Minor change to noise model

Author: tHarvey303

docs/source/noise_modelling/noise_models.ipynb

@@ -25,7 +25,7 @@
     "The simplest noise model is the Depth Noise Model, which adds Gaussian noise with a standard deviation determined by the depth of the observation. This takes a depth (in AB mag), and optionally a depth sigma (default 5.0), which sets the SNR at the given depth. The flux is this noise model is distributed in flux space *not* magnitude space.\n",
     "\n",
     "If you're using this model there are a few things to keep in mind:\n",
-    "- If you are training in magnitude space and conditioning the model on the magnitude errors, they are not well-behaved near the detection limit, as the flux errors become non-Gaussian in magnitude space. The code uses the approximation sigma_mag = 2.5 / ln(10) * (sigma_flux / flux) to estimate the magnitude errors, but this breaks down near the detection limit, and can lead to very large or undefined magnitude errors. Negative fluxes will lead to NaN magnitude errors.\n"
+    "- If you are training in magnitude space and conditioning the model on the magnitude errors, they are not well-behaved near the detection limit, as the flux errors become non-Gaussian in magnitude space. The code uses the approximation sigma_mag = 2.5 / ln(10) * (sigma_flux / flux) to estimate the magnitude errors, but this breaks down near the detection limit, and can lead to very large or undefined magnitude errors. Negative fluxes will lead to NaN magnitude errors, as they can't be representing logarithmically. \n"
    ]
   },
   {
@@ -389,7 +389,11 @@
    ]
   }
  ],
- "metadata": {},
+ "metadata": {
+  "language_info": {
+   "name": "python"
+  }
+ },
  "nbformat": 4,
  "nbformat_minor": 5
 }