diff --git a/docs/src/examples/complex.md b/docs/src/examples/complex.md
index a11a4d4d0f..ff9f1339a5 100644
--- a/docs/src/examples/complex.md
+++ b/docs/src/examples/complex.md
@@ -44,7 +44,7 @@ alg = NNODE(chain, opt, ps; strategy = StochasticTraining(500))
 sol = solve(problem, alg, verbose = false, maxiters = 5000, saveat = 0.01)
 ```
 
-Now, lets plot the predictions.
+Now, let's plot the predictions.
 
 `u1`:
 
diff --git a/docs/src/tutorials/Lotka_Volterra_BPINNs.md b/docs/src/tutorials/Lotka_Volterra_BPINNs.md
index b21d6685e5..a8a2bb0eb3 100644
--- a/docs/src/tutorials/Lotka_Volterra_BPINNs.md
+++ b/docs/src/tutorials/Lotka_Volterra_BPINNs.md
@@ -66,7 +66,7 @@ plot!(time, y, label = "noisy y")
 plot!(solution, labels = ["x" "y"])
 ```
 
-Lets define a PINN.
+Let's define a PINN.
 
 ```@example bpinn
 # Neural Networks must have 2 outputs as u -> [dx,dy] in function lotka_volterra()
diff --git a/docs/src/tutorials/dae.md b/docs/src/tutorials/dae.md
index a40d9c58d8..1f468caedd 100644
--- a/docs/src/tutorials/dae.md
+++ b/docs/src/tutorials/dae.md
@@ -37,7 +37,7 @@ alg = NNDAE(chain, opt; autodiff = false)
 sol = solve(prob, alg, verbose = true, dt = 1 / 100.0, maxiters = 3000, abstol = 1e-10)
 ```
 
-Now lets compare the predictions from the learned network with the ground truth which we can obtain by numerically solving the DAE.
+Now let's compare the predictions from the learned network with the ground truth which we can obtain by numerically solving the DAE.
 
 ```@example dae
 function example1(du, u, p, t)
diff --git a/docs/src/tutorials/ode.md b/docs/src/tutorials/ode.md
index 5b1afdfd2a..c78341c79e 100644
--- a/docs/src/tutorials/ode.md
+++ b/docs/src/tutorials/ode.md
@@ -61,7 +61,7 @@ Once these pieces are together, we call `solve` just like with any other `ODEPro
 sol = solve(prob, alg, verbose = true, maxiters = 2000, saveat = 0.01)
 ```
 
-Now lets compare the predictions from the learned network with the ground truth which we can obtain by numerically solving the ODE.
+Now let's compare the predictions from the learned network with the ground truth which we can obtain by numerically solving the ODE.
 
 ```@example nnode1
 using OrdinaryDiffEq, Plots
diff --git a/docs/src/tutorials/ode_parameter_estimation.md b/docs/src/tutorials/ode_parameter_estimation.md
index 93b1c99eca..c8243e9587 100644
--- a/docs/src/tutorials/ode_parameter_estimation.md
+++ b/docs/src/tutorials/ode_parameter_estimation.md
@@ -26,7 +26,7 @@ u0 = [5.0, 5.0]
 prob = ODEProblem(lv, u0, tspan, [1.0, 1.0, 1.0, 1.0])
 ```
 
-As we want to estimate the parameters as well, lets get some data.
+As we want to estimate the parameters as well, let's get some data.
 
 ```@example param_estim_lv
 true_p = [1.5, 1.0, 3.0, 1.0]
@@ -36,7 +36,7 @@ t_ = sol_data.t
 u_ = reduce(hcat, sol_data.u)
 ```
 
-Now, lets define a neural network for the PINN using [Lux.jl](https://lux.csail.mit.edu/).
+Now, let's define a neural network for the PINN using [Lux.jl](https://lux.csail.mit.edu/).
 
 ```@example param_estim_lv
 rng = Random.default_rng()
@@ -81,7 +81,7 @@ plot(sol, labels = ["u1_pinn" "u2_pinn"])
 plot!(sol_data, labels = ["u1_data" "u2_data"])
 ```
 
-We can see it is a good fit! Now lets see if we have the parameters of the equation also estimated correctly or not.
+We can see it is a good fit! Now let's see if we have the parameters of the equation also estimated correctly or not.
 
 ```@example param_estim_lv
 sol.k.u.p
diff --git a/src/training_strategies.jl b/src/training_strategies.jl
index 68aa9cd71f..858e93a237 100644
--- a/src/training_strategies.jl
+++ b/src/training_strategies.jl
@@ -160,7 +160,7 @@ that accelerate the convergence in high dimensional spaces over pure random sequ
 * `sampling_alg`: the quasi-Monte Carlo sampling algorithm,
 * `resampling`: if it's false - the full training set is generated in advance before training,
    and at each iteration, one subset is randomly selected out of the batch.
-   Ff it's true - the training set isn't generated beforehand, and one set of quasi-random
+   If it's true - the training set isn't generated beforehand, and one set of quasi-random
    points is generated directly at each iteration in runtime. In this case, `minibatch` has no effect,
 * `minibatch`: the number of subsets, if resampling == false.