using custom data configuration default-2ae3784a8d52f12b\n", - "Reusing dataset csv (/workspace/.cache/huggingface/datasets/csv/default-2ae3784a8d52f12b/0.0.0/6b9057d9e23d9d8a2f05b985917a0da84d70c5dae3d22ddd8a3f22fb01c69d9e)\n" + "Using custom data configuration default-ac5c5990503e1315\n" + ] + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "Downloading and preparing dataset csv/default to /workspace/.cache/huggingface/datasets/csv/default-ac5c5990503e1315/0.0.0/6b9057d9e23d9d8a2f05b985917a0da84d70c5dae3d22ddd8a3f22fb01c69d9e...\n" + ] + }, + { + "data": { + "application/vnd.jupyter.widget-view+json": { + "model_id": "53e8e843027a446bb619fa68a6dd87eb", + "version_major": 2, + "version_minor": 0 + }, + "text/plain": [ + " 0%| | 0/1 [00:00\n", - " \n", + " \n", " Your browser does not support the audio element.\n", " \n", " " @@ -19587,7 +19576,7 @@ "" ] }, - "execution_count": 17, + "execution_count": 16, "metadata": {}, "output_type": "execute_result" } @@ -19607,8 +19596,8 @@ }, { "cell_type": "code", - "execution_count": 18, - "id": "c39872d6", + "execution_count": 17, + "id": "6bf29928", "metadata": {}, "outputs": [], "source": [ @@ -19629,38 +19618,10 @@ }, { "cell_type": "code", - "execution_count": 19, - "id": "fef54a48", - "metadata": { - "collapsed": true, - "jupyter": { - "outputs_hidden": true - } - }, - "outputs": [ - { - "name": "stderr", - "output_type": "stream", - "text": [ - "Loading cached processed dataset at /workspace/.cache/huggingface/datasets/csv/default-decaf49f8e8b5be8/0.0.0/6b9057d9e23d9d8a2f05b985917a0da84d70c5dae3d22ddd8a3f22fb01c69d9e/cache-a8ad7f3bec152712.arrow\n", - "Loading cached processed dataset at /workspace/.cache/huggingface/datasets/csv/default-decaf49f8e8b5be8/0.0.0/6b9057d9e23d9d8a2f05b985917a0da84d70c5dae3d22ddd8a3f22fb01c69d9e/cache-5802da9af6ac9ac7.arrow\n", - "Loading cached processed dataset at /workspace/.cache/huggingface/datasets/csv/default-decaf49f8e8b5be8/0.0.0/6b9057d9e23d9d8a2f05b985917a0da84d70c5dae3d22ddd8a3f22fb01c69d9e/cache-418585d4baf07152.arrow\n", - "Loading cached processed dataset at /workspace/.cache/huggingface/datasets/csv/default-decaf49f8e8b5be8/0.0.0/6b9057d9e23d9d8a2f05b985917a0da84d70c5dae3d22ddd8a3f22fb01c69d9e/cache-c7e5028c91005615.arrow\n", - "Loading cached processed dataset at /workspace/.cache/huggingface/datasets/csv/default-decaf49f8e8b5be8/0.0.0/6b9057d9e23d9d8a2f05b985917a0da84d70c5dae3d22ddd8a3f22fb01c69d9e/cache-b96a6332cc5af3be.arrow\n", - "Loading cached processed dataset at /workspace/.cache/huggingface/datasets/csv/default-decaf49f8e8b5be8/0.0.0/6b9057d9e23d9d8a2f05b985917a0da84d70c5dae3d22ddd8a3f22fb01c69d9e/cache-6a1544fcabe8e1c5.arrow\n", - "Loading cached processed dataset at /workspace/.cache/huggingface/datasets/csv/default-decaf49f8e8b5be8/0.0.0/6b9057d9e23d9d8a2f05b985917a0da84d70c5dae3d22ddd8a3f22fb01c69d9e/cache-4926c7991e987d55.arrow\n", - "Loading cached processed dataset at /workspace/.cache/huggingface/datasets/csv/default-decaf49f8e8b5be8/0.0.0/6b9057d9e23d9d8a2f05b985917a0da84d70c5dae3d22ddd8a3f22fb01c69d9e/cache-b607d477202e12db.arrow\n", - "Loading cached processed dataset at /workspace/.cache/huggingface/datasets/csv/default-decaf49f8e8b5be8/0.0.0/6b9057d9e23d9d8a2f05b985917a0da84d70c5dae3d22ddd8a3f22fb01c69d9e/cache-a04833a515432724.arrow\n", - "Loading cached processed dataset at /workspace/.cache/huggingface/datasets/csv/default-decaf49f8e8b5be8/0.0.0/6b9057d9e23d9d8a2f05b985917a0da84d70c5dae3d22ddd8a3f22fb01c69d9e/cache-bee3e39e2b69f652.arrow\n", - "Loading cached processed dataset at /workspace/.cache/huggingface/datasets/csv/default-decaf49f8e8b5be8/0.0.0/6b9057d9e23d9d8a2f05b985917a0da84d70c5dae3d22ddd8a3f22fb01c69d9e/cache-ad833922c61a3f31.arrow\n", - "Loading