Upload 9 files
Browse files- config.json +82 -0
- model.safetensors +3 -0
- optimizer.pt +3 -0
- preprocessor_config.json +10 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- trainer_state.json +1218 -0
- training_args.bin +3 -0
- vocab.json +1 -0
config.json
ADDED
|
@@ -0,0 +1,82 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "facebook/w2v-bert-2.0",
|
| 3 |
+
"activation_dropout": 0.0,
|
| 4 |
+
"adapter_act": "relu",
|
| 5 |
+
"adapter_kernel_size": 3,
|
| 6 |
+
"adapter_stride": 2,
|
| 7 |
+
"add_adapter": true,
|
| 8 |
+
"apply_spec_augment": false,
|
| 9 |
+
"architectures": [
|
| 10 |
+
"Wav2Vec2BertForCTC"
|
| 11 |
+
],
|
| 12 |
+
"attention_dropout": 0.0,
|
| 13 |
+
"bos_token_id": 1,
|
| 14 |
+
"classifier_proj_size": 768,
|
| 15 |
+
"codevector_dim": 768,
|
| 16 |
+
"conformer_conv_dropout": 0.1,
|
| 17 |
+
"contrastive_logits_temperature": 0.1,
|
| 18 |
+
"conv_depthwise_kernel_size": 31,
|
| 19 |
+
"ctc_loss_reduction": "mean",
|
| 20 |
+
"ctc_zero_infinity": true,
|
| 21 |
+
"diversity_loss_weight": 0.1,
|
| 22 |
+
"eos_token_id": 2,
|
| 23 |
+
"feat_proj_dropout": 0.0,
|
| 24 |
+
"feat_quantizer_dropout": 0.0,
|
| 25 |
+
"feature_projection_input_dim": 160,
|
| 26 |
+
"final_dropout": 0.1,
|
| 27 |
+
"hidden_act": "swish",
|
| 28 |
+
"hidden_dropout": 0.0,
|
| 29 |
+
"hidden_size": 1024,
|
| 30 |
+
"initializer_range": 0.02,
|
| 31 |
+
"intermediate_size": 4096,
|
| 32 |
+
"layer_norm_eps": 1e-05,
|
| 33 |
+
"layerdrop": 0.0,
|
| 34 |
+
"left_max_position_embeddings": 64,
|
| 35 |
+
"mask_feature_length": 10,
|
| 36 |
+
"mask_feature_min_masks": 0,
|
| 37 |
+
"mask_feature_prob": 0.0,
|
| 38 |
+
"mask_time_length": 10,
|
| 39 |
+
"mask_time_min_masks": 2,
|
| 40 |
+
"mask_time_prob": 0.0,
|
| 41 |
+
"max_source_positions": 5000,
|
| 42 |
+
"model_type": "wav2vec2-bert",
|
| 43 |
+
"num_adapter_layers": 1,
|
| 44 |
+
"num_attention_heads": 16,
|
| 45 |
+
"num_codevector_groups": 2,
|
| 46 |
+
"num_codevectors_per_group": 320,
|
| 47 |
+
"num_hidden_layers": 24,
|
| 48 |
+
"num_negatives": 100,
|
| 49 |
+
"output_hidden_size": 1024,
|
| 50 |
+
"pad_token_id": 64,
|
| 51 |
+
"position_embeddings_type": "relative_key",
|
| 52 |
+
"proj_codevector_dim": 768,
|
| 53 |
+
"right_max_position_embeddings": 8,
|
| 54 |
+
"rotary_embedding_base": 10000,
|
| 55 |
+
"tdnn_dilation": [
|
| 56 |
+
1,
|
| 57 |
+
2,
|
| 58 |
+
3,
|
| 59 |
+
1,
|
| 60 |
+
1
|
| 61 |
+
],
|
| 62 |
+
"tdnn_dim": [
|
| 63 |
+
512,
|
| 64 |
+
512,
|
| 65 |
+
512,
|
| 66 |
+
512,
|
| 67 |
+
1500
|
| 68 |
+
],
|
| 69 |
+
"tdnn_kernel": [
|
| 70 |
+
5,
|
| 71 |
+
3,
|
| 72 |
+
3,
|
| 73 |
+
1,
|
| 74 |
+
1
|
| 75 |
+
],
|
| 76 |
+
"torch_dtype": "float32",
|
| 77 |
+
"transformers_version": "4.42.4",
|
| 78 |
+
"use_intermediate_ffn_before_adapter": false,
|
| 79 |
+
"use_weighted_layer_sum": false,
|
| 80 |
+
"vocab_size": 67,
|
| 81 |
+
"xvector_output_dim": 512
|
| 82 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:155c37ac72d0bba5739d459ccce59c55ea02d71c0897620e00d67797ef90c85e
|
| 3 |
+
size 2423089260
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fc1784a2b53b59419f2d8b1e1d3d32e8474595854fc49b083f0830140d040f77
|
| 3 |
+
size 4846639530
|
preprocessor_config.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"feature_extractor_type": "SeamlessM4TFeatureExtractor",
|
| 3 |
+
"feature_size": 80,
|
| 4 |
+
"num_mel_bins": 80,
|
| 5 |
+
"padding_side": "right",
|
| 6 |
+
"padding_value": 0.0,
|
| 7 |
+
"return_attention_mask": true,
|
| 8 |
+
"sampling_rate": 16000,
|
| 9 |
+
"stride": 2
|
| 10 |
+
}
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:647bef8e68eb2c12facf33dcc03c50a01444de59062ab22314c36c8a0a339d32
|
| 3 |
+
size 14244
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5e642b7a3aa10b1eec9dafcb1439b91de673591d3172a6ed1ab7db3390a9fc55
|
| 3 |
+
size 1256
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1218 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.23266415805525312,
|
| 3 |
+
"best_model_checkpoint": "results/punjabi_model/20240813_012221/checkpoint-5000",
|
| 4 |
+
"epoch": 4.815177439288638,
|
| 5 |
+
"eval_steps": 5000,
|
| 6 |
+
"global_step": 75000,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.03210118292859092,
|
| 13 |
+
"grad_norm": Infinity,
|
| 14 |
+
"learning_rate": 1.24e-05,
|
| 15 |
+
"loss": 2.1834,
|
| 16 |
+
"step": 500
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.06420236585718184,
|
| 20 |
+
"grad_norm": 9.310239791870117,
|
| 21 |
+
"learning_rate": 2.4900000000000002e-05,
|
| 22 |
+
"loss": 0.676,
|
| 23 |
+
"step": 1000
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.09630354878577276,
|
| 27 |
+
"grad_norm": 8.342071533203125,
|
| 28 |
+
"learning_rate": 3.74e-05,
|
| 29 |
+
"loss": 0.577,
|
| 30 |
+
"step": 1500
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.12840473171436367,
|
| 34 |
+
"grad_norm": 8.718728065490723,
|
| 35 |
+
"learning_rate": 4.99e-05,
|
| 36 |
+
"loss": 0.5662,
|
| 37 |
+
"step": 2000
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.1605059146429546,
|
| 41 |
+
"grad_norm": 8.425736427307129,
|
| 42 |
+
"learning_rate": 4.98390243902439e-05,
|
| 43 |
+
"loss": 0.5139,
|
| 44 |
+
"step": 2500
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.1926070975715455,
|
| 48 |
+
"grad_norm": 7.612177848815918,
|
| 49 |
+
"learning_rate": 4.967642276422764e-05,
|
| 50 |
+
"loss": 0.4737,
|
| 51 |
+
"step": 3000
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.22470828050013644,
