Upload Transformers files (#10)

- 0239ea0ed29c2e7efd73760d1a83bd7131243fb79cfcfbe6441ee00ea63a60d6 (c2496a537eff29277079f431128c748f2e82ddd2)
- 7e798c0f9215fd0587d45f7f7eebe375fa8dedd010053abca86960e842a24f93 (af7abdfe8be088cc14716b7cf3767d17c2aff350)
- 2e90982aba999d4786d9c627984626547ba6be8d7fc5272dd61996204981e199 (6ffcd0b0db10ee3502a5100e372a97332bea7850)
- 60a1fb4b40bb99fa9522269df7b34193e86f116bf016d525fedfca8132e48368 (f96d3e47d5a891033219b6611188f758ad6c0be0)
- eff96175fbebbb02b3c84aedc7038964efc508ef01ac14c6efa162aabb2af4f8 (5a7e572d823b17800ee15cdec1adcea536e37d9e)
- d5e64de3c4eeaa1a3dcd981f86a908b4db261d3036bca21ac4cf50d5bee59e77 (093f8c0c5aa9acc46c2a29dd920c2213d061578a)
- 79b499dfdb39f8b5086eafe13266d9eb93161371d407d0f09de77f318906c364 (07806a5b7b69f61f0a98b4e5191cfe28a9d28583)
- 9bc05ea4242c6d95640d5bcabaa576132f37012154f90a92c74890d15c6ef31d (ad3dd0eba489f1d66941358b563ec35db2d4760a)
- d6efae76c8767a8aaf68096607be0cf3a0557cfa2d3e48588ba5e47d36d69a98 (3da21df00389d39068a0df1f333a698b2607bc8b)
- 8f122d446944da12464e5def89fb400d83493546e7ccc119e9a42a15b7d1b232 (f4059c1c184fa6d604a3cbf1d1cfeabe60126be6)
- 7f97733f269c5fb3bc3a6356c7fd27718089b794ea86bd41106ce2e43ef16783 (082e1e45e515efd0f5f1f40196a275abed876235)
- f8eecb893de1103f3da5df2e92188e559c7105748b9acdf214f0ae746915c33f (762750547ba9416746dfb6f5897f7f821dfe2f90)
- 2a20f92b81001cf1c33ab21234f778799dd8e96a80daec2d12425a426fd4f24f (0b46e42a57f2c2bed6381cb7b5715d1bc41af989)
- 37439981bf15b4b72e5fdf4fec28ac261a08041eab8baf697c6b7da002782c5f (1e4d67a017e3aadd828ae3bc2f9c769a230b46bd)
- 2c90bbf864b575a299ce303355c3ff8aaf32f27505b58e7bca8ad3c131a44eab (7e985d10cd88e7900b88c41408b194835826cf4b)
- be5b3bcb480712fc236fed661b3541bff3adc689b104484ff5cf6870252994bd (e52a785d5590941b5b6ad07148f02f849db2590d)
- 448f9ad56731748a537d3d42ed08cfb72fd0fb585ffb6edbd21cb979b7e1e71e (5feae551b0298b2bb3118896deebe3de5d3e5285)


Co-authored-by: Yoach Lacombe <ylacombe@users.noreply.huggingface.co>

config.json

@@ -0,0 +1,340 @@
+{
+  "_commit_hash": null,
+  "architectures": [
+    "BarkModel"
+  ],
+  "coarse_acoustics_config": {
+    "_name_or_path": "",
+    "add_cross_attention": false,
+    "architectures": [
+      "BarkCoarseModel"
+    ],
+    "bad_words_ids": null,
+    "begin_suppress_tokens": null,
+    "bias": false,
+    "block_size": 1024,
+    "bos_token_id": null,
+    "chunk_size_feed_forward": 0,
+    "cross_attention_hidden_size": null,
+    "decoder_start_token_id": null,
+    "diversity_penalty": 0.0,
+    "do_sample": false,
+    "dropout": 0.0,
+    "early_stopping": false,
+    "encoder_no_repeat_ngram_size": 0,
+    "eos_token_id": null,
+    "exponential_decay_length_penalty": null,
+    "finetuning_task": null,
+    "forced_bos_token_id": null,
+    "forced_eos_token_id": null,
+    "hidden_size": 1024,
+    "id2label": {
+      "0": "LABEL_0",
+      "1": "LABEL_1"
+    },
+    "initializer_range": 0.02,
+    "input_vocab_size": 12096,
+    "is_decoder": false,
+    "is_encoder_decoder": false,
+    "label2id": {
+      "LABEL_0": 0,
+      "LABEL_1": 1
+    },
+    "length_penalty": 1.0,
+    "max_length": 20,
+    "min_length": 0,
+    "model_type": "coarse_acoustics",
+    "no_repeat_ngram_size": 0,
+    "num_beam_groups": 1,
