from models.modeling_flax_wav2vec2 import FlaxWav2Vec2ForCTC
from transformers import AutoConfig, AutoFeatureExtractor, AutoTokenizer
from flax.traverse_util import flatten_dict, unflatten_dict
import collections

model_id = "sanchit-gandhi/flax-wav2vec2-ctc-cv9-baseline-50k"

config = AutoConfig.from_pretrained(model_id)
config.use_scan = False

unrolled_model = FlaxWav2Vec2ForCTC.from_pretrained(model_id, config=config)
model = FlaxWav2Vec2ForCTC.from_pretrained(model_id)

def scanned_to_unrolled(params):
    new_params = collections.defaultdict(dict)
    # get the key of a scanned module
    for key, stacked_weights in flatten_dict(params['wav2vec2']['encoder']['layers']['FlaxWav2Vec2EncoderLayers']).items():
        for layer, weights in enumerate(stacked_weights):
            new_key = (str(layer),) + key
            new_params[new_key] = weights
            
    new_params = unflatten_dict({('wav2vec2', 'encoder', 'layers') : unflatten_dict(new_params)})
    new_params = flatten_dict(new_params)

    # append parameters for non-scanned modules (i.e. all modules that do not contain the key 'layers')
    for k in flatten_dict(params):
        if 'layers' not in k or 'adapter' in k:
            new_params[k] = flatten_dict(params)[k]

    return unflatten_dict(new_params)

unrolled_model.params = scanned_to_unrolled(model.params)

unrolled_model.save_pretrained("./")

feature_extractor = AutoFeatureExtractor.from_pretrained(model_id)
feature_extractor.save_pretrained("./")
tokenizer = AutoTokenizer.from_pretrained("patrickvonplaten/wav2vec2_ctc_cv9_tokenizer")
tokenizer.save_pretrained("./")