|
| 55 |
+
"grad_norm": 6.828492164611816,
|
| 56 |
+
"learning_rate": 4.951382113821139e-05,
|
| 57 |
+
"loss": 0.4696,
|
| 58 |
+
"step": 3500
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.25680946342872735,
|
| 62 |
+
"grad_norm": 25.135059356689453,
|
| 63 |
+
"learning_rate": 4.9351219512195126e-05,
|
| 64 |
+
"loss": 0.4487,
|
| 65 |
+
"step": 4000
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.2889106463573183,
|
| 69 |
+
"grad_norm": 6.97437047958374,
|
| 70 |
+
"learning_rate": 4.9188617886178865e-05,
|
| 71 |
+
"loss": 0.4259,
|
| 72 |
+
"step": 4500
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.3210118292859092,
|
| 76 |
+
"grad_norm": 4.6233744621276855,
|
| 77 |
+
"learning_rate": 4.9026016260162604e-05,
|
| 78 |
+
"loss": 0.3788,
|
| 79 |
+
"step": 5000
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.3210118292859092,
|
| 83 |
+
"eval_loss": 0.4430150091648102,
|
| 84 |
+
"eval_runtime": 597.4312,
|
| 85 |
+
"eval_samples_per_second": 11.601,
|
| 86 |
+
"eval_steps_per_second": 11.601,
|
| 87 |
+
"eval_wer": 0.23266415805525312,
|
| 88 |
+
"step": 5000
|
| 89 |
+
},
|
| 90 |
+
{
|
| 91 |
+
"epoch": 0.3531130122145001,
|
| 92 |
+
"grad_norm": 4.593179225921631,
|
| 93 |
+
"learning_rate": 4.886341463414634e-05,
|
| 94 |
+
"loss": 0.4035,
|
| 95 |
+
"step": 5500
|
| 96 |
+
},
|
| 97 |
+
{
|
| 98 |
+
"epoch": 0.385214195143091,
|
| 99 |
+
"grad_norm": 6.695476055145264,
|
| 100 |
+
"learning_rate": 4.870081300813009e-05,
|
| 101 |
+
"loss": 0.3882,
|
| 102 |
+
"step": 6000
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 0.41731537807168195,
|
| 106 |
+
"grad_norm": 14.509389877319336,
|
| 107 |
+
"learning_rate": 4.853853658536586e-05,
|
| 108 |
+
"loss": 0.3574,
|
| 109 |
+
"step": 6500
|
| 110 |
+
},
|
| 111 |
+
{
|
| 112 |
+
"epoch": 0.4494165610002729,
|
| 113 |
+
"grad_norm": 38.900367736816406,
|
| 114 |
+
"learning_rate": 4.8375934959349596e-05,
|
| 115 |
+
"loss": 0.3723,
|
| 116 |
+
"step": 7000
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.48151774392886376,
|
| 120 |
+
"grad_norm": 3.945736885070801,
|
| 121 |
+
"learning_rate": 4.8213333333333335e-05,
|
| 122 |
+
"loss": 0.3586,
|
| 123 |
+
"step": 7500
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 0.5136189268574547,
|
| 127 |
+
"grad_norm": 77.0046615600586,
|
| 128 |
+
"learning_rate": 4.8050731707317074e-05,
|
| 129 |
+
"loss": 0.3679,
|
| 130 |
+
"step": 8000
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 0.5457201097860456,
|
| 134 |
+
"grad_norm": 16.720748901367188,
|
| 135 |
+
"learning_rate": 4.788813008130081e-05,
|
| 136 |
+
"loss": 0.3352,
|
| 137 |
+
"step": 8500
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 0.5778212927146366,
|
| 141 |
+
"grad_norm": 17.356231689453125,
|
| 142 |
+
"learning_rate": 4.772552845528456e-05,
|
| 143 |
+
"loss": 0.341,
|
| 144 |
+
"step": 9000
|
| 145 |
+
},
|
| 146 |
+
{
|
| 147 |
+
"epoch": 0.6099224756432274,
|
| 148 |
+
"grad_norm": 9.849752426147461,
|
| 149 |
+
"learning_rate": 4.75629268292683e-05,
|
| 150 |
+
"loss": 0.3354,
|
| 151 |
+
"step": 9500
|
| 152 |
+
},
|
| 153 |
+
{
|
| 154 |
+
"epoch": 0.6420236585718184,
|
| 155 |
+
"grad_norm": 9.619921684265137,
|
| 156 |
+
"learning_rate": 4.7400650406504066e-05,
|
| 157 |
+
"loss": 0.3288,
|
| 158 |
+
"step": 10000
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.6420236585718184,
|
| 162 |
+
"eval_loss": 0.35230785608291626,
|
| 163 |
+
"eval_runtime": 596.403,
|
| 164 |
+
"eval_samples_per_second": 11.621,
|
| 165 |
+
"eval_steps_per_second": 11.621,
|
| 166 |
+
"eval_wer": 0.18700615692192107,
|
| 167 |
+
"step": 10000
|
| 168 |
+
},
|
| 169 |
+
{
|
| 170 |
+
"epoch": 0.6741248415004093,
|
| 171 |
+
"grad_norm": 7.539575099945068,
|
| 172 |
+
"learning_rate": 4.723804878048781e-05,
|
| 173 |
+
"loss": 0.3314,
|
| 174 |
+
"step": 10500
|
| 175 |
+
},
|
| 176 |
+
{
|
| 177 |
+
"epoch": 0.7062260244290002,
|
| 178 |
+
"grad_norm": 13.222583770751953,
|
| 179 |
+
"learning_rate": 4.7075447154471544e-05,
|
| 180 |
+
"loss": 0.3383,
|
| 181 |
+
"step": 11000
|
| 182 |
+
},
|
| 183 |
+
{
|
| 184 |
+
"epoch": 0.7383272073575912,
|
| 185 |
+
"grad_norm": 2.821357011795044,
|
| 186 |
+
"learning_rate": 4.691284552845528e-05,
|
| 187 |
+
"loss": 0.3319,
|
| 188 |
+
"step": 11500
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.770428390286182,
|
| 192 |
+
"grad_norm": 7.006532669067383,
|
| 193 |
+
"learning_rate": 4.675024390243903e-05,
|
| 194 |
+
"loss": 0.3218,
|
| 195 |
+
"step": 12000
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 0.8025295732147729,
|
| 199 |
+
"grad_norm": 6.498229503631592,
|
| 200 |
+
"learning_rate": 4.658764227642277e-05,
|
| 201 |
+
"loss": 0.3175,
|
| 202 |
+
"step": 12500
|
| 203 |
+
},
|
| 204 |
+
{
|
| 205 |
+
"epoch": 0.8346307561433639,
|
| 206 |
+
"grad_norm": 8.40401840209961,
|
| 207 |
+
"learning_rate": 4.6425365853658536e-05,
|
| 208 |
+
"loss": 0.3083,
|
| 209 |
+
"step": 13000
|
| 210 |
+
},
|
| 211 |
+
{
|
| 212 |
+
"epoch": 0.8667319390719548,
|
| 213 |
+
"grad_norm": 1.3094803094863892,
|
| 214 |
+
"learning_rate": 4.626276422764228e-05,
|
| 215 |
+
"loss": 0.3109,
|
| 216 |
+
"step": 13500
|
| 217 |
+
},
|
| 218 |
+
{
|
| 219 |
+
"epoch": 0.8988331220005458,
|
| 220 |
+
"grad_norm": 3.827925443649292,
|
| 221 |
+
"learning_rate": 4.610016260162602e-05,
|
| 222 |
+
"loss": 0.3203,
|
| 223 |
+
"step": 14000
|
| 224 |
+
},
|
| 225 |
+
{
|
| 226 |
+
"epoch": 0.9309343049291366,
|