+    "num_beams": 1,
+    "num_heads": 16,
+    "num_layers": 24,
+    "num_return_sequences": 1,
+    "output_attentions": false,
+    "output_hidden_states": false,
+    "output_scores": false,
+    "output_vocab_size": 12096,
+    "pad_token_id": null,
+    "prefix": null,
+    "problem_type": null,
+    "pruned_heads": {},
+    "remove_invalid_values": false,
+    "repetition_penalty": 1.0,
+    "return_dict": true,
+    "return_dict_in_generate": false,
+    "sep_token_id": null,
+    "suppress_tokens": null,
+    "task_specific_params": null,
+    "temperature": 1.0,
+    "tf_legacy_loss": false,
+    "tie_encoder_decoder": false,
+    "tie_word_embeddings": true,
+    "tokenizer_class": null,
+    "top_k": 50,
+    "top_p": 1.0,
+    "torch_dtype": "float32",
+    "torchscript": false,
+    "transformers_version": "4.31.0.dev0",
+    "typical_p": 1.0,
+    "use_bfloat16": false,
+    "use_cache": true
+  },
+  "codec_config": {
+    "_name_or_path": "ArthurZ/encodec_24khz",
+    "add_cross_attention": false,
+    "architectures": [
+      "EncodecModel"
+    ],
+    "audio_channels": 1,
+    "bad_words_ids": null,
+    "begin_suppress_tokens": null,
+    "bos_token_id": null,
+    "chunk_length_s": null,
+    "chunk_size_feed_forward": 0,
+    "codebook_dim": 128,
+    "codebook_size": 1024,
+    "compress": 2,
+    "cross_attention_hidden_size": null,
+    "decoder_start_token_id": null,
+    "dilation_growth_rate": 2,
+    "diversity_penalty": 0.0,
+    "do_sample": false,
+    "early_stopping": false,
+    "encoder_no_repeat_ngram_size": 0,
+    "eos_token_id": null,
+    "exponential_decay_length_penalty": null,
+    "finetuning_task": null,
+    "forced_bos_token_id": null,
+    "forced_eos_token_id": null,
+    "hidden_size": 128,
+    "id2label": {
+      "0": "LABEL_0",
+      "1": "LABEL_1"
+    },
+    "is_decoder": false,
+    "is_encoder_decoder": false,
+    "kernel_size": 7,
+    "label2id": {
+      "LABEL_0": 0,
+      "LABEL_1": 1
+    },
+    "last_kernel_size": 7,
+    "length_penalty": 1.0,
+    "max_length": 20,
+    "min_length": 0,
+    "model_type": "encodec",
+    "no_repeat_ngram_size": 0,
+    "norm_type": "weight_norm",
+    "normalize": false,
+    "num_beam_groups": 1,
+    "num_beams": 1,
+    "num_filters": 32,
+    "num_lstm_layers": 2,
+    "num_residual_layers": 1,
+    "num_return_sequences": 1,
+    "output_attentions": false,
+    "output_hidden_states": false,
+    "output_scores": false,
+    "overlap": null,
+    "pad_mode": "reflect",
+    "pad_token_id": null,
+    "prefix": null,
+    "problem_type": null,
+    "pruned_heads": {},
+    "remove_invalid_values": false,
+    "repetition_penalty": 1.0,
+    "residual_kernel_size": 3,
+    "return_dict": true,
+    "return_dict_in_generate": false,
+    "sampling_rate": 24000,
+    "sep_token_id": null,
+    "suppress_tokens": null,
+    "target_bandwidths": [
+      1.5,
+      3.0,
+      6.0,
+      12.0,
+      24.0
+    ],
+    "task_specific_params": null,
+    "temperature": 1.0,
+    "tf_legacy_loss": false,
+    "tie_encoder_decoder": false,
+    "tie_word_embeddings": true,
+    "tokenizer_class": null,
+    "top_k": 50,
+    "top_p": 1.0,
+    "torch_dtype": "float32",
+    "torchscript": false,