cached processed dataset at /workspace/.cache/huggingface/datasets/csv/default-decaf49f8e8b5be8/0.0.0/6b9057d9e23d9d8a2f05b985917a0da84d70c5dae3d22ddd8a3f22fb01c69d9e/cache-ecf0c779c655274d.arrow\n", - "Loading cached processed dataset at /workspace/.cache/huggingface/datasets/csv/default-decaf49f8e8b5be8/0.0.0/6b9057d9e23d9d8a2f05b985917a0da84d70c5dae3d22ddd8a3f22fb01c69d9e/cache-2ba17aaff236f685.arrow\n", - "Loading cached processed dataset at /workspace/.cache/huggingface/datasets/csv/default-decaf49f8e8b5be8/0.0.0/6b9057d9e23d9d8a2f05b985917a0da84d70c5dae3d22ddd8a3f22fb01c69d9e/cache-5ab7f29ea26c63ef.arrow\n", - "Loading cached processed dataset at /workspace/.cache/huggingface/datasets/csv/default-decaf49f8e8b5be8/0.0.0/6b9057d9e23d9d8a2f05b985917a0da84d70c5dae3d22ddd8a3f22fb01c69d9e/cache-240348bb32bdbb06.arrow\n", - "Loading cached processed dataset at /workspace/.cache/huggingface/datasets/csv/default-decaf49f8e8b5be8/0.0.0/6b9057d9e23d9d8a2f05b985917a0da84d70c5dae3d22ddd8a3f22fb01c69d9e/cache-38e1126fb0ce94f8.arrow\n" - ] - } - ], + "execution_count": 18, + "id": "6f96c9bd", + "metadata": {}, + "outputs": [], "source": [ "common_voice_train = common_voice_train.map(prepare_dataset, remove_columns=common_voice_train.column_names, num_proc=16)\n", "common_voice_valid = common_voice_valid.map(prepare_dataset, remove_columns=common_voice_valid.column_names, num_proc=16)" @@ -19669,7 +19630,7 @@ { "cell_type": "code", "execution_count": 20, - "id": "2f280b0d", + "id": "396d1e42", "metadata": {}, "outputs": [], "source": [ @@ -19680,8 +19641,8 @@ }, { "cell_type": "code", - "execution_count": 21, - "id": "c9dec52e", + "execution_count": 19, + "id": "bcab8d7a", "metadata": {}, "outputs": [], "source": [ @@ -19740,8 +19701,8 @@ }, { "cell_type": "code", - "execution_count": 22, - "id": "639dcc23", + "execution_count": 20, + "id": "c6051113", "metadata": {}, "outputs": [], "source": [ @@ -19750,8 +19711,8 @@ }, { "cell_type": "code", - "execution_count": 23, - "id": "3bb04288", + "execution_count": 21, + "id": "d6247fff", "metadata": {}, "outputs": [], "source": [ @@ -19761,8 +19722,8 @@ }, { "cell_type": "code", - "execution_count": 24, - "id": "9ba8858b", + "execution_count": 22, + "id": "ad02923a", "metadata": {}, "outputs": [], "source": [ @@ -19782,18 +19743,18 @@ }, { "cell_type": "code", - "execution_count": 25, - "id": "434869f9", + "execution_count": 23, + "id": "fb8f42c5", "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ - "Some weights of the model checkpoint at facebook/wav2vec2-xls-r-300m were not used when initializing Wav2Vec2ForCTC: ['project_hid.bias', 'quantizer.codevectors', 'quantizer.weight_proj.weight', 'project_q.weight', 'project_hid.weight', 'quantizer.weight_proj.bias', 'project_q.bias']\n", + "Some weights of the model checkpoint at facebook/wav2vec2-xls-r-300m were not used when initializing Wav2Vec2ForCTC: ['quantizer.codevectors', 'project_hid.weight', 'project_q.weight', 'project_q.bias', 'quantizer.weight_proj.bias', 'project_hid.bias', 'quantizer.weight_proj.weight']\n", "- This IS expected if you are initializing Wav2Vec2ForCTC from the checkpoint of a model trained on another task or with another architecture (e.g. initializing a BertForSequenceClassification model from a BertForPreTraining model).\n", "- This IS NOT expected if you are initializing Wav2Vec2ForCTC from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model).\n", - "Some weights of Wav2Vec2ForCTC were not initialized from the model checkpoint at facebook/wav2vec2-xls-r-300m and are newly initialized: ['lm_head.weight', 'lm_head.bias']\n", + "Some weights of Wav2Vec2ForCTC were not initialized from the model checkpoint at facebook/wav2vec2-xls-r-300m and are newly initialized: ['lm_head.bias', 'lm_head.weight']\n", "You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.