Upload trained RAVDESS emotion detection model
Browse files- checkpoint-5500/config.json +139 -0
- checkpoint-5500/model.safetensors +3 -0
- checkpoint-5500/optimizer.pt +3 -0
- checkpoint-5500/preprocessor_config.json +9 -0
- checkpoint-5500/rng_state.pth +3 -0
- checkpoint-5500/scheduler.pt +3 -0
- checkpoint-5500/trainer_state.json +1574 -0
- checkpoint-5500/training_args.bin +3 -0
- checkpoint-6000/config.json +139 -0
- checkpoint-6000/model.safetensors +3 -0
- checkpoint-6000/optimizer.pt +3 -0
- checkpoint-6000/preprocessor_config.json +9 -0
- checkpoint-6000/rng_state.pth +3 -0
- checkpoint-6000/scheduler.pt +3 -0
- checkpoint-6000/trainer_state.json +1714 -0
- checkpoint-6000/training_args.bin +3 -0
- checkpoint-6270/config.json +139 -0
- checkpoint-6270/model.safetensors +3 -0
- checkpoint-6270/optimizer.pt +3 -0
- checkpoint-6270/preprocessor_config.json +9 -0
- checkpoint-6270/rng_state.pth +3 -0
- checkpoint-6270/scheduler.pt +3 -0
- checkpoint-6270/trainer_state.json +1784 -0
- checkpoint-6270/training_args.bin +3 -0
- config.json +139 -0
- model.safetensors +3 -0
- preprocessor_config.json +9 -0
- special_tokens_map.json +6 -0
- tokenizer_config.json +51 -0
- training_args.bin +3 -0
- vocab.json +34 -0
checkpoint-5500/config.json
ADDED
|
@@ -0,0 +1,139 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"activation_dropout": 0.0,
|
| 3 |
+
"adapter_attn_dim": null,
|
| 4 |
+
"adapter_kernel_size": 3,
|
| 5 |
+
"adapter_stride": 2,
|
| 6 |
+
"add_adapter": false,
|
| 7 |
+
"apply_spec_augment": true,
|
| 8 |
+
"architectures": [
|
| 9 |
+
"Wav2Vec2ForSequenceClassification"
|
| 10 |
+
],
|
| 11 |
+
"attention_dropout": 0.1,
|
| 12 |
+
"bos_token_id": 1,
|
| 13 |
+
"classifier_proj_size": 256,
|
| 14 |
+
"codevector_dim": 256,
|
| 15 |
+
"contrastive_logits_temperature": 0.1,
|
| 16 |
+
"conv_bias": false,
|
| 17 |
+
"conv_dim": [
|
| 18 |
+
512,
|
| 19 |
+
512,
|
| 20 |
+
512,
|
| 21 |
+
512,
|
| 22 |
+
512,
|
| 23 |
+
512,
|
| 24 |
+
512
|
| 25 |
+
],
|
| 26 |
+
"conv_kernel": [
|
| 27 |
+
10,
|
| 28 |
+
3,
|
| 29 |
+
3,
|
| 30 |
+
3,
|
| 31 |
+
3,
|
| 32 |
+
2,
|
| 33 |
+
2
|
| 34 |
+
],
|
| 35 |
+
"conv_stride": [
|
| 36 |
+
5,
|
| 37 |
+
2,
|
| 38 |
+
2,
|
| 39 |
+
2,
|
| 40 |
+
2,
|
| 41 |
+
2,
|
| 42 |
+
2
|
| 43 |
+
],
|
| 44 |
+
"ctc_loss_reduction": "sum",
|
| 45 |
+
"ctc_zero_infinity": false,
|
| 46 |
+
"diversity_loss_weight": 0.1,
|
| 47 |
+
"do_stable_layer_norm": false,
|
| 48 |
+
"dtype": "float32",
|
| 49 |
+
"eos_token_id": 2,
|
| 50 |
+
"feat_extract_activation": "gelu",
|
| 51 |
+
"feat_extract_norm": "group",
|
| 52 |
+
"feat_proj_dropout": 0.1,
|
| 53 |
+
"feat_quantizer_dropout": 0.0,
|
| 54 |
+
"final_dropout": 0.0,
|
| 55 |
+
"finetuning_task": "wav2vec2_emotion",
|
| 56 |
+
"freeze_feat_extract_train": true,
|
| 57 |
+
"hidden_act": "gelu",
|
| 58 |
+
"hidden_dropout": 0.1,
|
| 59 |
+
"hidden_size": 768,
|
| 60 |
+
"id2label": {
|
| 61 |
+
"0": "neutral",
|
| 62 |
+
"1": "calm",
|
| 63 |
+
"2": "happy",
|
| 64 |
+
"3": "sad",
|
| 65 |
+
"4": "angry",
|
| 66 |
+
"5": "fearful",
|
| 67 |
+
"6": "disgust",
|
| 68 |
+
"7": "surprised"
|
| 69 |
+
},
|
| 70 |
+
"initializer_range": 0.02,
|
| 71 |
+
"intermediate_size": 3072,
|
| 72 |
+
"label2id": {
|
| 73 |
+
"angry": 4,
|
| 74 |
+
"calm": 1,
|
| 75 |
+
"disgust": 6,
|
| 76 |
+
"fearful": 5,
|
| 77 |
+
"happy": 2,
|
| 78 |
+
"neutral": 0,
|
| 79 |
+
"sad": 3,
|
| 80 |
+
"surprised": 7
|
| 81 |
+
},
|
| 82 |
+
"layer_norm_eps": 1e-05,
|
| 83 |
+
"layerdrop": 0.05,
|
| 84 |
+
"mask_channel_length": 10,
|
| 85 |
+
"mask_channel_min_space": 1,
|
| 86 |
+
"mask_channel_other": 0.0,
|
| 87 |
