ولې GradientTape په TensorFlow کې تر ټولو ناڅاپي ځانګړتیا دی

د موادو په څیر اتوماتیک مختلفې او gradients د جوړولو د کمپيوټر gradients ګرامر ټاپونه Gradients په اړه د يو ماډل کنترول چې ټیپ څارنه کوي Intermediate پایلې د فعالیتونو په اړه یادونه د non-scalar هدفونو gradients د جریان کنترول اتوماتیک مختلفې او gradients د ماشین زده کړې algorithms لکه د نوري شبکې روزنې لپاره. د اتوماتیک فرق پیژندل په دې لارښود کې، تاسو به د TensorFlow سره gradients محاسبه کړي، په ځانګړي ډول په لګښت کې. د جوړولو import numpy as np import matplotlib.pyplot as plt import tensorflow as tf 2024-08-15 01:30:07.003169: E external/local_xla/xla/stream_executor/cuda/cuda_fft.cc:485] Unable to register cuFFT factory: Attempting to register factory for plugin cuFFT when one has already been registered 2024-08-15 01:30:07.023862: E external/local_xla/xla/stream_executor/cuda/cuda_dnn.cc:8454] Unable to register cuDNN factory: Attempting to register factory for plugin cuDNN when one has already been registered 2024-08-15 01:30:07.029954: E external/local_xla/xla/stream_executor/cuda/cuda_blas.cc:1452] Unable to register cuBLAS factory: Attempting to register factory for plugin cuBLAS when one has already been registered د کمپيوټر gradients په اتوماتيک ډول مختلفولو لپاره، TensorFlow باید یاد ولرئ چې هغه عملونه چې په دې وخت کې په هغه ترتیب کې ترسره کیږي. پته. بيا، په وخت کې د ، TensorFlow د ګریډینټونه محاسبه کولو لپاره د عملیاتو په مخکښ ترتیب کې د دې لیست کې راټول کیږي. مخونه د پیښو پته ګرامر ټاپونه TensorFlow وړاندې کوي د اتوماتیک فرقولو لپاره API؛ که څه هم، د محاسبه د ځینو انډولونو په اړه د gradient محاسبه کول، معمولا s. TensorFlow "کښته کوي" د اړونده عملیاتونه چې د یو ځای کې ترسره شوي دي TensorFlow بیا د دې ټیپ کاروي ترڅو د "د ریکارډ شوی" محاسبه کولو ګراډینټونه د کارولو سره محاسبه کړي. . tf.GradientTape tf.Variable tf.GradientTape د مختلفو موډلونه دلته یو ساده مثال دی: x = tf.Variable(3.0) with tf.GradientTape() as tape: y = x**2 WARNING: All log messages before absl::InitializeLog() is called are written to STDERR I0000 00:00:1723685409.408818 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685409.412555 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685409.416343 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685409.420087 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685409.431667 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685409.435229 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685409.438777 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685409.442350 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685409.445712 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685409.449141 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685409.452491 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685409.456034 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685410.685265 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685410.687389 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685410.689411 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685410.691490 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685410.693542 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685410.695541 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685410.697441 