\n" ] } @@ -19818,8 +19779,8 @@ }, { "cell_type": "code", - "execution_count": 26, - "id": "9ffb97fd", + "execution_count": 24, + "id": "2b72369f", "metadata": {}, "outputs": [], "source": [ @@ -19828,8 +19789,8 @@ }, { "cell_type": "code", - "execution_count": 27, - "id": "c83b8d4e", + "execution_count": 25, + "id": "fe18c060", "metadata": {}, "outputs": [], "source": [ @@ -19843,11 +19804,11 @@ " evaluation_strategy=\"steps\",\n", " gradient_checkpointing=True,\n", " fp16=True,\n", - " num_train_epochs=100,\n", + " num_train_epochs=70,\n", " save_steps=400,\n", " eval_steps=400,\n", " logging_steps=100,\n", - " learning_rate=5e-5,\n", + " learning_rate=3e-5,\n", " warmup_steps=1000,\n", " save_total_limit=3,\n", " load_best_model_at_end=True\n", @@ -19856,8 +19817,8 @@ }, { "cell_type": "code", - "execution_count": 29, - "id": "7352a29a", + "execution_count": 26, + "id": "45fbfbbe", "metadata": {}, "outputs": [ { @@ -19884,8 +19845,8 @@ }, { "cell_type": "code", - "execution_count": 30, - "id": "5a73ff08", + "execution_count": 27, + "id": "8ae529c1", "metadata": {}, "outputs": [ { @@ -19896,12 +19857,12 @@ "/opt/conda/lib/python3.8/site-packages/transformers/optimization.py:306: FutureWarning: This implementation of AdamW is deprecated and will be removed in a future version. Use thePyTorch implementation torch.optim.AdamW instead, or set `no_deprecation_warning=True` to disable this warning\n", " warnings.warn(\n", "***** Running training *****\n", - " Num examples = 2353\n", - " Num Epochs = 100\n", + " Num examples = 2586\n", + " Num Epochs = 70\n", " Instantaneous batch size per device = 8\n", " Total train batch size (w. parallel, distributed & accumulation) = 32\n", " Gradient Accumulation steps = 4\n", - " Total optimization steps = 7300\n" + " Total optimization steps = 5670\n" ] }, { @@ -19910,8 +19871,8 @@ "\n", "
\n", " \n", - " \n", - " [7300/7300 4:05:50, Epoch 99/100]\n", + " \n", + " [5670/5670 3:13:57, Epoch 70/70]\n", "
\n", " \n", " \n", @@ -19925,111 +19886,87 @@ " \n", " \n", " \n", - " \n", - " \n", + " \n", + " \n", " \n", " \n", " \n", " \n", - " \n", - " \n", + " \n", + " \n", " \n", " \n", " \n", " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", " \n", " \n", " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", " \n", " \n", " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", " \n", " \n", " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", " \n", " \n", " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", " \n", " \n", " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", " \n", " \n", " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", " \n", " \n", " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", " \n", " \n", " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", " \n", " \n", " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", " \n", " \n", " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", " \n", " \n", " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", - " \n", + " \n", + " \n", + " \n", " \n", " \n", "
4005.0795004.4120576.1438005.3190121.000000
8003.5658003.5203373.6114003.5949151.000000
12003.3689002.8983760.9995963.5098003.5386261.000000
16002.0100001.0040760.7288143.3822003.2176741.000000
20001.6783000.6940640.5988702.3745001.5386380.861423
24001.5270000.5599430.5282492.0197001.1459860.711610
28001.4278000.4826760.4806301.8515000.9299400.677903
32001.3458000.4428690.4531881.7277000.8082370.617978
36001.2893000.4156090.4330101.6577000.7192690.558052
