up

config.json

@@ -0,0 +1,264 @@
+{
+  "architectures": [
+    "SpeechEncoderDecoderModel"
+  ],
+  "decoder": {
+    "_name_or_path": "gpt2-medium",
+    "activation_function": "gelu_new",
+    "add_cross_attention": true,
+    "architectures": [
+      "GPT2LMHeadModel"
+    ],
+    "attn_pdrop": 0.1,
+    "bad_words_ids": null,
+    "bos_token_id": 50256,
+    "chunk_size_feed_forward": 0,
+    "cross_attention_hidden_size": null,
+    "decoder_start_token_id": null,
+    "diversity_penalty": 0.0,
+    "do_sample": false,
+    "early_stopping": false,
+    "embd_pdrop": 0.1,
+    "encoder_no_repeat_ngram_size": 0,
+    "eos_token_id": 50256,
+    "finetuning_task": null,
+    "forced_bos_token_id": null,
+    "forced_eos_token_id": null,
+    "id2label": {
+      "0": "LABEL_0",
+      "1": "LABEL_1"
+    },
+    "initializer_range": 0.02,
+    "is_decoder": true,
+    "is_encoder_decoder": false,
+    "label2id": {
+      "LABEL_0": 0,
+      "LABEL_1": 1
+    },
+    "layer_norm_epsilon": 1e-05,
+    "length_penalty": 1.0,
+    "max_length": 20,
+    "min_length": 0,
+    "model_type": "gpt2",
+    "n_ctx": 1024,
+    "n_embd": 1024,
+    "n_head": 16,
+    "n_inner": null,
+    "n_layer": 24,
+    "n_positions": 1024,
+    "n_special": 0,
+    "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,
+    "predict_special_tokens": true,
+    "prefix": null,
+    "problem_type": null,
+    "pruned_heads": {},
+    "remove_invalid_values": false,
+    "reorder_and_upcast_attn": false,
+    "repetition_penalty": 1.0,
+    "resid_pdrop": 0.1,
+    "return_dict": true,
+    "return_dict_in_generate": false,
+    "scale_attn_by_inverse_layer_idx": false,
+    "scale_attn_weights": true,
+    "sep_token_id": null,
+    "summary_activation": null,
+    "summary_first_dropout": 0.1,
+    "summary_proj_to_labels": true,
+    "summary_type": "cls_index",
+    "summary_use_proj": true,
+    "task_specific_params": {
+      "text-generation": {
+        "do_sample": true,
+        "max_length": 50
+      }
+    },
+    "temperature": 1.0,
+    "tie_encoder_decoder": false,
+    "tie_word_embeddings": true,
+    "tokenizer_class": null,
+    "top_k": 50,
+    "top_p": 1.0,
+    "torch_dtype": null,
+    "torchscript": false,
+    "transformers_version": "4.18.0.dev0",
+    "typical_p": 1.0,
+    "use_bfloat16": false,
+    "use_cache": false,
+    "vocab_size": 50257
+  },
+  "decoder_start_token_id": 50256,
+  "encoder": {
+    "_name_or_path": "facebook/wav2vec2-large-lv60",
+    "activation_dropout": 0.1,
+    "adapter_kernel_size": 3,
+    "adapter_stride": 2,
+    "add_adapter": true,
+    "add_cross_attention": false,
+    "apply_spec_augment": true,
+    "architectures": [
+      "Wav2Vec2ForPreTraining"
+    ],
+    "attention_dropout": 0.1,
+    "bad_words_ids": null,
+    "bos_token_id": 1,
+    "chunk_size_feed_forward": 0,
+    "classifier_proj_size": 256,
+    "codevector_dim": 768,
+    "contrastive_logits_temperature": 0.1,
+    "conv_bias": true,
+    "conv_dim": [
+      512,
+      512,
+      512,
+      512,
+      512,
+      512,
+      512
+    ],
+    "conv_kernel": [
+      10,
+      3,
+      3,
+      3,
+      3,
+      2,
+      2
+    ],
+    "conv_stride": [
+      5,
+      2,
