Parler tts compatibility due to last updates (#2278)

Support compatibility with last updates in Parler TTS without support of
static cache.

notebooks/parler-tts-text-to-speech/parler-tts-text-to-speech.ipynb

@@ -350,6 +350,8 @@
  "source": [
  "from collections import namedtuple\n",
  "\n",
+ "import torch.nn as nn\n",
+ "\n",
  "EncoderOutput = namedtuple(\"EncoderOutput\", \"last_hidden_state\")\n",
  "DecoderOutput = namedtuple(\"DecoderOutput\", (\"last_hidden_state\", \"past_key_values\", \"hidden_states\", \"attentions\", \"cross_attentions\"))\n",
  "\n",
@@ -369,10 +371,14 @@
  "\n",
  "\n",
  "class DecoderWrapper(torch.nn.Module):\n",
- " def __init__(self, decoder_stage_1_ir_path, decoder_stage_2_ir_path):\n",
+ " def __init__(self, decoder_stage_1_ir_path, decoder_stage_2_ir_path, config):\n",
  " super().__init__()\n",
  " self.decoder_stage_1 = core.compile_model(decoder_stage_1_ir_path, device.value)\n",
  " self.decoder_stage_2 = core.compile_model(decoder_stage_2_ir_path, device.value)\n",
+ " self.config = config\n",
+ " self.embed_tokens = None\n",
+ " embed_dim = config.vocab_size + 1 # + 1 for pad token id\n",
+ " self.embed_tokens = nn.ModuleList([nn.Embedding(embed_dim, config.hidden_size) for _ in range(config.num_codebooks)])\n",
  "\n",
  " def __call__(self, input_ids=None, encoder_hidden_states=None, encoder_attention_mask=None, past_key_values=None, prompt_hidden_states=None, **kwargs):\n",
  " inputs = {}\n",
@@ -419,7 +425,9 @@
  "outputs": [],
  "source": [
  "model.text_encoder = TextEncoderModelWrapper(TEXT_ENCODER_OV_PATH, model.text_encoder.config)\n",
- "model.decoder.model.decoder = DecoderWrapper(DECODER_STAGE_1_OV_PATH, DECODER_STAGE_2_OV_PATH)"
+ "model.decoder.model.decoder = DecoderWrapper(DECODER_STAGE_1_OV_PATH, DECODER_STAGE_2_OV_PATH, model.decoder.model.decoder.config)\n",
+ "model._supports_cache_class = False\n",
+ "model._supports_static_cache = False"
  ]
  },
  {