update readme

README.md

@@ -116,13 +116,26 @@ automation. This dataset is generated by applying forced alignment on crowd-sour
 audio to produce per-word timing estimates for extraction.
 All alignments are included in the dataset.
 
-Data is provided in two formats: `wav` (16KHz) and `opus` (48KHz). To download a specific format pass it to the `format`
-argument (default format is `wav`):
+Data is provided in two formats: `wav` (16KHz) and `opus` (48KHz). Default configurations look like 
+`"{lang}_{format}"`, so to load, for example, Tatar in wav format do:
 
 ```python
-ds = load_dataset("polinaeterna/ml_spoken_words", languages="tt", format="opus")
+ds = load_dataset("polinaeterna/ml_spoken_words", "tt_wav")
 ```
 
+To download multiple languages in a single dataset pass list of languages to `languages` argument:
+```python
+ds = load_dataset("polinaeterna/ml_spoken_words", languages=["ar", "tt", "br"])
+```
+
+To download a specific format pass it to the `format` argument (default format is `wav`):
+```python
+ds = load_dataset("polinaeterna/ml_spoken_words", languages=["ar", "tt", "br"], format="opus")
+```
+Note that each time you provide different sets of languages, 
+examples are generated from scratch even if you already provided one or several of them before 
+because custom configurations are created each time (the data is **not** redownloaded though).
+
 ### Supported Tasks and Leaderboards
 
 Keyword spotting, Spoken term search
@@ -213,7 +226,7 @@ Hig-resourced (>100 hours):
 
 ### Data Fields
 
-* file: strinrelative audio path inside the archive **#TODO: change according to the new local path schema?**
+* file: strinrelative audio path inside the archive
 * is_valid: if a sample is valid
 * language: language of an instance. Makes sense only when providing multiple languages to the 
 dataset loader (for example, `load_dataset("ml_spoken_words", languages=["ar", "tt"])`)