| 227 |
+
"grad_norm": 2.329725503921509,
|
| 228 |
+
"learning_rate": 4.593756097560975e-05,
|
| 229 |
+
"loss": 0.2979,
|
| 230 |
+
"step": 14500
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.9630354878577275,
|
| 234 |
+
"grad_norm": 3.5614166259765625,
|
| 235 |
+
"learning_rate": 4.57749593495935e-05,
|
| 236 |
+
"loss": 0.3024,
|
| 237 |
+
"step": 15000
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 0.9630354878577275,
|
| 241 |
+
"eval_loss": 0.30602961778640747,
|
| 242 |
+
"eval_runtime": 592.7853,
|
| 243 |
+
"eval_samples_per_second": 11.692,
|
| 244 |
+
"eval_steps_per_second": 11.692,
|
| 245 |
+
"eval_wer": 0.16886647692905213,
|
| 246 |
+
"step": 15000
|
| 247 |
+
},
|
| 248 |
+
{
|
| 249 |
+
"epoch": 0.9951366707863185,
|
| 250 |
+
"grad_norm": 87.19066619873047,
|
| 251 |
+
"learning_rate": 4.561235772357724e-05,
|
| 252 |
+
"loss": 0.3088,
|
| 253 |
+
"step": 15500
|
| 254 |
+
},
|
| 255 |
+
{
|
| 256 |
+
"epoch": 1.0272378537149094,
|
| 257 |
+
"grad_norm": 2.4488048553466797,
|
| 258 |
+
"learning_rate": 4.544975609756098e-05,
|
| 259 |
+
"loss": 0.2761,
|
| 260 |
+
"step": 16000
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 1.0593390366435003,
|
| 264 |
+
"grad_norm": 2.8186426162719727,
|
| 265 |
+
"learning_rate": 4.5287154471544715e-05,
|
| 266 |
+
"loss": 0.2708,
|
| 267 |
+
"step": 16500
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 1.0914402195720911,
|
| 271 |
+
"grad_norm": 1.2367055416107178,
|
| 272 |
+
"learning_rate": 4.5124552845528454e-05,
|
| 273 |
+
"loss": 0.2814,
|
| 274 |
+
"step": 17000
|
| 275 |
+
},
|
| 276 |
+
{
|
| 277 |
+
"epoch": 1.1235414025006822,
|
| 278 |
+
"grad_norm": 1.350964903831482,
|
| 279 |
+
"learning_rate": 4.496227642276423e-05,
|
| 280 |
+
"loss": 0.2943,
|
| 281 |
+
"step": 17500
|
| 282 |
+
},
|
| 283 |
+
{
|
| 284 |
+
"epoch": 1.1556425854292731,
|
| 285 |
+
"grad_norm": 3.4485456943511963,
|
| 286 |
+
"learning_rate": 4.479967479674797e-05,
|
| 287 |
+
"loss": 0.2814,
|
| 288 |
+
"step": 18000
|
| 289 |
+
},
|
| 290 |
+
{
|
| 291 |
+
"epoch": 1.187743768357864,
|
| 292 |
+
"grad_norm": 2.5614452362060547,
|
| 293 |
+
"learning_rate": 4.463707317073171e-05,
|
| 294 |
+
"loss": 0.2624,
|
| 295 |
+
"step": 18500
|
| 296 |
+
},
|
| 297 |
+
{
|
| 298 |
+
"epoch": 1.2198449512864549,
|
| 299 |
+
"grad_norm": 5.0949625968933105,
|
| 300 |
+
"learning_rate": 4.447447154471545e-05,
|
| 301 |
+
"loss": 0.2808,
|
| 302 |
+
"step": 19000
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 1.2519461342150457,
|
| 306 |
+
"grad_norm": 1.0048319101333618,
|
| 307 |
+
"learning_rate": 4.431186991869919e-05,
|
| 308 |
+
"loss": 0.2534,
|
| 309 |
+
"step": 19500
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 1.2840473171436368,
|
| 313 |
+
"grad_norm": 1.9653562307357788,
|
| 314 |
+
"learning_rate": 4.4149268292682924e-05,
|
| 315 |
+
"loss": 0.2465,
|
| 316 |
+
"step": 20000
|
| 317 |
+
},
|
| 318 |
+
{
|
| 319 |
+
"epoch": 1.2840473171436368,
|
| 320 |
+
"eval_loss": 0.28473955392837524,
|
| 321 |
+
"eval_runtime": 595.99,
|
| 322 |
+
"eval_samples_per_second": 11.629,
|
| 323 |
+
"eval_steps_per_second": 11.629,
|
| 324 |
+
"eval_wer": 0.14723766227978022,
|
| 325 |
+
"step": 20000
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 1.3161485000722277,
|
| 329 |
+
"grad_norm": 0.9238675832748413,
|
| 330 |
+
"learning_rate": 4.3987317073170736e-05,
|
| 331 |
+
"loss": 0.2812,
|
| 332 |
+
"step": 20500
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 1.3482496830008186,
|
| 336 |
+
"grad_norm": 1.7164336442947388,
|
| 337 |
+
"learning_rate": 4.3824715447154475e-05,
|
| 338 |
+
"loss": 0.2716,
|
| 339 |
+
"step": 21000
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 1.3803508659294095,
|
| 343 |
+
"grad_norm": 1.9793908596038818,
|
| 344 |
+
"learning_rate": 4.3662113821138214e-05,
|
| 345 |
+
"loss": 0.2779,
|
| 346 |
+
"step": 21500
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 1.4124520488580004,
|
| 350 |
+
"grad_norm": 2.866766929626465,
|
| 351 |
+
"learning_rate": 4.349951219512195e-05,
|
| 352 |
+
"loss": 0.2467,
|
| 353 |
+
"step": 22000
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 1.4445532317865912,
|
| 357 |
+
"grad_norm": 1.6922576427459717,
|
| 358 |
+
"learning_rate": 4.333691056910569e-05,
|
| 359 |
+
"loss": 0.2616,
|
| 360 |
+
"step": 22500
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 1.4766544147151823,
|
| 364 |
+
"grad_norm": 2.3735485076904297,
|
| 365 |
+
"learning_rate": 4.317430894308943e-05,
|
| 366 |
+
"loss": 0.273,
|
| 367 |
+
"step": 23000
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 1.5087555976437732,
|
| 371 |
+
"grad_norm": 1.882097601890564,
|
| 372 |
+
"learning_rate": 4.3012032520325206e-05,
|
| 373 |
+
"loss": 0.2541,
|
| 374 |
+
"step": 23500
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 1.540856780572364,
|
| 378 |
+
"grad_norm": 1.2024834156036377,
|
| 379 |
+
"learning_rate": 4.2849430894308945e-05,
|
| 380 |
+
"loss": 0.2679,
|
| 381 |
+
"step": 24000
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 1.572957963500955,
|
| 385 |
+
"grad_norm": 1.467214584350586,
|
| 386 |
+
"learning_rate": 4.2686829268292684e-05,
|
| 387 |
+
"loss": 0.2788,
|
| 388 |
+
"step": 24500
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 1.605059146429546,
|
| 392 |
+
"grad_norm": 0.9247603416442871,
|
| 393 |
+
"learning_rate": 4.252422764227643e-05,
|
| 394 |
+
"loss": 0.2575,
|
| 395 |
+
"step": 25000
