bengali-speech-recognition.ipynb

{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"name":"python","version":"3.10.12","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"# This Python 3 environment comes with many helpful analytics libraries installed\n# It is defined by the kaggle/python Docker image: https://github.com/kaggle/docker-python\n# For example, here's several helpful packages to load\n\nimport numpy as np # linear algebra\nimport pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\n\n# Input data files are available in the read-only \"../input/\" directory\n# For example, running this (by clicking run or pressing Shift+Enter) will list all files under the input directory\n\nimport os\nfor dirname, _, filenames in os.walk('/kaggle/input'):\n    for filename in filenames:\n        print(os.path.join(dirname, filename))\n\n# You can write up to 20GB to the current directory (/kaggle/working/) that gets preserved as output when you create a version using \"Save & Run All\" \n# You can also write temporary files to /kaggle/temp/, but they won't be saved outside of the current session","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## try wav2vec with librosa","metadata":{}},{"cell_type":"code","source":"!pip install transformers librosa numpy scipy matplotlib torch","metadata":{"execution":{"iopub.status.busy":"2023-07-22T06:48:00.721868Z","iopub.execute_input":"2023-07-22T06:48:00.722292Z","iopub.status.idle":"2023-07-22T06:48:12.341418Z","shell.execute_reply.started":"2023-07-22T06:48:00.722257Z","shell.execute_reply":"2023-07-22T06:48:12.339984Z"},"trusted":true},"execution_count":5,"outputs":[{"name":"stdout","text":"Requirement already satisfied: transformers in /opt/conda/lib/python3.10/site-packages (4.30.2)\nRequirement already satisfied: librosa in /opt/conda/lib/python3.10/site-packages (0.10.0.post2)\nRequirement already satisfied: numpy in /opt/conda/lib/python3.10/site-packages (1.23.5)\nRequirement already satisfied: scipy in /opt/conda/lib/python3.10/site-packages (1.11.1)\nRequirement already satisfied: matplotlib in /opt/conda/lib/python3.10/site-packages (3.7.1)\nRequirement already satisfied: torch in /opt/conda/lib/python3.10/site-packages (2.0.0)\nRequirement already satisfied: filelock in /opt/conda/lib/python3.10/site-packages (from transformers) (3.12.2)\nRequirement already satisfied: huggingface-hub<1.0,>=0.14.1 in /opt/conda/lib/python3.10/site-packages (from transformers) (0.16.4)\nRequirement already satisfied: packaging>=20.0 in /opt/conda/lib/python3.10/site-packages (from transformers) (21.3)\nRequirement already satisfied: pyyaml>=5.1 in /opt/conda/lib/python3.10/site-packages (from transformers) (6.0)\nRequirement already satisfied: regex!=2019.12.17 in /opt/conda/lib/python3.10/site-packages (from transformers) (2023.6.3)\nRequirement already satisfied: requests in /opt/conda/lib/python3.10/site-packages (from transformers) (2.31.0)\nRequirement already satisfied: tokenizers!=0.11.3,<0.14,>=0.11.1 in /opt/conda/lib/python3.10/site-packages (from transformers) (0.13.3)\nRequirement already satisfied: safetensors>=0.3.1 in /opt/conda/lib/python3.10/site-packages (from transformers) (0.3.1)\nRequirement already satisfied: tqdm>=4.27 in /opt/conda/lib/python3.10/site-packages (from transformers) (4.65.0)\nRequirement already satisfied: audioread>=2.1.9 in /opt/conda/lib/python3.10/site-packages (from librosa) (3.0.0)\nRequirement already satisfied: scikit-learn>=0.20.0 in /opt/conda/lib/python3.10/site-packages (from librosa) (1.2.2)\nRequirement already