+
"mask_channel_prob": 0.0,
|
| 88 |
+
"mask_channel_selection": "static",
|
| 89 |
+
"mask_feature_length": 10,
|
| 90 |
+
"mask_feature_min_masks": 0,
|
| 91 |
+
"mask_feature_prob": 0.0,
|
| 92 |
+
"mask_time_length": 10,
|
| 93 |
+
"mask_time_min_masks": 2,
|
| 94 |
+
"mask_time_min_space": 1,
|
| 95 |
+
"mask_time_other": 0.0,
|
| 96 |
+
"mask_time_prob": 0.05,
|
| 97 |
+
"mask_time_selection": "static",
|
| 98 |
+
"model_type": "wav2vec2",
|
| 99 |
+
"no_mask_channel_overlap": false,
|
| 100 |
+
"no_mask_time_overlap": false,
|
| 101 |
+
"num_adapter_layers": 3,
|
| 102 |
+
"num_attention_heads": 12,
|
| 103 |
+
"num_codevector_groups": 2,
|
| 104 |
+
"num_codevectors_per_group": 320,
|
| 105 |
+
"num_conv_pos_embedding_groups": 16,
|
| 106 |
+
"num_conv_pos_embeddings": 128,
|
| 107 |
+
"num_feat_extract_layers": 7,
|
| 108 |
+
"num_hidden_layers": 12,
|
| 109 |
+
"num_negatives": 100,
|
| 110 |
+
"output_hidden_size": 768,
|
| 111 |
+
"pad_token_id": 0,
|
| 112 |
+
"proj_codevector_dim": 256,
|
| 113 |
+
"tdnn_dilation": [
|
| 114 |
+
1,
|
| 115 |
+
2,
|
| 116 |
+
3,
|
| 117 |
+
1,
|
| 118 |
+
1
|
| 119 |
+
],
|
| 120 |
+
"tdnn_dim": [
|
| 121 |
+
512,
|
| 122 |
+
512,
|
| 123 |
+
512,
|
| 124 |
+
512,
|
| 125 |
+
1500
|
| 126 |
+
],
|
| 127 |
+
"tdnn_kernel": [
|
| 128 |
+
5,
|
| 129 |
+
3,
|
| 130 |
+
3,
|
| 131 |
+
1,
|
| 132 |
+
1
|
| 133 |
+
],
|
| 134 |
+
"transformers_version": "4.57.1",
|
| 135 |
+
"use_cache": false,
|
| 136 |
+
"use_weighted_layer_sum": true,
|
| 137 |
+
"vocab_size": 32,
|
| 138 |
+
"xvector_output_dim": 512
|
| 139 |
+
}
|
checkpoint-5500/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:09f392b54932164c792cc8914621693b635ab82c3468157984201a99d604fe86
|
| 3 |
+
size 378308660
|
checkpoint-5500/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:02a1e676faefcb93be96420a683dc72165cecff9899910006bbbc8b5828e5fc6
|
| 3 |
+
size 756740235
|
checkpoint-5500/preprocessor_config.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_normalize": false,
|
| 3 |
+
"feature_extractor_type": "Wav2Vec2FeatureExtractor",
|
| 4 |
+
"feature_size": 1,
|
| 5 |
+
"padding_side": "right",
|
| 6 |
+
"padding_value": 0.0,
|
| 7 |
+
"return_attention_mask": true,
|
| 8 |
+
"sampling_rate": 16000
|
| 9 |
+
}
|
checkpoint-5500/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8d607223156c93d7364b87d1fd28545a218a521108b342deedbbdcfcf7609a85
|
| 3 |
+
size 14391
|
checkpoint-5500/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ad05c391e18917db5e823bd4d6accbaa4c8dc687f95feb992d6a8ef167586a2a
|
| 3 |
+
size 1465
|
checkpoint-5500/trainer_state.json
ADDED
|
@@ -0,0 +1,1574 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 19.29824561403509,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 5500,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.08771929824561403,
|
| 14 |
+
"grad_norm": 14.743780136108398,
|
| 15 |
+
"learning_rate": 1.4942583732057417e-05,
|
| 16 |
+
"loss": 0.9366,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.17543859649122806,
|
| 21 |
+
"grad_norm": 22.20017433166504,
|
| 22 |
+
"learning_rate": 1.4882775119617225e-05,
|
| 23 |
+
"loss": 1.2015,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.2631578947368421,
|
| 28 |
+
"grad_norm": 25.090787887573242,
|
| 29 |
+
"learning_rate": 1.4822966507177032e-05,
|
| 30 |
+
"loss": 1.0156,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.3508771929824561,
|
| 35 |
+
"grad_norm": 13.580609321594238,
|
| 36 |
+
"learning_rate": 1.4763157894736842e-05,
|
| 37 |
+
"loss": 1.117,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.43859649122807015,
|
| 42 |
+