+    "transformers_version": "4.31.0.dev0",
+    "trim_right_ratio": 1.0,
+    "typical_p": 1.0,
+    "upsampling_ratios": [
+      8,
+      5,
+      4,
+      2
+    ],
+    "use_bfloat16": false,
+    "use_causal_conv": true,
+    "use_conv_shortcut": true
+  },
+  "fine_acoustics_config": {
+    "_name_or_path": "",
+    "add_cross_attention": false,
+    "architectures": [
+      "BarkFineModel"
+    ],
+    "bad_words_ids": null,
+    "begin_suppress_tokens": null,
+    "bias": false,
+    "block_size": 1024,
+    "bos_token_id": null,
+    "chunk_size_feed_forward": 0,
+    "cross_attention_hidden_size": null,
+    "decoder_start_token_id": null,
+    "diversity_penalty": 0.0,
+    "do_sample": false,
+    "dropout": 0.0,
+    "early_stopping": false,
+    "encoder_no_repeat_ngram_size": 0,
+    "eos_token_id": null,
+    "exponential_decay_length_penalty": null,
+    "finetuning_task": null,
+    "forced_bos_token_id": null,
+    "forced_eos_token_id": null,
+    "hidden_size": 1024,
+    "id2label": {
+      "0": "LABEL_0",
+      "1": "LABEL_1"
+    },
+    "initializer_range": 0.02,
+    "input_vocab_size": 1056,
+    "is_decoder": false,
+    "is_encoder_decoder": false,
+    "label2id": {
+      "LABEL_0": 0,
+      "LABEL_1": 1
+    },
+    "length_penalty": 1.0,
+    "max_length": 20,
+    "min_length": 0,
+    "model_type": "fine_acoustics",
+    "n_codes_given": 1,
+    "n_codes_total": 8,
+    "no_repeat_ngram_size": 0,
+    "num_beam_groups": 1,
+    "num_beams": 1,
+    "num_heads": 16,
+    "num_layers": 24,
+    "num_return_sequences": 1,
+    "output_attentions": false,
+    "output_hidden_states": false,
+    "output_scores": false,
+    "output_vocab_size": 1056,
+    "pad_token_id": null,
+    "prefix": null,
+    "problem_type": null,
+    "pruned_heads": {},
+    "remove_invalid_values": false,
+    "repetition_penalty": 1.0,
+    "return_dict": true,
+    "return_dict_in_generate": false,
+    "sep_token_id": null,
+    "suppress_tokens": null,
+    "task_specific_params": null,
+    "temperature": 1.0,
+    "tf_legacy_loss": false,
+    "tie_encoder_decoder": false,
+    "tie_word_embeddings": true,
+    "tokenizer_class": null,
+    "top_k": 50,
+    "top_p": 1.0,
+    "torch_dtype": "float32",
+    "torchscript": false,
+    "transformers_version": "4.31.0.dev0",
+    "typical_p": 1.0,
+    "use_bfloat16": false,
+    "use_cache": true
+  },
+  "initializer_range": 0.02,
+  "model_type": "bark",
+  "semantic_config": {
+    "_name_or_path": "",
+    "add_cross_attention": false,
+    "architectures": [
+      "BarkSemanticModel"
+    ],
+    "bad_words_ids": null,
+    "begin_suppress_tokens": null,
+    "bias": false,
+    "block_size": 1024,
+    "bos_token_id": null,
+    "chunk_size_feed_forward": 0,
+    "cross_attention_hidden_size": null,
+    "decoder_start_token_id": null,
+    "diversity_penalty": 0.0,
+    "do_sample": false,
+    "dropout": 0.0,
+    "early_stopping": false,
+    "encoder_no_repeat_ngram_size": 0,
+    "eos_token_id": null,
+    "exponential_decay_length_penalty": null,
+    "finetuning_task": null,
+    "forced_bos_token_id": null,
+    "forced_eos_token_id": null,
+    "hidden_size": 1024,
+    "id2label": {
+      "0": "LABEL_0",
