add model override argument from load_ensemble_for_inference at generation...

# add the arguments: --pretrained True --pretrained-checkpoint path/to/checkpoint

# Generate:

$ python generate.py data-bin/writingPrompts --path /path/to/trained/model/checkpoint_best.pt --batch-size 32 --beam 1 --sampling --sampling-topk 10 --sampling-temperature 0.8 --nbest 1 

# Note: to load the pretrained model at generation time, you need to pass in a model-override argument to communicate to the fusion model at generation time where you have placed the pretrained checkpoint. By default, it will load the exact path of the fusion model's pretrained model from training time. You should use model-override if you have moved the pretrained model (or are using our provided models). If you are generating from a non-fusion model, the model-override argument is not necessary.

$ python generate.py data-bin/writingPrompts --path /path/to/trained/model/checkpoint_best.pt --batch-size 32 --beam 1 --sampling --sampling-topk 10 --sampling-temperature 0.8 --nbest 1 --model-overrides "{'pretrained_checkpoint':'/path/to/pretrained/model/checkpoint'}"

```

        trained_encoder, trained_decoder = None, None

        pretrained = eval(args.pretrained)

        if pretrained:

            print("| Loading pretrained model")

            print("| loading pretrained model")

            trained_model = utils.load_ensemble_for_inference(

                # not actually for inference, but loads pretrained model parameters

                filenames=[args.pretrained_checkpoint],

                       help='sample from top K likely next words instead of all words')

    group.add_argument('--sampling-temperature', default=1, type=float, metavar='N',

                       help='temperature for random sampling')

    group.add_argument('--model-overrides', default="{}", type=str, metavar='DICT',

                       help='a dictionary used to override model args at generation that were used during model training')

    return group

    # Load ensemble

    print('| loading model(s) from {}'.format(args.path))

    models, _ = utils.load_ensemble_for_inference(args.path.split(':'), task)

    models, _ = utils.load_ensemble_for_inference(args.path.split(':'), task, model_arg_overrides=eval(args.model_overrides))

    # Optimize ensemble for generation

    for model in models: