Add model files and logs. Update README.

README.md

@@ -1,3 +1,39 @@
 ---
+language:
+- am
 license: mit
+tags:
+- automatic-speech-recognition
+- speech
+metrics:
+- wer
+- cer
 ---
+
+# Amharic ASR using fine-tuned Wav2vec2 XLSR-53
+This is a finetuned version of [facebook/wav2vec2-large-xlsr-53](https://huggingface.co/facebook/wav2vec2-large-xlsr-53) trained on the [Amharic Speech Corpus](http://www.openslr.org/25/). This corpus was produced by [Abate et al. (2005)](https://www.isca-speech.org/archive/interspeech_2005/abate05_interspeech.html) (10.21437/Interspeech.2005-467).
+
+## Usage
+The model can be used as follows:
+```python
+import librosa
+from transformers import Wav2Vec2ForCTC, Wav2Vec2Processor
+
+model = Wav2Vec2ForCTC.from_pretrained("agkphysics/wav2vec2-large-xlsr-53-amharic")
+processor = Wav2Vec2Processor.from_pretrained("agkphysics/wav2vec2-large-xlsr-53-amharic")
+
+audio, _ = librosa.load("/path/to/audio.wav", sr=16000)
+
+input_values = processor(
+    audio.squeeze(),
+    sampling_rate=16000,
+    return_tensors="pt"
+).input_values
+
+model.eval()
+with torch.no_grad():
+    logits = model(input_values).logits
+    preds = logits.argmax(-1)
+    texts = processor.batch_decode(preds)
+print(texts[0])
+```

config.json

@@ -0,0 +1,115 @@
+{
+  "activation_dropout": 0.0,
+  "adapter_kernel_size": 3,
+  "adapter_stride": 2,
+  "add_adapter": false,
+  "apply_spec_augment": true,
+  "architectures": [
+    "Wav2Vec2ForCTC"
+  ],
+  "attention_dropout": 0.1,
+  "bos_token_id": 1,
+  "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
+  ],
+  "ctc_loss_reduction": "mean",
+  "ctc_zero_infinity": false,
+  "diversity_loss_weight": 0.1,
+  "do_stable_layer_norm": true,
+  "eos_token_id": 2,
+  "feat_extract_activation": "gelu",
+  "feat_extract_dropout": 0.0,
+  "feat_extract_norm": "layer",
+  "feat_proj_dropout": 0.1,
+  "feat_quantizer_dropout": 0.0,
+  "final_dropout": 0.0,
+  "gradient_checkpointing": false,
+  "hidden_act": "gelu",
+  "hidden_dropout": 0.1,
+  "hidden_size": 1024,
+  "initializer_range": 0.02,
+  "intermediate_size": 4096,
+  "layer_norm_eps": 1e-05,
+  "layerdrop": 0.1,
+  "mask_channel_length": 10,
+  "mask_channel_min_space": 1,
+  "mask_channel_other": 0.0,
+  "mask_channel_prob": 0.0,
+  "mask_channel_selection": "static",
+  "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_min_space": 1,
+  "mask_time_other": 0.0,
+  "mask_time_prob": 0.075,
+  "mask_time_selection": "static",
+  "model_type": "wav2vec2",
+  "num_adapter_layers": 3,
+  "num_attention_heads": 16,
+  "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,
+  "output_hidden_size": 1024,
+  "pad_token_id": 0,
+  "proj_codevector_dim": 768,
+  "tdnn_dilation": [
+    1,
+    2,
+    3,
+    1,
+    1
+  ],
+  "tdnn_dim": [
+    512,
+    512,
+    512,
+    512,
+    1500
+  ],
+  "tdnn_kernel": [
+    5,
+    3,
+    3,
+    1,
+    1
+  ],
+  "torch_dtype": "float32",
+  "transformers_version": "4.26.1",
+  "use_weighted_layer_sum": false,
+  "vocab_size": 234,
+  "xvector_output_dim": 512
+}

optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:007a6599f2e95d60f7d96669ed2062ca92ee8c71d5172ffa7e4d7cf029c5a63c
+size 2491980485

preprocessor_config.json

@@ -0,0 +1,10 @@
+{
+  "do_normalize": true,
+  "feature_extractor_type": "Wav2Vec2FeatureExtractor",
+  "feature_size": 1,
+  "padding_side": "right",
+  "padding_value": 0,
+  "processor_class": "Wav2Vec2Processor",
+  "return_attention_mask": true,
+  "sampling_rate": 16000
+}

