diff --git a/ASR/wav2vec2 TRT inference.ipynb b/ASR/wav2vec2 TRT inference.ipynb index fb546de..02aac94 100644 --- a/ASR/wav2vec2 TRT inference.ipynb +++ b/ASR/wav2vec2 TRT inference.ipynb @@ -2,12 +2,12 @@ "cells": [ { "cell_type": "code", - "execution_count": 1, + "execution_count": 2, "metadata": { "collapsed": false, "ExecuteTime": { - "end_time": "2023-05-28T10:58:35.369155100Z", - "start_time": "2023-05-28T10:58:33.253155400Z" + "end_time": "2023-06-02T14:22:01.393938100Z", + "start_time": "2023-06-02T14:21:59.704398700Z" } }, "outputs": [], @@ -29,8 +29,8 @@ "metadata": { "collapsed": false, "ExecuteTime": { - "end_time": "2023-05-28T10:58:42.538089900Z", - "start_time": "2023-05-28T10:58:40.178091100Z" + "end_time": "2023-06-02T14:22:06.515852100Z", + "start_time": "2023-06-02T14:22:01.988119500Z" } }, "outputs": [], @@ -41,12 +41,12 @@ }, { "cell_type": "code", - "execution_count": 7, + "execution_count": 4, "metadata": { "collapsed": false, "ExecuteTime": { - "end_time": "2023-05-28T10:59:29.613384100Z", - "start_time": "2023-05-28T10:59:29.592385500Z" + "end_time": "2023-06-02T14:22:07.261043Z", + "start_time": "2023-06-02T14:22:06.515852100Z" } }, "outputs": [], @@ -58,7 +58,7 @@ }, { "cell_type": "code", - "execution_count": null, + "execution_count": 5, "outputs": [], "source": [ "def clean(annotation):\n", @@ -69,31 +69,35 @@ " return annotation" ], "metadata": { - "collapsed": false + "collapsed": false, + "ExecuteTime": { + "end_time": "2023-06-02T14:22:10.504772200Z", + "start_time": "2023-06-02T14:22:10.487802100Z" + } } }, { "cell_type": "code", - "execution_count": 17, + "execution_count": 11, "metadata": { "collapsed": false, "ExecuteTime": { - "end_time": "2023-05-28T11:01:16.791265400Z", - "start_time": "2023-05-28T11:01:16.537264900Z" + "end_time": "2023-06-02T14:22:35.081080Z", + "start_time": "2023-06-02T14:22:34.853064100Z" } }, "outputs": [ { "data": { - "text/plain": "'SHE SCORED NINE FOR THE QUIZ HELP FROM HER FRIENDS'" + "text/plain": "'THE BATTERY LEFT ON THIS FHOND IS ZERO PERCENT TO AT THEIR JOBS'" }, - "execution_count": 17, + "execution_count": 11, "metadata": {}, "output_type": "execute_result" } ], "source": [ - "audio, sr = torchaudio.load('train/audio/train_00001.wav')\n", + "audio, sr = torchaudio.load('test/audio/evala_09977.wav')\n", "audio = processor(audio, sampling_rate=16000).input_values[0][0]\n", "audio = np.expand_dims(audio, axis=0)\n", "output = model({'input': audio})['output']\n", diff --git a/CV/SOLIDER-REID/TIL.yml b/CV/SOLIDER-REID/TIL.yml index 636e7d5..a7b1d50 100644 --- a/CV/SOLIDER-REID/TIL.yml +++ b/CV/SOLIDER-REID/TIL.yml @@ -35,7 +35,7 @@ SOLVER: # BASE_LR: 0.005 # 0.0008 BASE_LR: 5e-3 WARMUP_EPOCHS: 3 - IMS_PER_BATCH: 128 + IMS_PER_BATCH: 64 WARMUP_METHOD: 'cosine' LARGE_FC_LR: False CHECKPOINT_PERIOD: 1 diff --git a/requirements.txt b/requirements.txt index 63e54cb..6d44f31 100644 --- a/requirements.txt +++ b/requirements.txt @@ -32,5 +32,4 @@ timm mmcv-full yacs language_tool_python -ensemble-boxes audiomentations