40001.2441000.4020460.4039551.5752000.6624380.528090
44001.1880000.3777470.3866021.5401000.6389030.490637
48001.1628000.3607460.3857951.5248000.6022350.490637
52001.1324000.3533980.3603711.4982000.5868350.475655
56001.0969000.3428450.362389
60001.0897000.3387430.356739
64001.0625000.3338740.349879
68001.0601000.3288150.344633
72001.0474000.3280940.3462471.4859000.5801720.479401

" @@ -20047,33 +19984,34 @@ "text": [ "The following columns in the evaluation set don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.\n", "***** Running Evaluation *****\n", - " Num examples = 262\n", + " Num examples = 29\n", " Batch size = 8\n", "Saving model checkpoint to ./checkpoint-400\n", "Configuration saved in ./checkpoint-400/config.json\n", "Model weights saved in ./checkpoint-400/pytorch_model.bin\n", "Configuration saved in ./checkpoint-400/preprocessor_config.json\n", + "Deleting older checkpoint [checkpoint-6400] due to args.save_total_limit\n", "The following columns in the evaluation set don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.\n", "***** Running Evaluation *****\n", - " Num examples = 262\n", + " Num examples = 29\n", " Batch size = 8\n", "Saving model checkpoint to ./checkpoint-800\n", "Configuration saved in ./checkpoint-800/config.json\n", "Model weights saved in ./checkpoint-800/pytorch_model.bin\n", "Configuration saved in ./checkpoint-800/preprocessor_config.json\n", - "Deleting older checkpoint [checkpoint-3200] due to args.save_total_limit\n", + "Deleting older checkpoint [checkpoint-6800] due to args.save_total_limit\n", "The following columns in the evaluation set don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.\n", "***** Running Evaluation *****\n", - " Num examples = 262\n", + " Num examples = 29\n", " Batch size = 8\n", "Saving model checkpoint to ./checkpoint-1200\n", "Configuration saved in ./checkpoint-1200/config.json\n", "Model weights saved in ./checkpoint-1200/pytorch_model.bin\n", "Configuration saved in ./checkpoint-1200/preprocessor_config.json\n", - "Deleting older checkpoint [checkpoint-3600] due to args.save_total_limit\n", + "Deleting older checkpoint [checkpoint-7200] due to args.save_total_limit\n", "The following columns in the evaluation set don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.\n", "***** Running Evaluation *****\n", - " Num examples = 262\n", + " Num examples = 29\n", " Batch size = 8\n", "Saving model checkpoint to ./checkpoint-1600\n", "Configuration saved in ./checkpoint-1600/config.json\n", @@ -20082,7 +20020,7 @@ "Deleting older checkpoint [checkpoint-400] due to args.save_total_limit\n", "The following columns in the evaluation set don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.\n", "***** Running Evaluation *****\n", - " Num examples = 262\n", + " Num examples = 29\n", " Batch size = 8\n", "Saving model checkpoint to ./checkpoint-2000\n", "Configuration saved in ./checkpoint-2000/config.json\n", @@ -20091,7 +20029,7 @@ "Deleting older checkpoint [checkpoint-800] due to args.save_total_limit\n", "The following columns in the evaluation set don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.\n", "***** Running Evaluation *****\n", - " Num examples = 262\n", + " Num examples = 29\n", " Batch size = 8\n", "Saving model checkpoint to ./checkpoint-2400\n", "Configuration saved in ./checkpoint-2400/config.json\n", @@ -20100,7 +20038,7 @@ "Deleting older checkpoint [checkpoint-1200] due to args.save_total_limit\n", "The following columns in the evaluation set don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.