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 1.605059146429546,
|
| 399 |
+
"eval_loss": 0.26871952414512634,
|
| 400 |
+
"eval_runtime": 596.0677,
|
| 401 |
+
"eval_samples_per_second": 11.628,
|
| 402 |
+
"eval_steps_per_second": 11.628,
|
| 403 |
+
"eval_wer": 0.13321745331372287,
|
| 404 |
+
"step": 25000
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 1.637160329358137,
|
| 408 |
+
"grad_norm": 1.6064761877059937,
|
| 409 |
+
"learning_rate": 4.236162601626016e-05,
|
| 410 |
+
"loss": 0.2325,
|
| 411 |
+
"step": 25500
|
| 412 |
+
},
|
| 413 |
+
{
|
| 414 |
+
"epoch": 1.6692615122867278,
|
| 415 |
+
"grad_norm": 1.4238929748535156,
|
| 416 |
+
"learning_rate": 4.21990243902439e-05,
|
| 417 |
+
"loss": 0.2463,
|
| 418 |
+
"step": 26000
|
| 419 |
+
},
|
| 420 |
+
{
|
| 421 |
+
"epoch": 1.7013626952153187,
|
| 422 |
+
"grad_norm": 0.7233666181564331,
|
| 423 |
+
"learning_rate": 4.2036422764227646e-05,
|
| 424 |
+
"loss": 0.2537,
|
| 425 |
+
"step": 26500
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 1.7334638781439096,
|
| 429 |
+
"grad_norm": 0.9025558233261108,
|
| 430 |
+
"learning_rate": 4.1873821138211385e-05,
|
| 431 |
+
"loss": 0.2524,
|
| 432 |
+
"step": 27000
|
| 433 |
+
},
|
| 434 |
+
{
|
| 435 |
+
"epoch": 1.7655650610725004,
|
| 436 |
+
"grad_norm": 2.8646883964538574,
|
| 437 |
+
"learning_rate": 4.1711219512195124e-05,
|
| 438 |
+
"loss": 0.2536,
|
| 439 |
+
"step": 27500
|
| 440 |
+
},
|
| 441 |
+
{
|
| 442 |
+
"epoch": 1.7976662440010913,
|
| 443 |
+
"grad_norm": 1.5429887771606445,
|
| 444 |
+
"learning_rate": 4.154861788617886e-05,
|
| 445 |
+
"loss": 0.2664,
|
| 446 |
+
"step": 28000
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 1.8297674269296822,
|
| 450 |
+
"grad_norm": 2.8908677101135254,
|
| 451 |
+
"learning_rate": 4.138634146341464e-05,
|
| 452 |
+
"loss": 0.2636,
|
| 453 |
+
"step": 28500
|
| 454 |
+
},
|
| 455 |
+
{
|
| 456 |
+
"epoch": 1.8618686098582733,
|
| 457 |
+
"grad_norm": 2.929805278778076,
|
| 458 |
+
"learning_rate": 4.122373983739838e-05,
|
| 459 |
+
"loss": 0.257,
|
| 460 |
+
"step": 29000
|
| 461 |
+
},
|
| 462 |
+
{
|
| 463 |
+
"epoch": 1.8939697927868642,
|
| 464 |
+
"grad_norm": 1.2041730880737305,
|
| 465 |
+
"learning_rate": 4.1061138211382116e-05,
|
| 466 |
+
"loss": 0.2459,
|
| 467 |
+
"step": 29500
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"epoch": 1.9260709757154553,
|
| 471 |
+
"grad_norm": 1.7089420557022095,
|
| 472 |
+
"learning_rate": 4.0898536585365855e-05,
|
| 473 |
+
"loss": 0.2608,
|
| 474 |
+
"step": 30000
|
| 475 |
+
},
|
| 476 |
+
{
|
| 477 |
+
"epoch": 1.9260709757154553,
|
| 478 |
+
"eval_loss": 0.2512572705745697,
|
| 479 |
+
"eval_runtime": 594.6622,
|
| 480 |
+
"eval_samples_per_second": 11.655,
|
| 481 |
+
"eval_steps_per_second": 11.655,
|
| 482 |
+
"eval_wer": 0.1310080925002706,
|
| 483 |
+
"step": 30000
|
| 484 |
+
},
|
| 485 |
+
{
|
| 486 |
+
"epoch": 1.9581721586440461,
|
| 487 |
+
"grad_norm": 2.743481397628784,
|
| 488 |
+
"learning_rate": 4.073626016260163e-05,
|
| 489 |
+
"loss": 0.2381,
|
| 490 |
+
"step": 30500
|
| 491 |
+
},
|
| 492 |
+
{
|
| 493 |
+
"epoch": 1.990273341572637,
|
| 494 |
+
"grad_norm": 0.6640022993087769,
|
| 495 |
+
"learning_rate": 4.05739837398374e-05,
|
| 496 |
+
"loss": 0.2399,
|
| 497 |
+
"step": 31000
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 2.022374524501228,
|
| 501 |
+
"grad_norm": 1.1567021608352661,
|
| 502 |
+
"learning_rate": 4.041138211382114e-05,
|
| 503 |
+
"loss": 0.2238,
|
| 504 |
+
"step": 31500
|
| 505 |
+
},
|
| 506 |
+
{
|
| 507 |
+
"epoch": 2.0544757074298188,
|
| 508 |
+
"grad_norm": 1.3281149864196777,
|
| 509 |
+
"learning_rate": 4.0248780487804884e-05,
|
| 510 |
+
"loss": 0.2302,
|
| 511 |
+
"step": 32000
|
| 512 |
+
},
|
| 513 |
+
{
|
| 514 |
+
"epoch": 2.0865768903584097,
|
| 515 |
+
"grad_norm": 3.654982328414917,
|
| 516 |
+
"learning_rate": 4.008617886178862e-05,
|
| 517 |
+
"loss": 0.2195,
|
| 518 |
+
"step": 32500
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 2.1186780732870005,
|
| 522 |
+
"grad_norm": 1.8765735626220703,
|
| 523 |
+
"learning_rate": 3.9923577235772355e-05,
|
| 524 |
+
"loss": 0.2375,
|
| 525 |
+
"step": 33000
|
| 526 |
+
},
|
| 527 |
+
{
|
| 528 |
+
"epoch": 2.1507792562155914,
|
| 529 |
+
"grad_norm": 1.4444390535354614,
|
| 530 |
+
"learning_rate": 3.97609756097561e-05,
|
| 531 |
+
"loss": 0.2172,
|
| 532 |
+
"step": 33500
|
| 533 |
+
},
|
| 534 |
+
{
|
| 535 |
+
"epoch": 2.1828804391441823,
|
| 536 |
+
"grad_norm": 0.9096426367759705,
|
| 537 |
+
"learning_rate": 3.959837398373984e-05,
|
| 538 |
+
"loss": 0.2258,
|
| 539 |
+
"step": 34000
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 2.2149816220727736,
|
| 543 |
+
"grad_norm": 1.1495827436447144,
|
| 544 |
+
"learning_rate": 3.943577235772358e-05,
|
| 545 |
+
"loss": 0.223,
|
| 546 |
+
"step": 34500
|
| 547 |
+
},
|
| 548 |
+
{
|
| 549 |
+
"epoch": 2.2470828050013645,
|
| 550 |
+
"grad_norm": 4.880052089691162,
|
| 551 |
+
"learning_rate": 3.9273495934959354e-05,
|
| 552 |
+
"loss": 0.2465,
|
| 553 |
+
"step": 35000
|
| 554 |
+
},
|
| 555 |
+
{
|
| 556 |
+
"epoch": 2.2470828050013645,
|
| 557 |
+
"eval_loss": 0.2504861056804657,
|
| 558 |
+
"eval_runtime": 588.4465,
|
| 559 |
+
"eval_samples_per_second": 11.778,
|
| 560 |
+
"eval_steps_per_second": 11.778,
|
| 561 |
+
"eval_wer": 0.12438637709395832,
|
| 562 |
+
"step": 35000
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 2.2791839879299554,
|
| 566 |