+      2,
+      2,
+      2,
+      2,
+      2
+    ],
+    "cross_attention_hidden_size": null,
+    "ctc_loss_reduction": "sum",
+    "ctc_zero_infinity": false,
+    "decoder_start_token_id": null,
+    "diversity_loss_weight": 0.1,
+    "diversity_penalty": 0.0,
+    "do_sample": false,
+    "do_stable_layer_norm": true,
+    "early_stopping": false,
+    "encoder_no_repeat_ngram_size": 0,
+    "eos_token_id": 2,
+    "feat_extract_activation": "gelu",
+    "feat_extract_dropout": 0.0,
+    "feat_extract_norm": "layer",
+    "feat_proj_dropout": 0.0,
+    "feat_quantizer_dropout": 0.0,
+    "final_dropout": 0.0,
+    "finetuning_task": null,
+    "forced_bos_token_id": null,
+    "forced_eos_token_id": null,
+    "gradient_checkpointing": false,
+    "hidden_act": "gelu",
+    "hidden_dropout": 0.1,
+    "hidden_dropout_prob": 0.1,
+    "hidden_size": 1024,
+    "id2label": {
+      "0": "LABEL_0",
+      "1": "LABEL_1"
+    },
+    "initializer_range": 0.02,
+    "intermediate_size": 4096,
+    "is_decoder": false,
+    "is_encoder_decoder": false,
+    "label2id": {
+      "LABEL_0": 0,
+      "LABEL_1": 1
+    },
+    "layer_norm_eps": 1e-05,
+    "layerdrop": 0.0,
+    "length_penalty": 1.0,
+    "mask_feature_length": 10,
+    "mask_feature_min_masks": 0,
+    "mask_feature_prob": 0.0,
+    "mask_time_length": 10,
+    "mask_time_min_masks": 2,
+    "mask_time_prob": 0.1,
+    "max_length": 20,
+    "min_length": 0,
+    "model_type": "wav2vec2",
+    "no_repeat_ngram_size": 0,
+    "num_adapter_layers": 3,
+    "num_attention_heads": 16,
+    "num_beam_groups": 1,
+    "num_beams": 1,
+    "num_codevector_groups": 2,
+    "num_codevectors_per_group": 320,
+    "num_conv_pos_embedding_groups": 16,
+    "num_conv_pos_embeddings": 128,
+    "num_feat_extract_layers": 7,
+    "num_hidden_layers": 24,
+    "num_negatives": 100,
+    "num_return_sequences": 1,
+    "output_attentions": false,
+    "output_hidden_size": 1024,
+    "output_hidden_states": false,
+    "output_scores": false,
+    "pad_token_id": 0,
+    "prefix": null,
+    "problem_type": null,
+    "proj_codevector_dim": 768,
+    "pruned_heads": {},
+    "remove_invalid_values": false,
+    "repetition_penalty": 1.0,
+    "return_dict": true,
+    "return_dict_in_generate": false,
+    "sep_token_id": null,
+    "task_specific_params": null,
+    "tdnn_dilation": [
+      1,
+      2,
+      3,
+      1,
+      1
+    ],
+    "tdnn_dim": [
+      512,
+      512,
+      512,
+      512,
+      1500
+    ],
+    "tdnn_kernel": [
+      5,
+      3,
+      3,
+      1,
+      1
+    ],
+    "temperature": 1.0,
+    "tie_encoder_decoder": false,
+    "tie_word_embeddings": true,
+    "tokenizer_class": null,
+    "top_k": 50,
+    "top_p": 1.0,
+    "torch_dtype": null,
+    "torchscript": false,
+    "transformers_version": "4.18.0.dev0",
+    "typical_p": 1.0,
+    "use_bfloat16": false,
+    "use_weighted_layer_sum": false,
+    "vocab_size": 32,
+    "xvector_output_dim": 512
+  },
+  "eos_token_id": 50256,
+  "is_encoder_decoder": true,
+  "max_length": 50,
+  "model_type": "speech-encoder-decoder",
+  "pad_token_id": 50256,
+  "processor_class": "Wav2Vec2Processor",
+  "transformers_version": null,
+  "use_cache": false
+}