\n", "***** Running Evaluation *****\n", - " Num examples = 262\n", + " Num examples = 29\n", " Batch size = 8\n", "Saving model checkpoint to ./checkpoint-2800\n", "Configuration saved in ./checkpoint-2800/config.json\n", @@ -20109,7 +20047,7 @@ "Deleting older checkpoint [checkpoint-1600] due to args.save_total_limit\n", "The following columns in the evaluation set don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.\n", "***** Running Evaluation *****\n", - " Num examples = 262\n", + " Num examples = 29\n", " Batch size = 8\n", "Saving model checkpoint to ./checkpoint-3200\n", "Configuration saved in ./checkpoint-3200/config.json\n", @@ -20118,7 +20056,7 @@ "Deleting older checkpoint [checkpoint-2000] due to args.save_total_limit\n", "The following columns in the evaluation set don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.\n", "***** Running Evaluation *****\n", - " Num examples = 262\n", + " Num examples = 29\n", " Batch size = 8\n", "Saving model checkpoint to ./checkpoint-3600\n", "Configuration saved in ./checkpoint-3600/config.json\n", @@ -20127,7 +20065,7 @@ "Deleting older checkpoint [checkpoint-2400] due to args.save_total_limit\n", "The following columns in the evaluation set don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.\n", "***** Running Evaluation *****\n", - " Num examples = 262\n", + " Num examples = 29\n", " Batch size = 8\n", "Saving model checkpoint to ./checkpoint-4000\n", "Configuration saved in ./checkpoint-4000/config.json\n", @@ -20136,7 +20074,7 @@ "Deleting older checkpoint [checkpoint-2800] due to args.save_total_limit\n", "The following columns in the evaluation set don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.\n", "***** Running Evaluation *****\n", - " Num examples = 262\n", + " Num examples = 29\n", " Batch size = 8\n", "Saving model checkpoint to ./checkpoint-4400\n", "Configuration saved in ./checkpoint-4400/config.json\n", @@ -20145,7 +20083,7 @@ "Deleting older checkpoint [checkpoint-3200] due to args.save_total_limit\n", "The following columns in the evaluation set don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.\n", "***** Running Evaluation *****\n", - " Num examples = 262\n", + " Num examples = 29\n", " Batch size = 8\n", "Saving model checkpoint to ./checkpoint-4800\n", "Configuration saved in ./checkpoint-4800/config.json\n", @@ -20154,7 +20092,7 @@ "Deleting older checkpoint [checkpoint-3600] due to args.save_total_limit\n", "The following columns in the evaluation set don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.\n", "***** Running Evaluation *****\n", - " Num examples = 262\n", + " Num examples = 29\n", " Batch size = 8\n", "Saving model checkpoint to ./checkpoint-5200\n", "Configuration saved in ./checkpoint-5200/config.json\n", @@ -20163,64 +20101,28 @@ "Deleting older checkpoint [checkpoint-4000] due to args.save_total_limit\n", "The following columns in the evaluation set don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.\n", "***** Running Evaluation *****\n", - " Num examples = 262\n", + " Num examples = 29\n", " Batch size = 8\n", "Saving model checkpoint to ./checkpoint-5600\n", "Configuration saved in ./checkpoint-5600/config.json\n", "Model weights saved in ./checkpoint-5600/pytorch_model.bin\n", "Configuration saved in ./checkpoint-5600/preprocessor_config.json\n", "Deleting older checkpoint [checkpoint-4400] due to args.save_total_limit\n", - "The following columns in the evaluation set don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.\n", - "***** Running Evaluation *****\n", - " Num examples = 262\n", - " Batch size = 8\n", - "Saving model checkpoint to ./checkpoint-6000\n", - "Configuration saved in ./checkpoint-6000/config.json\n", - "Model weights saved in ./checkpoint-6000/pytorch_model.bin\n", - "Configuration saved in ./checkpoint-6000/preprocessor_config.json\n", - "Deleting older checkpoint [checkpoint-4800] due to args.save_total_limit\n", - "The following columns in the evaluation set don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.