Update tiny models for VibeVoiceAsrForConditionalGeneration

.gitattributes

@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+tokenizer.json filter=lfs diff=lfs merge=lfs -text

chat_template.jinja

@@ -0,0 +1,26 @@
+{%- set system_prompt = system_prompt | default("You are a helpful assistant that transcribes audio input into text output in JSON format.") -%}
+<|im_start|>system
+{{ system_prompt }}<|im_end|>
+{%- set audio_token = audio_token | default("<|box_start|>") -%}
+{%- set audio_start_token = "<|object_ref_start|>" -%}
+{%- set audio_end_token = "<|object_ref_end|>" -%}
+{%- for message in messages -%}
+    {%- if message['role'] == 'user' -%}
+{{ '
+' }}<|im_start|>user{{ '
+' }}{%- set text_items = message['content'] | selectattr('type', 'equalto', 'text') | list -%}
+        {%- set context_text = text_items[0]['text'] if text_items else none -%}
+        {%- for item in message['content'] -%}
+            {%- if item['type'] == 'audio' -%}
+{{ audio_start_token }}{{ audio_token }}{{ audio_end_token }}{{ "
+" }}{%- if context_text -%}
+This is a <|AUDIO_DURATION|> seconds audio, with extra info: {{ context_text }}
+
+Please transcribe it with these keys: Start time, End time, Speaker ID, Content{%- else -%}
+This is a <|AUDIO_DURATION|> seconds audio, please transcribe it with these keys: Start time, End time, Speaker ID, Content{%- endif -%}
+            {%- endif -%}
+        {%- endfor -%}
+<|im_end|>{{ '
+' }}
+    {%- endif -%}
+{%- endfor -%}

config.json

@@ -0,0 +1,84 @@
+{
+  "acoustic_tokenizer_chunk_size": 400,
+  "acoustic_tokenizer_encoder_config": {
+    "channels": 1,
+    "depths": [
+      1,
+      1
+    ],
+    "downsampling_ratios": [
+      2
+    ],
+    "ffn_expansion": 4,
+    "hidden_act": "gelu",
+    "hidden_size": 16,
+    "initializer_range": 0.01,
+    "kernel_size": 3,
+    "layer_scale_init_value": 1e-06,
+    "model_type": "vibevoice_acoustic_tokenizer_encoder",
+    "n_filters": 4,
+    "num_filters": 32,
+    "rms_norm_eps": 1e-05,
+    "vae_std": 0.625
+  },
+  "architectures": [
+    "VibeVoiceAsrForConditionalGeneration"
+  ],
+  "audio_bos_token_id": 151646,
+  "audio_eos_token_id": 151647,
+  "audio_token_id": 0,
+  "dtype": "float32",
+  "model_type": "vibevoice_asr",
+  "semantic_tokenizer_encoder_config": {
+    "channels": 1,
+    "depths": [
+      1,
+      1
+    ],
+    "downsampling_ratios": [
+      2
+    ],
+    "ffn_expansion": 4,
+    "hidden_act": "gelu",
+    "hidden_size": 32,
+    "initializer_range": 0.01,
+    "kernel_size": 3,
+    "layer_scale_init_value": 1e-06,
+    "model_type": "vibevoice_acoustic_tokenizer_encoder",
+    "n_filters": 4,
+    "num_filters": 32,
+    "rms_norm_eps": 1e-05,
+    "vae_std": 0.625
+  },
+  "text_config": {
+    "attention_dropout": 0.0,
+    "bos_token_id": null,
+    "eos_token_id": null,
+    "hidden_act": "silu",
+    "hidden_size": 32,
+    "initializer_range": 0.02,
+    "intermediate_size": 36,
+    "layer_types": [
+      "full_attention",
+      "full_attention"
+    ],
+    "max_position_embeddings": 200,
+    "max_window_layers": 28,
+    "model_type": "qwen2",
+    "num_attention_heads": 4,
+    "num_hidden_layers": 2,
+    "num_key_value_heads": 4,
+    "pad_token_id": 151643,
+    "rms_norm_eps": 1e-06,
+    "rope_parameters": {
+      "rope_theta": 10000.0,
+      "rope_type": "default"
+    },
+    "sliding_window": null,
+    "tie_word_embeddings": false,
+    "use_cache": true,
+    "use_sliding_window": false,
+    "vocab_size": 151665
+  },
+  "transformers_version": "5.16.0.dev0"
+}

generation_config.json

@@ -0,0 +1,8 @@
+{
+  "_from_model_config": true,
+  "output_attentions": false,
+  "output_hidden_states": false,
+  "pad_token_id": 151643,
+  "transformers_version": "5.16.0.dev0",
+  "use_cache": true
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f867d8182da0f171d34f5d3eb2f2802f5557c3a587cb84309650c24cc64452d5
+size 39355864

preprocessor_config.json

@@ -0,0 +1,11 @@
+{
+  "eps": 1e-06,
+  "feature_extractor_type": "VibeVoiceAcousticTokenizerFeatureExtractor",
+  "feature_size": 1,
+  "normalize_audio": true,
+  "padding_side": "right",
+  "padding_value": 0.0,
+  "return_attention_mask": true,
+  "sampling_rate": 24000,
+  "target_dB_FS": -25
+}

tokenizer.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3fd169731d2cbde95e10bf356d66d5997fd885dd8dbb6fb4684da3f23b2585d8
+size 11421892

tokenizer_config.json

@@ -0,0 +1,31 @@
+{
+  "add_prefix_space": false,
+  "backend": "tokenizers",
+  "bos_token": null,
+  "clean_up_tokenization_spaces": false,
+  "eos_token": "<|endoftext|>",
+  "errors": "replace",
+  "extra_special_tokens": [
+    "<|im_start|>",
+    "<|im_end|>",
+    "<|object_ref_start|>",
+    "<|object_ref_end|>",
+    "<|box_start|>",
+    "<|box_end|>",
+    "<|quad_start|>",
+    "<|quad_end|>",
+    "<|vision_start|>",
+    "<|vision_end|>",
+    "<|vision_pad|>",
+    "<|image_pad|>",
+    "<|video_pad|>"
+  ],
+  "is_local": true,
+  "local_files_only": false,
+  "model_max_length": 200,
+  "pad_token": "<|endoftext|>",
+  "processor_class": "VibeVoiceAsrProcessor",
+  "split_special_tokens": false,
+  "tokenizer_class": "Qwen2Tokenizer",
+  "unk_token": null
+}