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685410.699432 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685410.701351 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685410.703333 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685410.705229 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685410.707222 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685410.744994 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685410.747037 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685410.749507 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685410.751538 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685410.753500 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685410.755501 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685410.757421 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685410.759404 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685410.761363 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685410.763858 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685410.766199 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 I0000 00:00:1723685410.768560 20970 cuda_executor.cc:1015] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero. See more at https://github.com/torvalds/linux/blob/v6.0/Documentation/ABI/testing/sysfs-bus-pci#L344-L355 کله چې تاسو ځینې عملیاتو ثبت کړئ، د د ځینې هدفونو (د اغیزمن ډول د خسارت) د ځینې سرچینې (د اغیزمن ډول د نمونې) په اړه د ګراډینټ محاسبه کولو لپاره: GradientTape.gradient(target, sources) # dy = 2x * dx dy_dx = tape.gradient(y, x) dy_dx.numpy() 6.0 په دې مثال کې د scalars کاروي، مګر په هر ټینسر کې په آسانه توګه کار کوي: tf.GradientTape w = tf.Variable(tf.random.normal((3, 2)), name='w') b = tf.Variable(tf.zeros(2, dtype=tf.float32), name='b') x = [[1., 2., 3.]] with tf.GradientTape(persistent=True) as tape: y = x @ w + b loss = tf.reduce_mean(y**2) د gradient څخه ترلاسه کولو لپاره د هر دوو متغیرونو په اړه، تاسو کولی شئ د هر دوو په توګه د سرچینې په توګه د ټیپ په اړه انعطاف دی چې څنګه سرچینو انتقال شوي دي او به د لیستونو یا کلماتو هر نښلیدو ترکیب راټول شي او د ګادینټ په ورته لاره کې جوړ شوی دی (نور) همدارنګه loss gradient tf.nest [dl_dw, dl_db] = tape.gradient(loss, [w, b]) د هر سرچینې په اړه د ګراډینټ د سرچینې شکل لري: print(w.shape) print(dl_dw.shape) (3, 2) (3, 2) دلته د ګراډینټ محاسبه بیا، په دې وخت کې د ګراډینټ محاسبه: my_vars = { 'w': w, 'b': b } grad = tape.gradient(loss, my_vars) grad['b'] Gradients په اړه د يو ماډل دا عادي ده چې د جمع د A یا د دې زیربنا څخه یو ( د ) د او . tf.Variables tf.Module layers.Layer keras.Model د Checkpoint د صادراتو په ډیری صورتونو کې، تاسو به غواړئ د يو ماډل د تمرین وړ متغیرونو په اړه gradients محاسبه کړئ. Since all subclasses of د ګمرکونو په ګډه د د ملکیت، تاسو کولی شئ په ځینې کرښه کوډ کې د دغو gradients محاسبه کړئ: tf.Module Module.trainable_variables layer = tf.keras.layers.Dense(2, activation='relu') x = tf.constant([[1., 2., 3.]]) with tf.GradientTape() as tape: # Forward pass y = layer(x) loss = tf.reduce_mean(y**2) # Calculate gradients with respect to every trainable variable grad = tape.gradient(loss, layer.trainable_variables) for var, g in