Training in progress, step 5

config.json

@@ -1,5 +1,5 @@
 {
-  "_name_or_path": "hf-test/xls-r-dummy",
+  "_name_or_path": "facebook/wav2vec2-xls-r-300m",
   "activation_dropout": 0.0,
   "adapter_kernel_size": 3,
   "adapter_stride": 2,
@@ -11,23 +11,35 @@
   "attention_dropout": 0.0,
   "bos_token_id": 1,
   "classifier_proj_size": 256,
-  "codevector_dim": 256,
+  "codevector_dim": 768,
   "contrastive_logits_temperature": 0.1,
-  "conv_bias": false,
+  "conv_bias": true,
   "conv_dim": [
-    32,
-    32,
-    32
+    512,
+    512,
+    512,
+    512,
+    512,
+    512,
+    512
   ],
   "conv_kernel": [
-    8,
-    8,
-    8
+    10,
+    3,
+    3,
+    3,
+    3,
+    2,
+    2
   ],
   "conv_stride": [
-    4,
-    4,
-    4
+    5,
+    2,
+    2,
+    2,
+    2,
+    2,
+    2
   ],
   "ctc_loss_reduction": "mean",
   "ctc_zero_infinity": false,
@@ -42,10 +54,9 @@
   "final_dropout": 0.0,
   "hidden_act": "gelu",
   "hidden_dropout": 0.0,
-  "hidden_dropout_prob": 0.1,
-  "hidden_size": 16,
+  "hidden_size": 1024,
   "initializer_range": 0.02,
-  "intermediate_size": 20,
+  "intermediate_size": 4096,
   "layer_norm_eps": 1e-05,
   "layerdrop": 0.0,
   "mask_feature_length": 10,
@@ -56,17 +67,17 @@
   "mask_time_prob": 0.05,
   "model_type": "wav2vec2",
   "num_adapter_layers": 3,
-  "num_attention_heads": 2,
+  "num_attention_heads": 16,
   "num_codevector_groups": 2,
   "num_codevectors_per_group": 320,
-  "num_conv_pos_embedding_groups": 2,
-  "num_conv_pos_embeddings": 16,
-  "num_feat_extract_layers": 3,
-  "num_hidden_layers": 4,
-  "num_negatives": 10,
-  "output_hidden_size": 16,
+  "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": 108,
-  "proj_codevector_dim": 256,
+  "proj_codevector_dim": 768,
   "tdnn_dilation": [
     1,
     2,

preprocessor_config.json

@@ -3,7 +3,7 @@
   "feature_extractor_type": "Wav2Vec2FeatureExtractor",
   "feature_size": 1,
   "padding_side": "right",
-  "padding_value": 0.0,
-  "return_attention_mask": false,
+  "padding_value": 0,
+  "return_attention_mask": true,
   "sampling_rate": 16000
 }

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5a3cb1b2b64a17f1c10451a526278e2e3072fe6444c43ef4f6c91091065317ce
-size 147814
+oid sha256:b84b86b71a70d39e8fdda0f7e05e065028e99b889512345becf3fdc8a56c4106
+size 1262378737

run.sh

@@ -1,6 +1,6 @@
 python run_speech_recognition_ctc.py \
 --dataset_name="mozilla-foundation/common_voice_7_0" \
---model_name_or_path="hf-test/xls-r-dummy" \
+--model_name_or_path="facebook/wav2vec2-xls-r-300m" \
 --dataset_config_name="ta" \
 --output_dir="./" \
 --overwrite_output_dir \

special_tokens_map.json

@@ -1 +1 @@
-{"bos_token": "<s>", "eos_token": "</s>", "unk_token": "[UNK]", "pad_token": "[PAD]", "additional_special_tokens": [{"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}]}
+{"bos_token": "<s>", "eos_token": "</s>", "unk_token": "[UNK]", "pad_token": "[PAD]", "additional_special_tokens": [{"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, {"content": "<s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}, {"content": "</s>", "single_word": false, "lstrip": false, "rstrip": false, "normalized": true}]}

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e1b06503fbf2f5e66af00bf54614c1db9e1bd34fea3fd357f93c9d7590e9670c
+oid sha256:b381b5987de34382e58ae6875e3f7b8c627bb1466de9cb0bd135536993bb72ab
 size 2991