+
"grad_norm": 0.8823996782302856,
|
| 567 |
+
"learning_rate": 3.911089430894309e-05,
|
| 568 |
+
"loss": 0.2123,
|
| 569 |
+
"step": 35500
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 2.3112851708585462,
|
| 573 |
+
"grad_norm": 0.8461039066314697,
|
| 574 |
+
"learning_rate": 3.894861788617886e-05,
|
| 575 |
+
"loss": 0.2255,
|
| 576 |
+
"step": 36000
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 2.343386353787137,
|
| 580 |
+
"grad_norm": 1.3658907413482666,
|
| 581 |
+
"learning_rate": 3.878601626016261e-05,
|
| 582 |
+
"loss": 0.2166,
|
| 583 |
+
"step": 36500
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 2.375487536715728,
|
| 587 |
+
"grad_norm": 5.125720500946045,
|
| 588 |
+
"learning_rate": 3.8623414634146346e-05,
|
| 589 |
+
"loss": 0.2174,
|
| 590 |
+
"step": 37000
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 2.407588719644319,
|
| 594 |
+
"grad_norm": 0.960725724697113,
|
| 595 |
+
"learning_rate": 3.846081300813008e-05,
|
| 596 |
+
"loss": 0.2148,
|
| 597 |
+
"step": 37500
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 2.4396899025729097,
|
| 601 |
+
"grad_norm": 3.487365484237671,
|
| 602 |
+
"learning_rate": 3.829853658536586e-05,
|
| 603 |
+
"loss": 0.225,
|
| 604 |
+
"step": 38000
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 2.4717910855015006,
|
| 608 |
+
"grad_norm": 2.1244943141937256,
|
| 609 |
+
"learning_rate": 3.813593495934959e-05,
|
| 610 |
+
"loss": 0.2237,
|
| 611 |
+
"step": 38500
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 2.5038922684300915,
|
| 615 |
+
"grad_norm": 128.5044708251953,
|
| 616 |
+
"learning_rate": 3.797333333333333e-05,
|
| 617 |
+
"loss": 0.2298,
|
| 618 |
+
"step": 39000
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 2.5359934513586824,
|
| 622 |
+
"grad_norm": 2.0013089179992676,
|
| 623 |
+
"learning_rate": 3.781073170731708e-05,
|
| 624 |
+
"loss": 0.2184,
|
| 625 |
+
"step": 39500
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 2.5680946342872737,
|
| 629 |
+
"grad_norm": 4.2432541847229,
|
| 630 |
+
"learning_rate": 3.7648130081300816e-05,
|
| 631 |
+
"loss": 0.211,
|
| 632 |
+
"step": 40000
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 2.5680946342872737,
|
| 636 |
+
"eval_loss": 0.2507112920284271,
|
| 637 |
+
"eval_runtime": 589.626,
|
| 638 |
+
"eval_samples_per_second": 11.755,
|
| 639 |
+
"eval_steps_per_second": 11.755,
|
| 640 |
+
"eval_wer": 0.12312570435314117,
|
| 641 |
+
"step": 40000
|
| 642 |
+
},
|
| 643 |
+
{
|
| 644 |
+
"epoch": 2.6001958172158646,
|
| 645 |
+
"grad_norm": 7.432988166809082,
|
| 646 |
+
"learning_rate": 3.7485528455284555e-05,
|
| 647 |
+
"loss": 0.2079,
|
| 648 |
+
"step": 40500
|
| 649 |
+
},
|
| 650 |
+
{
|
| 651 |
+
"epoch": 2.6322970001444554,
|
| 652 |
+
"grad_norm": 0.998031735420227,
|
| 653 |
+
"learning_rate": 3.732325203252033e-05,
|
| 654 |
+
"loss": 0.2193,
|
| 655 |
+
"step": 41000
|
| 656 |
+
},
|
| 657 |
+
{
|
| 658 |
+
"epoch": 2.6643981830730463,
|
| 659 |
+
"grad_norm": 0.9233012199401855,
|
| 660 |
+
"learning_rate": 3.716065040650407e-05,
|
| 661 |
+
"loss": 0.2219,
|
| 662 |
+
"step": 41500
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 2.696499366001637,
|
| 666 |
+
"grad_norm": 2.8505234718322754,
|
| 667 |
+
"learning_rate": 3.69980487804878e-05,
|
| 668 |
+
"loss": 0.2117,
|
| 669 |
+
"step": 42000
|
| 670 |
+
},
|
| 671 |
+
{
|
| 672 |
+
"epoch": 2.728600548930228,
|
| 673 |
+
"grad_norm": 5.0437912940979,
|
| 674 |
+
"learning_rate": 3.683544715447155e-05,
|
| 675 |
+
"loss": 0.195,
|
| 676 |
+
"step": 42500
|
| 677 |
+
},
|
| 678 |
+
{
|
| 679 |
+
"epoch": 2.760701731858819,
|
| 680 |
+
"grad_norm": 3.6365315914154053,
|
| 681 |
+
"learning_rate": 3.6672845528455286e-05,
|
| 682 |
+
"loss": 0.2184,
|
| 683 |
+
"step": 43000
|
| 684 |
+
},
|
| 685 |
+
{
|
| 686 |
+
"epoch": 2.79280291478741,
|
| 687 |
+
"grad_norm": 6.058934211730957,
|
| 688 |
+
"learning_rate": 3.651024390243903e-05,
|
| 689 |
+
"loss": 0.2132,
|
| 690 |
+
"step": 43500
|
| 691 |
+
},
|
| 692 |
+
{
|
| 693 |
+
"epoch": 2.8249040977160007,
|
| 694 |
+
"grad_norm": 34.89470291137695,
|
| 695 |
+
"learning_rate": 3.6347642276422764e-05,
|
| 696 |
+
"loss": 0.2241,
|
| 697 |
+
"step": 44000
|
| 698 |
+
},
|
| 699 |
+
{
|
| 700 |
+
"epoch": 2.857005280644592,
|
| 701 |
+
"grad_norm": 1.1790027618408203,
|
| 702 |
+
"learning_rate": 3.61850406504065e-05,
|
| 703 |
+
"loss": 0.2196,
|
| 704 |
+
"step": 44500
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 2.8891064635731825,
|
| 708 |
+
"grad_norm": 1.6652076244354248,
|
| 709 |
+
"learning_rate": 3.602243902439025e-05,
|
| 710 |
+
"loss": 0.2178,
|
| 711 |
+
"step": 45000
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 2.8891064635731825,
|
| 715 |
+
"eval_loss": 0.2256719172000885,
|
| 716 |
+
"eval_runtime": 582.4547,
|
| 717 |
+
"eval_samples_per_second": 11.9,
|
| 718 |
+
"eval_steps_per_second": 11.9,
|
| 719 |
+
"eval_wer": 0.11872608382837023,
|
| 720 |
+
"step": 45000
|
| 721 |
+
},
|
| 722 |
+
{
|
| 723 |
+
"epoch": 2.921207646501774,
|
| 724 |
+
"grad_norm": 1.9215317964553833,
|
| 725 |
+
"learning_rate": 3.585983739837399e-05,
|
| 726 |
+
"loss": 0.2213,
|
| 727 |
+
"step": 45500
|
| 728 |
+
},
|
| 729 |
+
{
|
| 730 |
+
"epoch": 2.9533088294303647,
|
| 731 |
+
"grad_norm": 13.67776107788086,
|
| 732 |
+
"learning_rate": 3.5697560975609756e-05,
|
| 733 |
+
"loss": 0.2233,
|
| 734 |
+
"step": 46000
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 2.9854100123589555,