zip(layer.trainable_variables, grad): print(f'{var.name}, shape: {g.shape}') kernel, shape: (3, 2) bias, shape: (2,) کنترول چې ټیپ څارنه کوي د معياري عمل دی چې د یو trainable ته لاس رسی کولو وروسته ټول فعالیتونه ریکارډ کړي. د دې سببونه دي: tf.Variable د ټیپ ته اړتیا لري چې پوه شي چې هغه عملونه چې په مخکښ پاس کې ریکارډ شي ترڅو په مخکښ پاس کې ګریډینټونه محاسبه کړي. په ټیپ کې د منځني محصولاتو په اړه پیژندل کیږي، نو تاسو غواړئ د غیرقانوني عملونو ریکارډ کړئ. د اغیزمنو د کارولو په صورت کې شامل دي چې د ټولو د يو ماډل د تمرین وړ بدلونونو په اړه د خسارت ګراډینټ محاسبه کوي. د مثال په توګه، په لاندې ډول د gradient محاسبه نلري ځکه چې نه ده "پړاو" په لټه کې، او د نندارتون نه لري: tf.Tensor tf.Variable # A trainable variable x0 = tf.Variable(3.0, name='x0') # Not trainable x1 = tf.Variable(3.0, name='x1', trainable=False) # Not a Variable: A variable + tensor returns a tensor. x2 = tf.Variable(2.0, name='x2') + 1.0 # Not a variable x3 = tf.constant(3.0, name='x3') with tf.GradientTape() as tape: y = (x0**2) + (x1**2) + (x2**2) grad = tape.gradient(y, [x0, x1, x2, x3]) for g in grad: print(g) tf.Tensor(6.0, shape=(), dtype=float32) None None None تاسو کولی شئ د بدلونونو لیست کړئ چې د ټیپ په کارولو سره وګورئ د نمونوي: GradientTape.watched_variables [var.name for var in tape.watched_variables()] ['x0:0'] د کوکونو وړاندې کوي چې د کاروونکي کنترول ورکوي چې څه وګورئ یا نه. tf.GradientTape د ګریډینټونو په اړه د A تاسو باید د تماس : tf.Tensor GradientTape.watch(x) x = tf.constant(3.0) with tf.GradientTape() as tape: tape.watch(x) y = x**2 # dy = 2x * dx dy_dx = tape.gradient(y, x) print(dy_dx.numpy()) 6.0 برعکس، د ټولو وګورئ د معياري عمل د غیر فعالولو لپاره د Set کله چې د gradient ټیپ جوړ کړئ. دا محاسبه د دوو بدلونونو کاروي، مګر یوازې د بدلونونو یو لپاره د gradient سره اړیکه ونیسئ: tf.Variables watch_accessed_variables=False x0 = tf.Variable(0.0) x1 = tf.Variable(10.0) with tf.GradientTape(watch_accessed_variables=False) as tape: tape.watch(x1) y0 = tf.math.sin(x0) y1 = tf.nn.softplus(x1) y = y0 + y1 ys = tf.reduce_sum(y) له خوا نه په تماس کې ، د دې په اړه د gradient نه محاسبه کیږي: GradientTape.watch x0 # dys/dx1 = exp(x1) / (1 + exp(x1)) = sigmoid(x1) grad = tape.gradient(ys, {'x0': x0, 'x1': x1}) print('dy/dx0:', grad['x0']) print('dy/dx1:', grad['x1'].numpy()) dy/dx0: None dy/dx1: 0.9999546 Intermediate پایلې تاسو کولی شئ د پایلو د منځني ارزښتونو په اړه د پایلو ګراډینټونه پوښتنه وکړئ چې په کور کې محاسبه کیږي. په کنټرول کې tf.GradientTape x = tf.constant(3.0) with tf.GradientTape() as tape: tape.watch(x) y = x * x z = y * y # Use the tape to compute the gradient of z with respect to the # intermediate value y. # dz_dy = 2 * y and y = x ** 2 = 9 print(tape.gradient(z, y).numpy()) 18.0 د معياري لخوا، د سرچینې د A له همدې کبله چې د د ډیرو ګراډینټونه په ورته محاسبه کې محاسبه کولو لپاره، د ګراډینټ ټیپ سره . دا اجازه ورکوي چې د ډیری تماسونو ته لکه څنګه چې د سرچینو په توګه په وخت کې چې د ټیپ اجزاو د سپما د جمع شوي دي. د مثال په توګه: GradientTape GradientTape.gradient persistent=True gradient x = tf.constant([1, 3.0]) with tf.GradientTape(persistent=True) as tape: tape.watch(x) y = x * x z = y * y print(tape.gradient(z, x).numpy()) # [4.0, 108.0] (4 * x**3 at x = [1.0, 3.0]) print(tape.gradient(y, x).numpy()) # [2.0, 6.0] (2 * x at x = [1.0, 3.0]) [ 4. 