satisfied: joblib>=0.14 in /opt/conda/lib/python3.10/site-packages (from librosa) (1.2.0)\nRequirement already satisfied: decorator>=4.3.0 in /opt/conda/lib/python3.10/site-packages (from librosa) (5.1.1)\nRequirement already satisfied: numba>=0.51.0 in /opt/conda/lib/python3.10/site-packages (from librosa) (0.57.1)\nRequirement already satisfied: soundfile>=0.12.1 in /opt/conda/lib/python3.10/site-packages (from librosa) (0.12.1)\nRequirement already satisfied: pooch<1.7,>=1.0 in /opt/conda/lib/python3.10/site-packages (from librosa) (1.6.0)\nRequirement already satisfied: soxr>=0.3.2 in /opt/conda/lib/python3.10/site-packages (from librosa) (0.3.5)\nRequirement already satisfied: typing-extensions>=4.1.1 in /opt/conda/lib/python3.10/site-packages (from librosa) (4.6.3)\nRequirement already satisfied: lazy-loader>=0.1 in /opt/conda/lib/python3.10/site-packages (from librosa) (0.2)\nRequirement already satisfied: msgpack>=1.0 in /opt/conda/lib/python3.10/site-packages (from librosa) (1.0.5)\nRequirement already satisfied: contourpy>=1.0.1 in /opt/conda/lib/python3.10/site-packages (from matplotlib) (1.1.0)\nRequirement already satisfied: cycler>=0.10 in /opt/conda/lib/python3.10/site-packages (from matplotlib) (0.11.0)\nRequirement already satisfied: fonttools>=4.22.0 in /opt/conda/lib/python3.10/site-packages (from matplotlib) (4.40.0)\nRequirement already satisfied: kiwisolver>=1.0.1 in /opt/conda/lib/python3.10/site-packages (from matplotlib) (1.4.4)\nRequirement already satisfied: pillow>=6.2.0 in /opt/conda/lib/python3.10/site-packages (from matplotlib) (9.5.0)\nRequirement already satisfied: pyparsing>=2.3.1 in /opt/conda/lib/python3.10/site-packages (from matplotlib) (3.0.9)\nRequirement already satisfied: python-dateutil>=2.7 in /opt/conda/lib/python3.10/site-packages (from matplotlib) (2.8.2)\nRequirement already satisfied: sympy in /opt/conda/lib/python3.10/site-packages (from torch) (1.12)\nRequirement already satisfied: networkx in /opt/conda/lib/python3.10/site-packages (from torch) (3.1)\nRequirement already satisfied: jinja2 in /opt/conda/lib/python3.10/site-packages (from torch) (3.1.2)\nRequirement already satisfied: fsspec in /opt/conda/lib/python3.10/site-packages (from huggingface-hub<1.0,>=0.14.1->transformers) (2023.6.0)\nRequirement already satisfied: llvmlite<0.41,>=0.40.0dev0 in /opt/conda/lib/python3.10/site-packages (from numba>=0.51.0->librosa) (0.40.1)\nRequirement already satisfied: appdirs>=1.3.0 in /opt/conda/lib/python3.10/site-packages (from pooch<1.7,>=1.0->librosa) (1.4.4)\nRequirement already satisfied: six>=1.5 in /opt/conda/lib/python3.10/site-packages (from python-dateutil>=2.7->matplotlib) (1.16.0)\nRequirement already satisfied: charset-normalizer<4,>=2 in /opt/conda/lib/python3.10/site-packages (from requests->transformers) (3.1.0)\nRequirement already satisfied: idna<4,>=2.5 in /opt/conda/lib/python3.10/site-packages (from requests->transformers) (3.4)\nRequirement already satisfied: urllib3<3,>=1.21.1 in /opt/conda/lib/python3.10/site-packages (from requests->transformers) (1.26.15)\nRequirement already satisfied: certifi>=2017.4.17 in /opt/conda/lib/python3.10/site-packages (from requests->transformers) (2023.5.7)\nRequirement already satisfied: threadpoolctl>=2.0.0 in /opt/conda/lib/python3.10/site-packages (from scikit-learn>=0.20.0->librosa) (3.1.0)\nRequirement already satisfied: cffi>=1.0 in /opt/conda/lib/python3.10/site-packages (from soundfile>=0.12.1->librosa) (1.15.1)\nRequirement already satisfied: MarkupSafe>=2.0 in /opt/conda/lib/python3.10/site-packages (from