pytorch_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:19438fd56ddaa550eae67241f7f3bd042315e302dd51388f15c1c713eb49fa36
+size 1262861229

rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0b3a0a5d11e512a58a0b6622a9d1dc932091ce202dcc19cbc4c29834312a3019
+size 17641

runs/May17_11-43-47_hgx1.novalocal/1684280642.7338111/events.out.tfevents.1684280642.hgx1.novalocal.436095.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:879228333a7024d0ae8074be96899070a111697eac7ad483f90f05cc4d4ec01d
+size 5732

runs/May17_11-43-47_hgx1.novalocal/events.out.tfevents.1684280642.hgx1.novalocal.436095.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b4b52cdefd1977d0d3c1cefc2a4f9ca9fed1a877d37da5d82440393692ec9545
+size 597176

scaler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3a569bd05dae496939d6f816ab83c4619af7025b75bcdb078cd9542c7e570b9a
+size 557

scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7be0030e8a72dfd3795e709437759f1cfb2fcfdb16e2781eede417e5896550ad
+size 627

special_tokens_map.json

@@ -0,0 +1,6 @@
+{
+  "bos_token": "<s>",
+  "eos_token": "</s>",
+  "pad_token": "<pad>",
+  "unk_token": "<unk>"
+}

tokenizer_config.json

@@ -0,0 +1,12 @@
+{
+  "bos_token": "<s>",
+  "do_lower_case": false,
+  "eos_token": "</s>",
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "<pad>",
+  "processor_class": "Wav2Vec2Processor",
+  "replace_word_delimiter_char": " ",
+  "tokenizer_class": "Wav2Vec2CTCTokenizer",
+  "unk_token": "<unk>",
+  "word_delimiter_token": " "
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9af8d0388d3afd7f7d0062165b3d2429f1fd23b1ebee144559698a59b949ead2
+size 3515