108.] [2. 6.] del tape # Drop the reference to the tape د فعالیتونو په اړه یادونه There is a tiny overhead associated with doing operations inside a gradient tape context. For most eager execution this will not be a noticeable cost, but you should still use tape context around the areas only where it is required. Gradient tapes use memory to store intermediate results, including inputs and outputs, for use during the backwards pass. For efficiency, some ops (like ) don't need to keep their intermediate results and they are pruned during the forward pass. However, if you use on your tape, and your peak memory usage will be higher. ReLU persistent=True nothing is discarded د non-scalar هدفونو gradients د gradient په اصل کې د scalar په اړه یو عمل دی. x = tf.Variable(2.0) with tf.GradientTape(persistent=True) as tape: y0 = x**2 y1 = 1 / x print(tape.gradient(y0, x).numpy()) print(tape.gradient(y1, x).numpy()) 4.0 -0.25 په دې توګه، که تاسو د څو هدفونو د ګراډینټ لپاره پوښتنه کوئ، د هر سرچینې لپاره پایله ده: د هدفونو په شمول د gradient، یا برابر د هر هدف د ګراډینټونو شمیره. x = tf.Variable(2.0) with tf.GradientTape() as tape: y0 = x**2 y1 = 1 / x print(tape.gradient({'y0': y0, 'y1': y1}, x).numpy()) 3.75 همدارنګه، که هدف(s) نه scalar دي، د مجموعې gradient محاسبه کیږي: x = tf.Variable(2.) with tf.GradientTape() as tape: y = x * [3., 4.] print(tape.gradient(y, x).numpy()) 7.0 دا اسانه کوي چې د کټګورۍ د کټګورۍ د کټګورۍ په شمول د gradient، یا د element-wise کټګورۍ په شمول د gradient ترلاسه کړي. که تاسو د هر موضوع لپاره د انفرادي gradient ته اړتیا لرئ، د . د جاکوبین په ځینو حالتونو کې، تاسو کولی شئ د جاکوبینګ له لاسه ورکړئ. د عنصر ډول محاسبه کولو لپاره، د شمیره ګریډینټ د هر عنصر د انټرنټ عنصر په اړه د هر عنصر د منلو ورکړئ، ځکه چې هر عنصر مستقل دی: x = tf.linspace(-10.0, 10.0, 200+1) with tf.GradientTape() as tape: tape.watch(x) y = tf.nn.sigmoid(x) dy_dx = tape.gradient(y, x) plt.plot(x, y, label='y') plt.plot(x, dy_dx, label='dy/dx') plt.legend() _ = plt.xlabel('x') د جریان کنترول ځکه چې د ګراډینټ ټیپ د عملیاتو ریکارډ کوي لکه څنګه چې دوی ترسره کیږي، د پیټون کنترول د جریان په طبيعي توګه کارول کیږي (د مثال په توګه، او د اعلانونو). if while دلته د یو مختلف متغیر په هر زون کې کارول کیږي د ګراډینټ یوازې د ګراډینټ سره تړاو لري چې کارول شوی: if x = tf.constant(1.0) v0 = tf.Variable(2.0) v1 = tf.Variable(2.0) with tf.GradientTape(persistent=True) as tape: tape.watch(x) if x > 0.0: result = v0 else: result = v1**2 dv0, dv1 = tape.gradient(result, [v0, v1]) print(dv0) print(dv1) tf.Tensor(1.0, shape=(), dtype=float32) None یوازې په یاد ولرئ چې د کنترول بیاناتو په ځان کې د مختلفو نه دي، نو دوی د gradient-based optimizers لپاره غیرقانوني دي. پر بنسټ د ارزښت په دې مثال کې، د ټیپ یا ریکارډونه یا . د gradient په اړه همدارنګه . x result = v0 result = v1**2 x None dx = tape.gradient(result, x) print(dx) None په اصل کې په TensorFlow ویب پاڼه کې چاپ شوی، دا مقاله دلته د نوي نوم لاندې ښودل کیږي او د CC BY 4.0 لسیس شوی دی. د کوډ نمونې د Apache 2.0 لسیس لاندې شریک شوي دي. په اصل کې په TensorFlow ویب پاڼه کې چاپ شوی، دا مقاله دلته د نوي نوم لاندې ښودل کیږي او د CC BY 4.0 لسیس شوی دی. د کوډ نمونې د Apache 2.0 لسیس لاندې شریک شوي دي. د TensorFlow د TensorFlow