|
| 738 |
+
"grad_norm": 5.060420989990234,
|
| 739 |
+
"learning_rate": 3.55349593495935e-05,
|
| 740 |
+
"loss": 0.2118,
|
| 741 |
+
"step": 46500
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 3.0175111952875464,
|
| 745 |
+
"grad_norm": 0.8249053955078125,
|
| 746 |
+
"learning_rate": 3.537268292682927e-05,
|
| 747 |
+
"loss": 0.2147,
|
| 748 |
+
"step": 47000
|
| 749 |
+
},
|
| 750 |
+
{
|
| 751 |
+
"epoch": 3.0496123782161373,
|
| 752 |
+
"grad_norm": 0.7037186622619629,
|
| 753 |
+
"learning_rate": 3.521008130081301e-05,
|
| 754 |
+
"loss": 0.1941,
|
| 755 |
+
"step": 47500
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"epoch": 3.081713561144728,
|
| 759 |
+
"grad_norm": 0.7166640758514404,
|
| 760 |
+
"learning_rate": 3.5047479674796755e-05,
|
| 761 |
+
"loss": 0.1903,
|
| 762 |
+
"step": 48000
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 3.113814744073319,
|
| 766 |
+
"grad_norm": 0.8476791977882385,
|
| 767 |
+
"learning_rate": 3.488487804878049e-05,
|
| 768 |
+
"loss": 0.2004,
|
| 769 |
+
"step": 48500
|
| 770 |
+
},
|
| 771 |
+
{
|
| 772 |
+
"epoch": 3.14591592700191,
|
| 773 |
+
"grad_norm": 0.4231201410293579,
|
| 774 |
+
"learning_rate": 3.472260162601626e-05,
|
| 775 |
+
"loss": 0.1894,
|
| 776 |
+
"step": 49000
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 3.178017109930501,
|
| 780 |
+
"grad_norm": 0.40067389607429504,
|
| 781 |
+
"learning_rate": 3.456e-05,
|
| 782 |
+
"loss": 0.1773,
|
| 783 |
+
"step": 49500
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 3.2101182928590917,
|
| 787 |
+
"grad_norm": 0.7153622508049011,
|
| 788 |
+
"learning_rate": 3.439739837398374e-05,
|
| 789 |
+
"loss": 0.2046,
|
| 790 |
+
"step": 50000
|
| 791 |
+
},
|
| 792 |
+
{
|
| 793 |
+
"epoch": 3.2101182928590917,
|
| 794 |
+
"eval_loss": 0.22042368352413177,
|
| 795 |
+
"eval_runtime": 575.7584,
|
| 796 |
+
"eval_samples_per_second": 12.038,
|
| 797 |
+
"eval_steps_per_second": 12.038,
|
| 798 |
+
"eval_wer": 0.11246728936259622,
|
| 799 |
+
"step": 50000
|
| 800 |
+
},
|
| 801 |
+
{
|
| 802 |
+
"epoch": 3.2422194757876825,
|
| 803 |
+
"grad_norm": 2.9656982421875,
|
| 804 |
+
"learning_rate": 3.4235121951219515e-05,
|
| 805 |
+
"loss": 0.1929,
|
| 806 |
+
"step": 50500
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 3.274320658716274,
|
| 810 |
+
"grad_norm": 1.7078142166137695,
|
| 811 |
+
"learning_rate": 3.4072520325203254e-05,
|
| 812 |
+
"loss": 0.193,
|
| 813 |
+
"step": 51000
|
| 814 |
+
},
|
| 815 |
+
{
|
| 816 |
+
"epoch": 3.3064218416448647,
|
| 817 |
+
"grad_norm": 7.298386573791504,
|
| 818 |
+
"learning_rate": 3.390991869918699e-05,
|
| 819 |
+
"loss": 0.1933,
|
| 820 |
+
"step": 51500
|
| 821 |
+
},
|
| 822 |
+
{
|
| 823 |
+
"epoch": 3.3385230245734556,
|
| 824 |
+
"grad_norm": 0.4513390064239502,
|
| 825 |
+
"learning_rate": 3.374731707317073e-05,
|
| 826 |
+
"loss": 0.2013,
|
| 827 |
+
"step": 52000
|
| 828 |
+
},
|
| 829 |
+
{
|
| 830 |
+
"epoch": 3.3706242075020465,
|
| 831 |
+
"grad_norm": 0.4610413610935211,
|
| 832 |
+
"learning_rate": 3.358471544715447e-05,
|
| 833 |
+
"loss": 0.1828,
|
| 834 |
+
"step": 52500
|
| 835 |
+
},
|
| 836 |
+
{
|
| 837 |
+
"epoch": 3.4027253904306374,
|
| 838 |
+
"grad_norm": 0.921280026435852,
|
| 839 |
+
"learning_rate": 3.342211382113821e-05,
|
| 840 |
+
"loss": 0.1897,
|
| 841 |
+
"step": 53000
|
| 842 |
+
},
|
| 843 |
+
{
|
| 844 |
+
"epoch": 3.4348265733592283,
|
| 845 |
+
"grad_norm": 0.9576189517974854,
|
| 846 |
+
"learning_rate": 3.325951219512195e-05,
|
| 847 |
+
"loss": 0.184,
|
| 848 |
+
"step": 53500
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 3.466927756287819,
|
| 852 |
+
"grad_norm": 0.47692379355430603,
|
| 853 |
+
"learning_rate": 3.3096910569105695e-05,
|
| 854 |
+
"loss": 0.1904,
|
| 855 |
+
"step": 54000
|
| 856 |
+
},
|
| 857 |
+
{
|
| 858 |
+
"epoch": 3.49902893921641,
|
| 859 |
+
"grad_norm": 1.5692068338394165,
|
| 860 |
+
"learning_rate": 3.2934308943089434e-05,
|
| 861 |
+
"loss": 0.1905,
|
| 862 |
+
"step": 54500
|
| 863 |
+
},
|
| 864 |
+
{
|
| 865 |
+
"epoch": 3.531130122145001,
|
| 866 |
+
"grad_norm": 6.266742706298828,
|
| 867 |
+
"learning_rate": 3.277170731707317e-05,
|
| 868 |
+
"loss": 0.1887,
|
| 869 |
+
"step": 55000
|
| 870 |
+
},
|
| 871 |
+
{
|
| 872 |
+
"epoch": 3.531130122145001,
|
| 873 |
+
"eval_loss": 0.2165077179670334,
|
| 874 |
+
"eval_runtime": 597.4689,
|
| 875 |
+
"eval_samples_per_second": 11.601,
|
| 876 |
+
"eval_steps_per_second": 11.601,
|
| 877 |
+
"eval_wer": 0.11000324718736271,
|
| 878 |
+
"step": 55000
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 3.563231305073592,
|
| 882 |
+
"grad_norm": 1.031356692314148,
|
| 883 |
+
"learning_rate": 3.260910569105691e-05,
|
| 884 |
+
"loss": 0.1883,
|
| 885 |
+
"step": 55500
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 3.5953324880021826,
|
| 889 |
+
"grad_norm": 1.9417179822921753,
|
| 890 |
+
"learning_rate": 3.244682926829268e-05,
|
| 891 |
+
"loss": 0.1956,
|
| 892 |
+
"step": 56000
|
| 893 |
+
},
|
| 894 |
+
{
|
| 895 |
+
"epoch": 3.627433670930774,
|
| 896 |
+
"grad_norm": 0.7637033462524414,
|
| 897 |
+
"learning_rate": 3.2284227642276426e-05,
|
| 898 |
+
"loss": 0.1877,
|
| 899 |
+
"step": 56500
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 3.659534853859365,
|
| 903 |
+
"grad_norm": 0.5973896384239197,
|
| 904 |
+
"learning_rate": 3.2121626016260165e-05,
|
| 905 |
+
"loss": 0.1756,
|
| 906 |
+
"step": 57000