vocab.json

@@ -0,0 +1,236 @@
+{
+  " ": 4,
+  "</s>": 2,
+  "<pad>": 0,
+  "<s>": 1,
+  "<unk>": 3,
+  "ሀ": 5,
+  "ሁ": 6,
+  "ሂ": 7,
+  "ሃ": 8,
+  "ሄ": 9,
+  "ህ": 10,
+  "ሆ": 11,
+  "ለ": 12,
+  "ሉ": 13,
+  "ሊ": 14,
+  "ላ": 15,
+  "ሌ": 16,
+  "ል": 17,
+  "ሎ": 18,
+  "ሏ": 19,
+  "መ": 20,
+  "ሙ": 21,
+  "ሚ": 22,
+  "ማ": 23,
+  "ሜ": 24,
+  "ም": 25,
+  "ሞ": 26,
+  "ሟ": 27,
+  "ረ": 28,
+  "ሩ": 29,
+  "ሪ": 30,
+  "ራ": 31,
+  "ሬ": 32,
+  "ር": 33,
+  "ሮ": 34,
+  "ሯ": 35,
+  "ሰ": 36,
+  "ሱ": 37,
+  "ሲ": 38,
+  "ሳ": 39,
+  "ሴ": 40,
+  "ስ": 41,
+  "ሶ": 42,
+  "ሷ": 43,
+  "ሸ": 44,
+  "ሹ": 45,
+  "ሺ": 46,
+  "ሻ": 47,
+  "ሼ": 48,
+  "ሽ": 49,
+  "ሾ": 50,
+  "ሿ": 51,
+  "ቀ": 52,
+  "ቁ": 53,
+  "ቂ": 54,
+  "ቃ": 55,
+  "ቄ": 56,
+  "ቅ": 57,
+  "ቆ": 58,
+  "ቋ": 59,
+  "በ": 60,
+  "ቡ": 61,
+  "ቢ": 62,
+  "ባ": 63,
+  "ቤ": 64,
+  "ብ": 65,
+  "ቦ": 66,
+  "ቧ": 67,
+  "ቨ": 68,
+  "ቩ": 69,
+  "ቪ": 70,
+  "ቫ": 71,
+  "ቬ": 72,
+  "ቭ": 73,
+  "ቮ": 74,
+  "ቯ": 75,
+  "ተ": 76,
+  "ቱ": 77,
+  "ቲ": 78,
+  "ታ": 79,
+  "ቴ": 80,
+  "ት": 81,
+  "ቶ": 82,
+  "ቷ": 83,
+  "ቸ": 84,
+  "ቹ": 85,
+  "ቺ": 86,
+  "ቻ": 87,
+  "ቼ": 88,
+  "ች": 89,
+  "ቾ": 90,
+  "ቿ": 91,
+  "ኋ": 92,
+  "ነ": 93,
+  "ኑ": 94,
+  "ኒ": 95,
+  "ና": 96,
+  "ኔ": 97,
+  "ን": 98,
+  "ኖ": 99,
+  "ኗ": 100,
+  "ኘ": 101,
+  "ኙ": 102,
+  "ኚ": 103,
+  "ኛ": 104,
+  "ኜ": 105,
+  "ኝ": 106,
+  "ኞ": 107,
+  "ኟ": 108,
+  "አ": 109,
+  "ኡ": 110,
+  "ኢ": 111,
+  "ኣ": 112,
+  "ኤ": 113,
+  "እ": 114,
+  "ኦ": 115,
+  "ከ": 116,
+  "ኩ": 117,
+  "ኪ": 118,
+  "ካ": 119,
+  "ኬ": 120,
+  "ክ": 121,
+  "ኮ": 122,
+  "ኰ": 123,
+  "ኲ": 124,
+  "ኳ": 125,
+  "ኸ": 126,
+  "ወ": 127,
+  "ዉ": 128,
+  "ዊ": 129,
+  "ዋ": 130,
+  "ዌ": 131,
+  "ው": 132,
+  "ዎ": 133,
+  "ዘ": 134,
+  "ዙ": 135,
+  "ዚ": 136,
+  "ዛ": 137,
+  "ዜ": 138,
+  "ዝ": 139,
+  "ዞ": 140,
+  "ዟ": 141,
+  "ዠ": 142,
+  "ዡ": 143,
+  "ዢ": 144,
+  "ዣ": 145,
+  "ዤ": 146,
+  "ዥ": 147,
+  "ዦ": 148,
+  "ዧ": 149,
+  "የ": 150,
+  "ዩ": 151,
+  "ዪ": 152,
+  "ያ": 153,
+  "ዬ": 154,
+  "ይ": 155,
+  "ዮ": 156,
+  "ደ": 157,
+  "ዱ": 158,
+  "ዲ": 159,
+  "ዳ": 160,
+  "ዴ": 161,
+  "ድ": 162,
+  "ዶ": 163,
+  "ዷ": 164,
+  "ዸ": 165,
+  "ዺ": 166,
+  "ዻ": 167,
+  "ዼ": 168,
+  "ዽ": 169,
+  "ዾ": 170,
+  "ጀ": 171,
+  "ጁ": 172,
+  "ጂ": 173,
+  "ጃ": 174,
+  "ጄ": 175,
+  "ጅ": 176,
+  "ጆ": 177,
+  "ጇ": 178,
+  "ገ": 179,
+  "ጉ": 180,
+  "ጊ": 181,
+  "ጋ": 182,
+  "ጌ": 183,
+  "ግ": 184,
+  "ጐ": 185,
+  "ጓ": 186,
+  "ጔ": 187,
+  "ጠ": 188,
+  "ጡ": 189,
+  "ጢ": 190,
+  "ጣ": 191,
+  "ጤ": 192,
+  "ጥ": 193,
+  "ጦ": 194,
+  "ጧ": 195,
+  "ጨ": 196,
+  "ጩ": 197,
+  "ጪ": 198,
+  "ጫ": 199,
+  "ጬ": 200,
+  "ጭ": 201,
+  "ጮ": 202,
+  "ጯ": 203,
+  "ጰ": 204,
+  "ጱ": 205,
+  "ጲ": 206,
+  "ጳ": 207,
+  "ጴ": 208,
+  "ጵ": 209,
+  "ጶ": 210,
+  "ጸ": 211,
+  "ጹ": 212,
+  "ጺ": 213,
+  "ጻ": 214,
+  "ጼ": 215,
+  "ጽ": 216,
+  "ጾ": 217,
+  "ጿ": 218,
+  "ፈ": 219,
+  "ፉ": 220,
+  "ፊ": 221,
+  "ፋ": 222,
+  "ፌ": 223,
+  "ፍ": 224,
+  "ፎ": 225,
+  "ፏ": 226,
+  "ፐ": 227,
+  "ፑ": 228,
+  "ፒ": 229,
+  "ፓ": 230,
+  "ፔ": 231,
+  "ፕ": 232,
+  "ፖ": 233
+}