create_model.py

@@ -0,0 +1,41 @@
+import numpy as np
+from transformers import AutoFeatureExtractor, AutoTokenizer, FlaxSpeechEncoderDecoderModel, GPT2Tokenizer
+
+encoder_id = "facebook/wav2vec2-large-lv60"
+decoder_id = "gpt2-medium"
+
+model = FlaxSpeechEncoderDecoderModel.from_encoder_decoder_pretrained(encoder_id, decoder_id, encoder_add_adapter=True)
+
+# force GPT2 to append EOS to begin and end of seq
+def build_inputs_with_special_tokens(self, token_ids_0, token_ids_1=None):
+    outputs = [self.bos_token_id] + token_ids_0 + [self.eos_token_id]
+    return outputs
+
+GPT2Tokenizer.build_inputs_with_special_tokens = build_inputs_with_special_tokens
+gpt2_tokenizer = GPT2Tokenizer.from_pretrained(decoder_id)
+# set pad_token_id to unk_token_id, note: unk_token_id == eos_token_id == bos_token_id
+gpt2_tokenizer.pad_token = gpt2_tokenizer.unk_token
+gpt2_tokenizer.save_pretrained("./")
+
+model.config.pad_token_id = gpt2_tokenizer.pad_token_id
+model.config.decoder_start_token_id = model.config.decoder.bos_token_id
+model.config.eos_token_id = model.config.decoder.eos_token_id
+model.config.max_length = 50
+model.config.num_beams = 1
+
+# same enc regularisation as wav2vec2-2-bart
+model.config.encoder.feat_proj_dropout = 0.0
+model.config.encoder.final_dropout = 0.0
+model.config.encoder.mask_time_prob = 0.1
+
+model.config.use_cache = False
+model.config.decoder.use_cache = False
+model.config.processor_class = "Wav2Vec2Processor"
+
+# check if generation works
+out = model.generate(np.ones((1, 2000)))
+
+model.save_pretrained("./")
+
+feature_extractor = AutoFeatureExtractor.from_pretrained(encoder_id)
+feature_extractor.save_pretrained("./")

flax_model.msgpack

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:240e5d0555dce67f7eb5f3d95f29de05de215e913ea754d408ac72d91b1f95b7
+size 3159844591

preprocessor_config.json

@@ -0,0 +1,9 @@
+{
+  "do_normalize": true,
+  "feature_extractor_type": "Wav2Vec2FeatureExtractor",
+  "feature_size": 1,
+  "padding_side": "right",
+  "padding_value": 0.0,
+  "return_attention_mask": true,
+  "sampling_rate": 16000
+}

run_flax_speech_recognition_seq2seq.py

@@ -0,0 +1 @@
+/home/sanchitgandhi/transformers/examples/flax/speech-recognition/run_flax_speech_recognition_seq2seq.py

run_librispeech.sh

@@ -0,0 +1,29 @@
+#!/usr/bin/env bash
+JAX_DEFAULT_MATMUL_PRECISION=float32 python run_flax_speech_recognition_seq2seq.py \
+	--dataset_name="librispeech_asr" \
+	--model_name_or_path="./" \
+	--dataset_config_name="clean" \
+	--train_split_name="train.100[:5%]" \
+	--eval_split_name="validation[:5%]" \
+	--output_dir="./" \
+	--preprocessing_num_workers="16" \
+	--length_column_name="input_length" \
+	--overwrite_output_dir \
+	--num_train_epochs="1" \
+	--per_device_train_batch_size="1" \
+	--per_device_eval_batch_size="1" \
+	--logging_steps="1" \
+	--max_duration_in_seconds="10" \
+	--max_target_length="32" \
+	--generation_max_length="40" \
+	--generation_num_beams="1" \
+	--learning_rate="3e-4" \
+	--warmup_steps="500" \
+	--text_column_name="text" \
+	--save_total_limit="1" \
+	--freeze_feature_encoder \
+	--predict_with_generate \
+	--do_lower_case \
+	--do_eval --do_train \
+	--push_to_hub \
+	--use_auth_token 

special_tokens_map.json

@@ -0,0 +1 @@
+{"bos_token": {"content": "<|endoftext|>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, "eos_token": {"content": "<|endoftext|>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, "unk_token": {"content": "<|endoftext|>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, "pad_token": "<|endoftext|>"}

tokenizer_config.json

@@ -0,0 +1 @@
+{"errors": "replace", "unk_token": {"content": "<|endoftext|>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "bos_token": {"content": "<|endoftext|>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "eos_token": {"content": "<|endoftext|>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true, "__type": "AddedToken"}, "add_prefix_space": false, "model_max_length": 1024, "special_tokens_map_file": null, "name_or_path": "gpt2-medium", "tokenizer_class": "GPT2Tokenizer"}