+      "1": "LABEL_1"
+    },
+    "initializer_range": 0.02,
+    "input_vocab_size": 129600,
+    "is_decoder": false,
+    "is_encoder_decoder": false,
+    "label2id": {
+      "LABEL_0": 0,
+      "LABEL_1": 1
+    },
+    "length_penalty": 1.0,
+    "max_length": 20,
+    "min_length": 0,
+    "model_type": "semantic",
+    "no_repeat_ngram_size": 0,
+    "num_beam_groups": 1,
+    "num_beams": 1,
+    "num_heads": 16,
+    "num_layers": 24,
+    "num_return_sequences": 1,
+    "output_attentions": false,
+    "output_hidden_states": false,
+    "output_scores": false,
+    "output_vocab_size": 10048,
+    "pad_token_id": null,
+    "prefix": null,
+    "problem_type": null,
+    "pruned_heads": {},
+    "remove_invalid_values": false,
+    "repetition_penalty": 1.0,
+    "return_dict": true,
+    "return_dict_in_generate": false,
+    "sep_token_id": null,
+    "suppress_tokens": null,
+    "task_specific_params": null,
+    "temperature": 1.0,
+    "tf_legacy_loss": false,
+    "tie_encoder_decoder": false,
+    "tie_word_embeddings": true,
+    "tokenizer_class": null,
+    "top_k": 50,
+    "top_p": 1.0,
+    "torch_dtype": "float32",
+    "torchscript": false,
+    "transformers_version": "4.31.0.dev0",
+    "typical_p": 1.0,
+    "use_bfloat16": false,
+    "use_cache": true
+  },
+  "torch_dtype": "float32",
+  "transformers_version": null
+}

generation_config.json

@@ -0,0 +1,172 @@
+{
+  "coarse_acoustics_config": {
+    "_from_model_config": false,
+    "bad_words_ids": null,
+    "begin_suppress_tokens": null,
+    "bos_token_id": null,
+    "coarse_infer_token": 12050,
+    "coarse_rate_hz": 75,
+    "coarse_semantic_pad_token": 12048,
+    "constraints": null,
+    "decoder_start_token_id": null,
+    "diversity_penalty": 0.0,
+    "do_sample": true,
+    "early_stopping": false,
+    "encoder_no_repeat_ngram_size": 0,
+    "encoder_repetition_penalty": 1.0,
+    "eos_token_id": null,
+    "epsilon_cutoff": 0.0,
+    "eta_cutoff": 0.0,
+    "exponential_decay_length_penalty": null,
+    "force_words_ids": null,
+    "forced_bos_token_id": null,
+    "forced_decoder_ids": null,
+    "forced_eos_token_id": null,
+    "generation_kwargs": {},
+    "guidance_scale": null,
+    "length_penalty": 1.0,
+    "max_coarse_history": 630,
+    "max_coarse_input_length": 256,
+    "max_length": 20,
+    "max_new_tokens": null,
+    "max_time": null,
+    "min_length": 0,
+    "min_new_tokens": null,
+    "n_coarse_codebooks": 2,
+    "no_repeat_ngram_size": 0,
+    "num_beam_groups": 1,
+    "num_beams": 1,
+    "num_return_sequences": 1,
+    "output_attentions": false,
+    "output_hidden_states": false,
+    "output_scores": false,
+    "pad_token_id": null,
+    "penalty_alpha": null,
+    "remove_invalid_values": false,
+    "renormalize_logits": true,
+    "repetition_penalty": 1.0,
+    "return_dict_in_generate": false,
+    "sequence_bias": null,
+    "sliding_window_len": 60,
+    "suppress_tokens": null,
+    "temperature": 0.7,
+    "top_k": 50,
+    "top_p": 1.0,
+    "transformers_version": "4.31.0.dev0",
+    "typical_p": 1.0,
+    "use_cache": true
+  },
+  "codebook_size": 1024,
+  "fine_acoustics_config": {
+    "_from_model_config": false,
+    "bad_words_ids": null,