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"epoch": 3.6916360367879557,
|
| 910 |
+
"grad_norm": 1.629115343093872,
|
| 911 |
+
"learning_rate": 3.1959024390243904e-05,
|
| 912 |
+
"loss": 0.1867,
|
| 913 |
+
"step": 57500
|
| 914 |
+
},
|
| 915 |
+
{
|
| 916 |
+
"epoch": 3.7237372197165466,
|
| 917 |
+
"grad_norm": 0.4055944085121155,
|
| 918 |
+
"learning_rate": 3.179642276422764e-05,
|
| 919 |
+
"loss": 0.1766,
|
| 920 |
+
"step": 58000
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 3.7558384026451375,
|
| 924 |
+
"grad_norm": 3.1317808628082275,
|
| 925 |
+
"learning_rate": 3.163382113821138e-05,
|
| 926 |
+
"loss": 0.1928,
|
| 927 |
+
"step": 58500
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 3.7879395855737283,
|
| 931 |
+
"grad_norm": 1.9987761974334717,
|
| 932 |
+
"learning_rate": 3.147121951219512e-05,
|
| 933 |
+
"loss": 0.1775,
|
| 934 |
+
"step": 59000
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 3.820040768502319,
|
| 938 |
+
"grad_norm": 1.7839641571044922,
|
| 939 |
+
"learning_rate": 3.1308617886178867e-05,
|
| 940 |
+
"loss": 0.171,
|
| 941 |
+
"step": 59500
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 3.85214195143091,
|
| 945 |
+
"grad_norm": 0.4871620535850525,
|
| 946 |
+
"learning_rate": 3.1146016260162605e-05,
|
| 947 |
+
"loss": 0.1874,
|
| 948 |
+
"step": 60000
|
| 949 |
+
},
|
| 950 |
+
{
|
| 951 |
+
"epoch": 3.85214195143091,
|
| 952 |
+
"eval_loss": 0.20520555973052979,
|
| 953 |
+
"eval_runtime": 598.1599,
|
| 954 |
+
"eval_samples_per_second": 11.587,
|
| 955 |
+
"eval_steps_per_second": 11.587,
|
| 956 |
+
"eval_wer": 0.1052216046199199,
|
| 957 |
+
"step": 60000
|
| 958 |
+
},
|
| 959 |
+
{
|
| 960 |
+
"epoch": 3.884243134359501,
|
| 961 |
+
"grad_norm": 1.415724754333496,
|
| 962 |
+
"learning_rate": 3.098341463414634e-05,
|
| 963 |
+
"loss": 0.1839,
|
| 964 |
+
"step": 60500
|
| 965 |
+
},
|
| 966 |
+
{
|
| 967 |
+
"epoch": 3.9163443172880923,
|
| 968 |
+
"grad_norm": 0.7966309189796448,
|
| 969 |
+
"learning_rate": 3.0820813008130083e-05,
|
| 970 |
+
"loss": 0.1676,
|
| 971 |
+
"step": 61000
|
| 972 |
+
},
|
| 973 |
+
{
|
| 974 |
+
"epoch": 3.9484455002166827,
|
| 975 |
+
"grad_norm": 0.7828078866004944,
|
| 976 |
+
"learning_rate": 3.065821138211382e-05,
|
| 977 |
+
"loss": 0.175,
|
| 978 |
+
"step": 61500
|
| 979 |
+
},
|
| 980 |
+
{
|
| 981 |
+
"epoch": 3.980546683145274,
|
| 982 |
+
"grad_norm": 0.8037332892417908,
|
| 983 |
+
"learning_rate": 3.049593495934959e-05,
|
| 984 |
+
"loss": 0.1993,
|
| 985 |
+
"step": 62000
|
| 986 |
+
},
|
| 987 |
+
{
|
| 988 |
+
"epoch": 4.0126478660738645,
|
| 989 |
+
"grad_norm": 1.3568751811981201,
|
| 990 |
+
"learning_rate": 3.0333333333333337e-05,
|
| 991 |
+
"loss": 0.1747,
|
| 992 |
+
"step": 62500
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 4.044749049002456,
|
| 996 |
+
"grad_norm": 4.724090576171875,
|
| 997 |
+
"learning_rate": 3.0170731707317072e-05,
|
| 998 |
+
"loss": 0.16,
|
| 999 |
+
"step": 63000
|
| 1000 |
+
},
|
| 1001 |
+
{
|
| 1002 |
+
"epoch": 4.076850231931046,
|
| 1003 |
+
"grad_norm": 4.713302135467529,
|
| 1004 |
+
"learning_rate": 3.0008130081300818e-05,
|
| 1005 |
+
"loss": 0.1544,
|
| 1006 |
+
"step": 63500
|
| 1007 |
+
},
|
| 1008 |
+
{
|
| 1009 |
+
"epoch": 4.1089514148596376,
|
| 1010 |
+
"grad_norm": 2.2513558864593506,
|
| 1011 |
+
"learning_rate": 2.9845528455284553e-05,
|
| 1012 |
+
"loss": 0.1673,
|
| 1013 |
+
"step": 64000
|
| 1014 |
+
},
|
| 1015 |
+
{
|
| 1016 |
+
"epoch": 4.141052597788229,
|
| 1017 |
+
"grad_norm": 0.5379019379615784,
|
| 1018 |
+
"learning_rate": 2.9682926829268292e-05,
|
| 1019 |
+
"loss": 0.1598,
|
| 1020 |
+
"step": 64500
|
| 1021 |
+
},
|
| 1022 |
+
{
|
| 1023 |
+
"epoch": 4.173153780716819,
|
| 1024 |
+
"grad_norm": 2.381577968597412,
|
| 1025 |
+
"learning_rate": 2.9520325203252035e-05,
|
| 1026 |
+
"loss": 0.1598,
|
| 1027 |
+
"step": 65000
|
| 1028 |
+
},
|
| 1029 |
+
{
|
| 1030 |
+
"epoch": 4.173153780716819,
|
| 1031 |
+
"eval_loss": 0.20186389982700348,
|
| 1032 |
+
"eval_runtime": 597.2934,
|
| 1033 |
+
"eval_samples_per_second": 11.604,
|
| 1034 |
+
"eval_steps_per_second": 11.604,
|
| 1035 |
+
"eval_wer": 0.10251561515099421,
|
| 1036 |
+
"step": 65000
|
| 1037 |
+
},
|
| 1038 |
+
{
|
| 1039 |
+
"epoch": 4.205254963645411,
|
| 1040 |
+
"grad_norm": 0.5360216498374939,
|
| 1041 |
+
"learning_rate": 2.9357723577235774e-05,
|
| 1042 |
+
"loss": 0.1459,
|
| 1043 |
+
"step": 65500
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"epoch": 4.237356146574001,
|
| 1047 |
+
"grad_norm": 0.7315216064453125,
|
| 1048 |
+
"learning_rate": 2.9195121951219513e-05,
|
| 1049 |
+
"loss": 0.1702,
|
| 1050 |
+
"step": 66000
|
| 1051 |
+
},
|
| 1052 |
+
{
|
| 1053 |
+
"epoch": 4.269457329502592,
|
| 1054 |
+
"grad_norm": 1.0547746419906616,
|
| 1055 |
+
"learning_rate": 2.9032520325203255e-05,
|
| 1056 |
+
"loss": 0.1632,
|
| 1057 |
+
"step": 66500
|
| 1058 |
+
},
|
| 1059 |
+
{
|
| 1060 |
+
"epoch": 4.301558512431183,
|
| 1061 |
+
"grad_norm": 2.5075976848602295,
|
| 1062 |
+
"learning_rate": 2.8869918699186994e-05,
|
| 1063 |
+
"loss": 0.1595,
|
| 1064 |
+
"step": 67000
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 4.333659695359774,
|
| 1068 |
+
"grad_norm": 1.5994772911071777,
|
| 1069 |
+
"learning_rate": 2.870731707317073e-05,
|
| 1070 |
+
"loss": 0.1665,
|
| 1071 |
+
"step": 67500
|
| 1072 |
+
},
|
| 1073 |
+
{
|
| 1074 |
+
"epoch": 4.365760878288365,
|
| 1075 |
+
"grad_norm": 0.7722116112709045,
