Loosen PIP requirements (#136)

* Make requirements less strict

* Added clarification on file name, minor corrections

python-sdk/nuscenes/eval/detection/README.md

@@ -5,7 +5,7 @@ In this document we present the rules, result format, classes, evaluation metric
 ## Overview
 - [Introduction](#introduction)
 - [Challenges](#challenges)
-- [General rules](#general-rules)
+- [Submission rules](#submission-rules)
 - [Results format](#results-format)
 - [Classes and attributes](#classes-attributes-and-detection-ranges)
 - [Evaluation metrics](#evaluation-metrics)
@@ -37,6 +37,7 @@ We define a standardized detection result format that serves as an input to the
 The detection results for a particular evaluation set (train/val/test) are stored in a single JSON file. 
 For the train and val sets the evaluation can be performed by the user on their local machine.
 For the test set the user needs to zip the JSON results file and submit it to the official evaluation server.
+The ZIP file and the JSON file must have the exact same name, except for the file extension.
 The JSON file includes meta data `meta` on the type of inputs used for this method.
 Furthermore it includes a dictionary `results` that maps each sample_token to a list of `sample_result` entries.
 Each `sample_token` from the current evaluation set must be included in `results`, although the list of predictions may be empty if no object is detected.

python-sdk/nuscenes/eval/detection/evaluate.py

@@ -71,6 +71,8 @@ def __init__(self,
             os.makedirs(self.plot_dir)
 
         # Load data.
+        if verbose:
+            print('Initializing nuScenes evaluation')
         self.pred_boxes, self.meta = load_prediction(self.result_path, self.cfg.max_boxes_per_sample, verbose=verbose)
         self.gt_boxes = load_gt(self.nusc, self.eval_set, verbose=verbose)
 

python-sdk/nuscenes/nuscenes.py

@@ -57,7 +57,7 @@ def __init__(self,
 
         start_time = time.time()
         if verbose:
-            print("======\nLoading NuScenes tables for version {} ...".format(self.version))
+            print("======\nLoading NuScenes tables for version {}...".format(self.version))
 
         # Explicitly assign tables to help the IDE determine valid class members.
         self.category = self.__load_table__('category')

setup/installation.md

@@ -61,7 +61,7 @@ export PYTHONPATH="${PYTHONPATH}:$HOME/nuscenes-devkit/python-sdk"
 
 To install the required packages, run the following command in your favourite virtual environment:
 ```
-pip install -r setup/requirements.txt
+pip install -r nuscenes-devkit/setup/requirements.txt
 ```
 
 ### Verify install

setup/requirements.txt

@@ -1,11 +1,11 @@
-jupyter==1.0.0
-matplotlib==2.2.3
-numpy==1.14.5
-opencv-python==3.4.2.17
-Pillow==5.2.0
-pyquaternion==0.9.5
-scikit-learn==0.19.2
-tqdm==4.25.0
-scipy==1.1.0
-cachetools==3.1.0
-Shapely==1.6.4.post2
+jupyter
+matplotlib
+numpy
+opencv-python
+Pillow
+pyquaternion>=0.9.5
+scikit-learn
+tqdm
+scipy
+cachetools
+Shapely