GEE (Unrecognized argument type to convert to a FeatureCollection: {'collection': [<ee.feature.Feature object at 0x7f5acea2a890>)

[ernest19]

Please am trying to playing around with the random forest algorithm which was successful in the code editor, now I try to convert the codes to python and run in colab and having this error which I don't seem to understand.
below is a snapshot of the error and the my script

`
import ee
ee.Initialize()
geometry = ee.Geometry.Polygon([[-2.5130642919845525,9.530810667281315],
[-2.0406521826095525,9.530810667281315],
[-2.0406521826095525,9.958508332761049],
[-2.5130642919845525,9.958508332761049],
[-2.5130642919845525,9.530810667281315]])
# geometry=ee.Feature(ee.Geometry.Rectangle(-2.5130642919845525,9.530810667281315, -2.5130642919845525,9.958508332761049))
# /Loading Sentinel 2 data and filtering by date and bounds
Sentinel2Collection=ee.ImageCollection("COPERNICUS/S2_SR")
.filterDate('2019-01-01','2019-01-28')
.filterBounds(geometry)
.sort('CLOUD_COVER', True)

feat = [

ee.Feature(ee.Geometry.Point(-2.336156720273055,9.617736781866462), {'landcover': 0}),
ee.Feature(ee.Geometry.Point(-2.336156720273055,9.627736781866462), {'landcover': 1}),
ee.Feature(ee.Geometry.Point(-2.4336156720273055,9.627736781866462), {'landcover': 2})
];

Create a FeatureCollection from the list and print it.

forest = ee.FeatureCollection(forest)
urban = ee.FeatureCollection(urban)

vegetation = ee.FeatureCollection(vegetation)

print(fromList)

Filter the collection by the IMAGE_QUALITY property.

filtered = Sentinel2Collection
.filterMetadata('IMAGE_QUALITY', 'equals', 9)

Sentinel_med = Sentinel2Collection.median()
Sentinel_crop = Sentinel_med.clip(geometry)

bands = ['B2', 'B3', 'B4', 'B5', 'B6', 'B7','B8','B8A']

training = Sentinel_crop.select(bands).sampleRegions({
'collection': feat,
'properties':['landcover'],
'scale': 20
})

print(training, 'training')

Making a Random Forest classifier and training it.

classifier= ee.Classifier.smileRandomForest(4).train({
'features': training,
'classProperty': 'landcover',
'inputProperties': bands
})

print(classifier)

#Classifying the input imagery
classification= Sentinel_crop.select(bands).classify(classifier)

Define a palette for the Land Use classification.

palette = [
'cyan', # forest (0) # red
'red', # urban (1) # green
'green' # agriculture (2) # blue
]
print(classification)`
.

[yasasmahima]

Something like this will work...

training_data = image.sampleRegions(collection=training_points,properties=["Property"],scale=30)

[luyao5550]

I also encountered this problem, and finally solved it after checking and trying.
Try this method : add ** before {
classifier= ee.Classifier.smileRandomForest(4).train(
**{'features': training,
'classProperty': 'landcover',
'inputProperties':bands})

[Eors-ui]

Thank you very much!