|
| 1076 |
+
"learning_rate": 2.8544715447154475e-05,
|
| 1077 |
+
"loss": 0.1642,
|
| 1078 |
+
"step": 68000
|
| 1079 |
+
},
|
| 1080 |
+
{
|
| 1081 |
+
"epoch": 4.397862061216956,
|
| 1082 |
+
"grad_norm": 1.0250564813613892,
|
| 1083 |
+
"learning_rate": 2.838211382113821e-05,
|
| 1084 |
+
"loss": 0.1687,
|
| 1085 |
+
"step": 68500
|
| 1086 |
+
},
|
| 1087 |
+
{
|
| 1088 |
+
"epoch": 4.429963244145547,
|
| 1089 |
+
"grad_norm": 0.7502961158752441,
|
| 1090 |
+
"learning_rate": 2.8219837398373983e-05,
|
| 1091 |
+
"loss": 0.15,
|
| 1092 |
+
"step": 69000
|
| 1093 |
+
},
|
| 1094 |
+
{
|
| 1095 |
+
"epoch": 4.462064427074138,
|
| 1096 |
+
"grad_norm": 2.5044519901275635,
|
| 1097 |
+
"learning_rate": 2.8057560975609758e-05,
|
| 1098 |
+
"loss": 0.1748,
|
| 1099 |
+
"step": 69500
|
| 1100 |
+
},
|
| 1101 |
+
{
|
| 1102 |
+
"epoch": 4.494165610002729,
|
| 1103 |
+
"grad_norm": 0.7988201975822449,
|
| 1104 |
+
"learning_rate": 2.7894959349593497e-05,
|
| 1105 |
+
"loss": 0.1565,
|
| 1106 |
+
"step": 70000
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 4.494165610002729,
|
| 1110 |
+
"eval_loss": 0.19880281388759613,
|
| 1111 |
+
"eval_runtime": 595.5489,
|
| 1112 |
+
"eval_samples_per_second": 11.638,
|
| 1113 |
+
"eval_steps_per_second": 11.638,
|
| 1114 |
+
"eval_wer": 0.09975232237566774,
|
| 1115 |
+
"step": 70000
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 4.526266792931319,
|
| 1119 |
+
"grad_norm": 0.47625595331192017,
|
| 1120 |
+
"learning_rate": 2.7732357723577236e-05,
|
| 1121 |
+
"loss": 0.1603,
|
| 1122 |
+
"step": 70500
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 4.558367975859911,
|
| 1126 |
+
"grad_norm": 0.46868839859962463,
|
| 1127 |
+
"learning_rate": 2.7569756097560978e-05,
|
| 1128 |
+
"loss": 0.1623,
|
| 1129 |
+
"step": 71000
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 4.590469158788501,
|
| 1133 |
+
"grad_norm": 7.4738335609436035,
|
| 1134 |
+
"learning_rate": 2.7407154471544717e-05,
|
| 1135 |
+
"loss": 0.152,
|
| 1136 |
+
"step": 71500
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 4.6225703417170925,
|
| 1140 |
+
"grad_norm": 1.6319257020950317,
|
| 1141 |
+
"learning_rate": 2.724455284552846e-05,
|
| 1142 |
+
"loss": 0.1473,
|
| 1143 |
+
"step": 72000
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 4.654671524645683,
|
| 1147 |
+
"grad_norm": 1.34164559841156,
|
| 1148 |
+
"learning_rate": 2.70819512195122e-05,
|
| 1149 |
+
"loss": 0.1428,
|
| 1150 |
+
"step": 72500
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 4.686772707574274,
|
| 1154 |
+
"grad_norm": 0.5560343861579895,
|
| 1155 |
+
"learning_rate": 2.6919349593495934e-05,
|
| 1156 |
+
"loss": 0.1599,
|
| 1157 |
+
"step": 73000
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 4.718873890502865,
|
| 1161 |
+
"grad_norm": 0.8048732280731201,
|
| 1162 |
+
"learning_rate": 2.6757073170731713e-05,
|
| 1163 |
+
"loss": 0.1656,
|
| 1164 |
+
"step": 73500
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 4.750975073431456,
|
| 1168 |
+
"grad_norm": 0.577629804611206,
|
| 1169 |
+
"learning_rate": 2.6594471544715448e-05,
|
| 1170 |
+
"loss": 0.1533,
|
| 1171 |
+
"step": 74000
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 4.783076256360047,
|
| 1175 |
+
"grad_norm": 1.2069358825683594,
|
| 1176 |
+
"learning_rate": 2.6431869918699187e-05,
|
| 1177 |
+
"loss": 0.1411,
|
| 1178 |
+
"step": 74500
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 4.815177439288638,
|
| 1182 |
+
"grad_norm": 2.0387120246887207,
|
| 1183 |
+
"learning_rate": 2.626926829268293e-05,
|
| 1184 |
+
"loss": 0.1498,
|
| 1185 |
+
"step": 75000
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 4.815177439288638,
|
| 1189 |
+
"eval_loss": 0.18954835832118988,
|
| 1190 |
+
"eval_runtime": 595.37,
|
| 1191 |
+
"eval_samples_per_second": 11.642,
|
| 1192 |
+
"eval_steps_per_second": 11.642,
|
| 1193 |
+
"eval_wer": 0.09902011346054668,
|
| 1194 |
+
"step": 75000
|
| 1195 |
+
}
|
| 1196 |
+
],
|
| 1197 |
+
"logging_steps": 500,
|
| 1198 |
+
"max_steps": 155750,
|
| 1199 |
+
"num_input_tokens_seen": 0,
|
| 1200 |
+
"num_train_epochs": 10,
|
| 1201 |
+
"save_steps": 5000,
|
| 1202 |
+
"stateful_callbacks": {
|
| 1203 |
+
"TrainerControl": {
|
| 1204 |
+
"args": {
|
| 1205 |
+
"should_epoch_stop": false,
|
| 1206 |
+
"should_evaluate": false,
|
| 1207 |
+
"should_log": false,
|
| 1208 |
+
"should_save": true,
|
| 1209 |
+
"should_training_stop": false
|
| 1210 |
+
},
|
| 1211 |
+
"attributes": {}
|
| 1212 |
+
}
|
| 1213 |
+
},
|
| 1214 |
+
"total_flos": 6.299548382938926e+19,
|
| 1215 |
+
"train_batch_size": 1,
|
| 1216 |
+
"trial_name": null,
|
| 1217 |
+
"trial_params": null
|
| 1218 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:17e701a2666b059164627cbdcf780a7bc4685100c06c549201703aadc5788ce9
|
| 3 |
+
size 5368
|
vocab.json
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{"ਂ": 1, "ਅ": 2, "ਆ": 3, "ਇ": 4, "ਈ": 5, "ਉ": 6, "ਊ": 7, "ਏ": 8, "ਐ": 9, "ਓ": 10, "ਔ": 11, "ਕ": 12, "ਖ": 13, "ਗ": 14, "ਘ": 15, "ਚ": 16, "ਛ": 17, "ਜ": 18, "ਝ": 19, "ਟ": 20, "ਠ": 21, "ਡ": 22, "ਢ": 23, "ਣ": 24, "ਤ": 25, "ਥ": 26, "ਦ": 27, "ਧ": 28, "ਨ": 29, "ਪ": 30, "ਫ": 31, "ਬ": 32, "ਭ": 33, "ਮ": 34, "ਯ": 35, "ਰ": 36, "ਲ": 37, "ਲ਼": 38, "ਵ": 39, "ਸ਼": 40, "ਸ": 41, "ਹ": 42, "਼": 43, "ਾ": 44, "ਿ": 45, "ੀ": 46, "ੁ": 47, "ੂ": 48, "ੇ": 49, "ੈ": 50, "ੋ": 51, "ੌ": 52, "੍": 53, "ਖ਼": 54, "ਗ਼": 55, "ਜ਼": 56, "ੜ": 57, "ਫ਼": 58, "ੰ": 59, "ੱ": 60, "ੲ": 61, "ੳ": 62, "|": 0, "[UNK]": 63, "[PAD]": 64}
|