Updated model info

README.md

@@ -46,7 +46,9 @@ pip install https://github.com/cheind/autoregressive.git#egg=autoregressive[dev]
 which requires Python 3.9 and a recent PyTorch > 1.9
 
 ## Usage
-The library comes with a set of pre-trained models in `models/`. The following commands use those models to make various predictions. Many commands come with additional parameters, use `--help` to get additional help.
+The library comes with a set of pre-trained models in [`models/`](models/). The following commands use those models to make various predictions. 
+
+Many listed commands come with additional parameters; use `--help` to get additional information.
 
 ### 1D Fourier
 

models/Readme.md

@@ -1,8 +1,17 @@
 # Autoregressive Pretrained Models
 
-The following pre-trained models are provided (R=receptive field, Q=number of quantization channels):
-- `mnist_q2`: WaveNet trained on MNIST with R=699, Q=2
-- `mnist_q256`: WaveNet trained on MNIST with R=699, Q=256
-- `fourier`: WaveNet trained on random Fourier dataset.
+Autoregressive library comes with a set of pre-trained models. Neither of models was carefully tuned. In the following
+ - `R` refers to the models' receptive field,
+ - `Q` to the number of input/output quantization channels.
+
+### Model `mnist_q256`
+WaveNet trained on MNIST. R=699, Q=256, conditioned on the digit class.
+
+### Model `mnist_q2`
+WaveNet trained on binarized MNIST. R=699, Q=2, conditioned on the digit class.
+
+### Model `fseries_q127`
+WaveNet trained on random Fourier series dataset. R=699, Q=127, conditioned on the periodicity of the signal. Training parameters allow periods (integer) between 5-10secs which maps to conditions [0..4].
+
+For usage see [main](../) Readme.
 
-Each model has roughly 4M parameters. Please note that the models have not been tuned with few parameters in mind.