\n", - "***** Running Evaluation *****\n", - " Num examples = 262\n", - " Batch size = 8\n", - "Saving model checkpoint to ./checkpoint-6400\n", - "Configuration saved in ./checkpoint-6400/config.json\n", - "Model weights saved in ./checkpoint-6400/pytorch_model.bin\n", - "Configuration saved in ./checkpoint-6400/preprocessor_config.json\n", - "Deleting older checkpoint [checkpoint-5200] due to args.save_total_limit\n", - "The following columns in the evaluation set don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.\n", - "***** Running Evaluation *****\n", - " Num examples = 262\n", - " Batch size = 8\n", - "Saving model checkpoint to ./checkpoint-6800\n", - "Configuration saved in ./checkpoint-6800/config.json\n", - "Model weights saved in ./checkpoint-6800/pytorch_model.bin\n", - "Configuration saved in ./checkpoint-6800/preprocessor_config.json\n", - "Deleting older checkpoint [checkpoint-5600] due to args.save_total_limit\n", - "The following columns in the evaluation set don't have a corresponding argument in `Wav2Vec2ForCTC.forward` and have been ignored: input_length.\n", - "***** Running Evaluation *****\n", - " Num examples = 262\n", - " Batch size = 8\n", - "Saving model checkpoint to ./checkpoint-7200\n", - "Configuration saved in ./checkpoint-7200/config.json\n", - "Model weights saved in ./checkpoint-7200/pytorch_model.bin\n", - "Configuration saved in ./checkpoint-7200/preprocessor_config.json\n", - "Deleting older checkpoint [checkpoint-6000] due to args.save_total_limit\n", "\n", "\n", "Training completed. Do not forget to share your model on huggingface.co/models =)\n", "\n", "\n", - "Loading best model from ./checkpoint-7200 (score: 0.3280937969684601).\n" + "Loading best model from ./checkpoint-5600 (score: 0.580172061920166).\n" ] }, { "data": { "text/plain": [ - "TrainOutput(global_step=7300, training_loss=2.0282830110314776, metrics={'train_runtime': 14754.0737, 'train_samples_per_second': 15.948, 'train_steps_per_second': 0.495, 'total_flos': 3.5572390287970673e+19, 'train_loss': 2.0282830110314776, 'epoch': 99.99})" + "TrainOutput(global_step=5670, training_loss=2.8286902425991376, metrics={'train_runtime': 11640.9835, 'train_samples_per_second': 15.55, 'train_steps_per_second': 0.487, 'total_flos': 2.7418278911692476e+19, 'train_loss': 2.8286902425991376, 'epoch': 70.0})" ] }, - "execution_count": 30, + "execution_count": 27, "metadata": {}, "output_type": "execute_result" } @@ -20232,7 +20134,7 @@ { "cell_type": "code", "execution_count": 31, - "id": "967962d1", + "id": "5d4ffe69", "metadata": {}, "outputs": [ { @@ -20253,7 +20155,7 @@ { "cell_type": "code", "execution_count": 32, - "id": "da40a75c", + "id": "a2691979", "metadata": {}, "outputs": [ { @@ -20286,7 +20188,7 @@ { "cell_type": "code", "execution_count": 34, - "id": "24166e72", + "id": "edf10e59", "metadata": {}, "outputs": [], "source": [ @@ -20303,7 +20205,7 @@ { "cell_type": "code", "execution_count": 35, - "id": "95d69b2e", + "id": "535c0037", "metadata": {}, "outputs": [ { @@ -20322,7 +20224,7 @@ { "cell_type": "code", "execution_count": 36, - "id": "d60a731d", + "id": "7a7896ca", "metadata": {}, "outputs": [ { @@ -20374,7 +20276,7 @@ { "cell_type": "code", "execution_count": 38, - "id": "beca9a8c", + "id": "81e31a96", "metadata": {}, "outputs": [ { @@ -20395,7 +20297,7 @@ { "cell_type": "code", "execution_count": null, - "id": "20063dbc", + "id": "b387df57", "metadata": {}, "outputs": [], "source": []