+    "begin_suppress_tokens": null,
+    "bos_token_id": null,
+    "constraints": null,
+    "decoder_start_token_id": null,
+    "diversity_penalty": 0.0,
+    "do_sample": false,
+    "early_stopping": false,
+    "encoder_no_repeat_ngram_size": 0,
+    "encoder_repetition_penalty": 1.0,
+    "eos_token_id": null,
+    "epsilon_cutoff": 0.0,
+    "eta_cutoff": 0.0,
+    "exponential_decay_length_penalty": null,
+    "force_words_ids": null,
+    "forced_bos_token_id": null,
+    "forced_decoder_ids": null,
+    "forced_eos_token_id": null,
+    "generation_kwargs": {},
+    "guidance_scale": null,
+    "length_penalty": 1.0,
+    "max_fine_history_length": 512,
+    "max_fine_input_length": 1024,
+    "max_length": 20,
+    "max_new_tokens": null,
+    "max_time": null,
+    "min_length": 0,
+    "min_new_tokens": null,
+    "n_fine_codebooks": 8,
+    "no_repeat_ngram_size": 0,
+    "num_beam_groups": 1,
+    "num_beams": 1,
+    "num_return_sequences": 1,
+    "output_attentions": false,
+    "output_hidden_states": false,
+    "output_scores": false,
+    "pad_token_id": null,
+    "penalty_alpha": null,
+    "remove_invalid_values": false,
+    "renormalize_logits": false,
+    "repetition_penalty": 1.0,
+    "return_dict_in_generate": false,
+    "sequence_bias": null,
+    "suppress_tokens": null,
+    "temperature": 0.5,
+    "top_k": 50,
+    "top_p": 1.0,
+    "transformers_version": "4.31.0.dev0",
+    "typical_p": 1.0,
+    "use_cache": true
+  },
+  "model_type": "bark",
+  "sample_rate": 24000,
+  "semantic_config": {
+    "_from_model_config": false,
+    "bad_words_ids": null,
+    "begin_suppress_tokens": null,
+    "bos_token_id": null,
+    "constraints": null,
+    "decoder_start_token_id": null,
+    "diversity_penalty": 0.0,
+    "do_sample": true,
+    "early_stopping": false,
+    "encoder_no_repeat_ngram_size": 0,
+    "encoder_repetition_penalty": 1.0,
+    "eos_token_id": 10000,
+    "epsilon_cutoff": 0.0,
+    "eta_cutoff": 0.0,
+    "exponential_decay_length_penalty": null,
+    "force_words_ids": null,
+    "forced_bos_token_id": null,
+    "forced_decoder_ids": null,
+    "forced_eos_token_id": null,
+    "generation_kwargs": {},
+    "guidance_scale": null,
+    "length_penalty": 1.0,
+    "max_input_semantic_length": 256,
+    "max_length": 20,
+    "max_new_tokens": 768,
+    "max_time": null,
+    "min_length": 0,
+    "min_new_tokens": null,
+    "no_repeat_ngram_size": 0,
+    "num_beam_groups": 1,
+    "num_beams": 1,
+    "num_return_sequences": 1,
+    "output_attentions": false,
+    "output_hidden_states": false,
+    "output_scores": false,
+    "pad_token_id": null,
+    "penalty_alpha": null,
+    "remove_invalid_values": false,
+    "renormalize_logits": true,
+    "repetition_penalty": 1.0,
+    "return_dict_in_generate": false,
+    "semantic_infer_token": 129599,
+    "semantic_pad_token": 10000,
+    "semantic_rate_hz": 49.9,
+    "semantic_vocab_size": 10000,
+    "sequence_bias": null,
+    "suppress_tokens": null,
+    "temperature": 0.7,
+    "text_encoding_offset": 10048,
+    "text_pad_token": 129595,
+    "top_k": 50,
+    "top_p": 1.0,
+    "transformers_version": "4.31.0.dev0",
+    "typical_p": 1.0,
+    "use_cache": true
+  }
+}