Add ONNX FP32 models

.gitattributes

@@ -33,3 +33,6 @@ 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
+onnx/decoder_model.onnx.data filter=lfs diff=lfs merge=lfs -text
+onnx/decoder_with_past_model.onnx.data filter=lfs diff=lfs merge=lfs -text
+onnx/encoder_model.onnx.data filter=lfs diff=lfs merge=lfs -text

onnx/config.json

@@ -0,0 +1,73 @@
+{
+  "attention_bias": false,
+  "attention_dropout": 0.0,
+  "bos_token_id": 1,
+  "decoder_start_token_id": 1,
+  "dtype": "float32",
+  "encoder_config": {
+    "attention_bias": false,
+    "attention_dropout": 0.0,
+    "dtype": "float32",
+    "frame_ms": 5.0,
+    "head_dim": 40,
+    "hidden_act": "gelu",
+    "hidden_size": 320,
+    "intermediate_size": 1280,
+    "max_position_embeddings": 4096,
+    "model_type": "moonshine_streaming_encoder",
+    "num_attention_heads": 8,
+    "num_hidden_layers": 6,
+    "num_key_value_heads": 8,
+    "sample_rate": 16000,
+    "sliding_windows": [
+      [
+        16,
+        4
+      ],
+      [
+        16,
+        4
+      ],
+      [
+        16,
+        0
+      ],
+      [
+        16,
+        0
+      ],
+      [
+        16,
+        4
+      ],
+      [
+        16,
+        4
+      ]
+    ]
+  },
+  "encoder_hidden_size": 320,
+  "eos_token_id": 2,
+  "ffn_mult": 4,
+  "head_dim": 40,
+  "hidden_act": "silu",
+  "hidden_size": 320,
+  "intermediate_size": 1280,
+  "is_encoder_decoder": true,
+  "max_position_embeddings": 4096,
+  "model_type": "moonshine_streaming",
+  "num_attention_heads": 8,
+  "num_hidden_layers": 6,
+  "num_key_value_heads": 8,
+  "pad_head_dim_to_multiple_of": null,
+  "pad_token_id": 0,
+  "rope_parameters": {
+    "partial_rotary_factor": 0.8,
+    "rope_theta": 10000.0,
+    "rope_type": "default"
+  },
+  "tie_word_embeddings": false,
+  "transformers_version": "5.6.2",
+  "use_cache": true,
+  "vocab_size": 32768
+}

onnx/decoder_model.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:740bad720c0f5c9a85d136e8dcf663ae1e0c1983d08b02fd16c72f319cc412ed
+size 1133251

onnx/decoder_model.onnx.data

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:706c2c14ea7b82968111bfa7d8aa2d80a4f37c2bb35eaa1e0bfc7cd9873207e8
+size 138412032

onnx/decoder_with_past_model.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f5291822f80aa9b76a7c2511290a6edb54a256e7ca7ae4fd93e7213f63ae80ca
+size 1163100

onnx/decoder_with_past_model.onnx.data

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c03f554a4180eba9171ffe0d8ba698c02cee25d727d7f04add71fdf3e14b8157
+size 128253952

onnx/encoder_model.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:de20c3c602636498980194646bbc87b219c74c5d8eaa769ff8328da0929f7d05
+size 699936

onnx/encoder_model.onnx.data

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b990c5914f11fe7b5a393afebdc1cc41c2f295e3aee04d3e4f6aef65b1f46dc9
+size 37912576

onnx/processor_config.json

@@ -0,0 +1,13 @@
+{
+  "feature_extractor": {
+    "do_normalize": false,
+    "feature_extractor_type": "Wav2Vec2FeatureExtractor",
+    "feature_size": 1,
+    "pad_to_multiple_of": 80,
+    "padding_side": "right",
+    "padding_value": 0.0,
+    "return_attention_mask": true,
+    "sampling_rate": 16000
+  },
+  "processor_class": "MoonshineStreamingProcessor"
+}

onnx/tokenizer_config.json

@@ -0,0 +1,12 @@
+{
+  "backend": "tokenizers",
+  "bos_token": "<s>",
+  "eos_token": "</s>",
+  "is_local": false,
+  "local_files_only": false,
+  "model_max_length": 4096,
+  "pad_token": "<unk>",
+  "processor_class": "MoonshineStreamingProcessor",
+  "tokenizer_class": "TokenizersBackend",
+  "unk_token": "<unk>"
+}