jinja2->torch) (2.1.3)\nRequirement already satisfied: mpmath>=0.19 in /opt/conda/lib/python3.10/site-packages (from sympy->torch) (1.3.0)\nRequirement already satisfied: pycparser in /opt/conda/lib/python3.10/site-packages (from cffi>=1.0->soundfile>=0.12.1->librosa) (2.21)\n","output_type":"stream"}]},{"cell_type":"code","source":"import numpy as np\nimport pandas as pd\nimport matplotlib.pyplot as plt\n\nimport librosa\n\nimport torch\n\n#Importing Wav2Vec\nfrom transformers import Wav2Vec2ForCTC, Wav2Vec2Tokenizer","metadata":{"execution":{"iopub.status.busy":"2023-07-22T06:48:20.718089Z","iopub.execute_input":"2023-07-22T06:48:20.718487Z","iopub.status.idle":"2023-07-22T06:48:20.723952Z","shell.execute_reply.started":"2023-07-22T06:48:20.718457Z","shell.execute_reply":"2023-07-22T06:48:20.722904Z"},"trusted":true},"execution_count":7,"outputs":[]},{"cell_type":"code","source":"audio, rate = librosa.load(\"/kaggle/input/bengaliai-speech/train_mp3s/000024b3d810.mp3\", sr = 16000)\naudio","metadata":{"execution":{"iopub.status.busy":"2023-07-22T06:49:07.492071Z","iopub.execute_input":"2023-07-22T06:49:07.492975Z","iopub.status.idle":"2023-07-22T06:49:16.736877Z","shell.execute_reply.started":"2023-07-22T06:49:07.492940Z","shell.execute_reply":"2023-07-22T06:49:16.735733Z"},"trusted":true},"execution_count":8,"outputs":[{"execution_count":8,"output_type":"execute_result","data":{"text/plain":"array([ 6.9388939e-18, -6.9388939e-18, -1.0408341e-17, ...,\n       -4.5474735e-13, -1.1368684e-13, -7.9580786e-13], dtype=float32)"},"metadata":{}}]},{"cell_type":"code","source":"tokenizer = Wav2Vec2Tokenizer.from_pretrained(\"facebook/wav2vec2-base-960h\")\nmodel = Wav2Vec2ForCTC.from_pretrained(\"facebook/wav2vec2-base-960h\")","metadata":{"execution":{"iopub.status.busy":"2023-07-22T06:53:40.633620Z","iopub.execute_input":"2023-07-22T06:53:40.634003Z","iopub.status.idle":"2023-07-22T06:53:41.694335Z","shell.execute_reply.started":"2023-07-22T06:53:40.633970Z","shell.execute_reply":"2023-07-22T06:53:41.693326Z"},"trusted":true},"execution_count":14,"outputs":[{"name":"stderr","text":"The tokenizer class you load from this checkpoint is not the same type as the class this function is called from. It may result in unexpected tokenization. \nThe tokenizer class you load from this checkpoint is 'Wav2Vec2CTCTokenizer'. \nThe class this function is called from is 'Wav2Vec2Tokenizer'.\nSome weights of Wav2Vec2ForCTC were not initialized from the model checkpoint at facebook/wav2vec2-base-960h and are newly initialized: ['wav2vec2.masked_spec_embed']\nYou should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.\n","output_type":"stream"}]},{"cell_type":"code","source":"input_values = tokenizer(audio, return_tensors = \"pt\").input_values\n# Storing logits (non-normalized prediction values)\nlogits = model(input_values).logits\n\n\n\n# Storing predicted ids\nprediction = torch.argmax(logits, dim = -1)\n\n\n# Passing the prediction to the tokenzer decode to get the transcription\ntranscription = tokenizer.batch_decode(prediction)[0]\n\n# Printing the transcription\nprint(transcription)","metadata":{"execution":{"iopub.status.busy":"2023-07-22T06:53:43.971929Z","iopub.execute_input":"2023-07-22T06:53:43.972325Z","iopub.status.idle":"2023-07-22T06:53:45.722885Z","shell.execute_reply.started":"2023-07-22T06:53:43.972294Z","shell.execute_reply":"2023-07-22T06:53:45.721915Z"},"trusted":true},"execution_count":15,"outputs":[{"name":"stdout","text":"NAT SHISH HOR POR CHACOLICHORIS DRIE ASHONGIBHUDUN KORIN\n","output_type":"stream"}]}]}