Training in progress, step 200

config.json

@@ -1,7 +1,7 @@
 {
-  "_name_or_path": "facebook/wav2vec2-base-960h",
-  "activation_dropout": 0.1,
-  "adapter_attn_dim": null,
+  "_name_or_path": "facebook/mms-1b-all",
+  "activation_dropout": 0.05,
+  "adapter_attn_dim": 16,
   "adapter_kernel_size": 3,
   "adapter_stride": 2,
   "add_adapter": false,
@@ -12,9 +12,9 @@
   "attention_dropout": 0.0,
   "bos_token_id": 1,
   "classifier_proj_size": 256,
-  "codevector_dim": 256,
+  "codevector_dim": 1024,
   "contrastive_logits_temperature": 0.1,
-  "conv_bias": false,
+  "conv_bias": true,
   "conv_dim": [
     512,
     512,
@@ -45,21 +45,19 @@
   "ctc_loss_reduction": "mean",
   "ctc_zero_infinity": false,
   "diversity_loss_weight": 0.1,
-  "do_stable_layer_norm": false,
+  "do_stable_layer_norm": true,
   "eos_token_id": 2,
   "feat_extract_activation": "gelu",
   "feat_extract_dropout": 0.0,
-  "feat_extract_norm": "group",
+  "feat_extract_norm": "layer",
   "feat_proj_dropout": 0.0,
   "feat_quantizer_dropout": 0.0,
-  "final_dropout": 0.1,
-  "gradient_checkpointing": false,
+  "final_dropout": 0.05,
   "hidden_act": "gelu",
   "hidden_dropout": 0.0,
-  "hidden_dropout_prob": 0.1,
-  "hidden_size": 768,
+  "hidden_size": 1280,
   "initializer_range": 0.02,
-  "intermediate_size": 3072,
+  "intermediate_size": 5120,
   "layer_norm_eps": 1e-05,
   "layerdrop": 0.0,
   "mask_feature_length": 10,
@@ -70,17 +68,17 @@
   "mask_time_prob": 0.05,
   "model_type": "wav2vec2",
   "num_adapter_layers": 3,
-  "num_attention_heads": 12,
+  "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": 12,
+  "num_hidden_layers": 48,
   "num_negatives": 100,
-  "output_hidden_size": 768,
+  "output_hidden_size": 1280,
   "pad_token_id": 108,
-  "proj_codevector_dim": 256,
+  "proj_codevector_dim": 1024,
   "tdnn_dilation": [
     1,
     2,

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:87860264d273b876496932f16980143765fe9f3cf667076032b700dc6bf68fd8
-size 377854132
+oid sha256:2ee18645f5661dfb101aa11d294c0fcb76baab6651909879d38ba66f2a8794d0
+size 3859300844

runs/Dec04_11-43-31_8cb8a5bb18ab/events.out.tfevents.1701690370.8cb8a5bb18ab.219.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:afd91b7bbd026f4b2f9b604f6d99cffd91c11bab93f34562a450b9bb8b168cfb
+size 6715

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:838af9a902d351815fb954d4d4c22ddbe17d7492d75d821638524c24af3d287f
-size 4536
+oid sha256:120a92afde0ac02470e8f72df58261711ea8621af766d6071c4b3f4cdb3ec5f4
+size 4664