"grad_norm": 12.062329292297363,
|
| 43 |
+
"learning_rate": 1.470334928229665e-05,
|
| 44 |
+
"loss": 0.8312,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.5263157894736842,
|
| 49 |
+
"grad_norm": 18.625316619873047,
|
| 50 |
+
"learning_rate": 1.464354066985646e-05,
|
| 51 |
+
"loss": 0.9938,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.6140350877192983,
|
| 56 |
+
"grad_norm": 11.628118515014648,
|
| 57 |
+
"learning_rate": 1.4583732057416269e-05,
|
| 58 |
+
"loss": 0.8922,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.7017543859649122,
|
| 63 |
+
"grad_norm": 12.121639251708984,
|
| 64 |
+
"learning_rate": 1.4523923444976078e-05,
|
| 65 |
+
"loss": 0.9787,
|
| 66 |
+
"step": 200
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 0.7894736842105263,
|
| 70 |
+
"grad_norm": 14.091375350952148,
|
| 71 |
+
"learning_rate": 1.4464114832535886e-05,
|
| 72 |
+
"loss": 0.806,
|
| 73 |
+
"step": 225
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.8771929824561403,
|
| 77 |
+
"grad_norm": 9.919096946716309,
|
| 78 |
+
"learning_rate": 1.4404306220095696e-05,
|
| 79 |
+
"loss": 0.9935,
|
| 80 |
+
"step": 250
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.9649122807017544,
|
| 84 |
+
"grad_norm": 13.80988597869873,
|
| 85 |
+
"learning_rate": 1.4344497607655503e-05,
|
| 86 |
+
"loss": 0.8116,
|
| 87 |
+
"step": 275
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 1.0526315789473684,
|
| 91 |
+
"grad_norm": 13.865740776062012,
|
| 92 |
+
"learning_rate": 1.4284688995215311e-05,
|
| 93 |
+
"loss": 0.7288,
|
| 94 |
+
"step": 300
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 1.1403508771929824,
|
| 98 |
+
"grad_norm": 9.76303768157959,
|
| 99 |
+
"learning_rate": 1.4224880382775121e-05,
|
| 100 |
+
"loss": 0.8283,
|
| 101 |
+
"step": 325
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 1.2280701754385965,
|
| 105 |
+
"grad_norm": 9.364381790161133,
|
| 106 |
+
"learning_rate": 1.4165071770334929e-05,
|
| 107 |
+
"loss": 0.6827,
|
| 108 |
+
"step": 350
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 1.3157894736842106,
|
| 112 |
+
"grad_norm": 13.79258918762207,
|
| 113 |
+
"learning_rate": 1.4105263157894738e-05,
|
| 114 |
+
"loss": 0.7745,
|
| 115 |
+
"step": 375
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 1.4035087719298245,
|
| 119 |
+
"grad_norm": 9.642690658569336,
|
| 120 |
+
"learning_rate": 1.4045454545454546e-05,
|
| 121 |
+
"loss": 0.7514,
|
| 122 |
+
"step": 400
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 1.4912280701754386,
|
| 126 |
+
"grad_norm": 14.490239143371582,
|
| 127 |
+
"learning_rate": 1.3985645933014356e-05,
|
| 128 |
+
"loss": 0.6594,
|
| 129 |
+
"step": 425
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 1.5789473684210527,
|
| 133 |
+
"grad_norm": 13.364012718200684,
|
| 134 |
+
"learning_rate": 1.3925837320574163e-05,
|
| 135 |
+
"loss": 0.6036,
|
| 136 |
+
"step": 450
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 1.6666666666666665,
|
| 140 |
+
"grad_norm": 6.50977087020874,
|
| 141 |
+
"learning_rate": 1.3866028708133971e-05,
|
| 142 |
+
"loss": 0.7616,
|
| 143 |
+
"step": 475
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 1.7543859649122808,
|
| 147 |
+
"grad_norm": 30.46417808532715,
|
| 148 |
+
"learning_rate": 1.380622009569378e-05,
|
| 149 |
+
"loss": 0.6735,
|
| 150 |
+
"step": 500
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 1.8421052631578947,
|
| 154 |
+
"grad_norm": 17.846593856811523,
|
| 155 |
+
"learning_rate": 1.3746411483253589e-05,
|
| 156 |
+
"loss": 0.5902,
|
| 157 |
+
"step": 525
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 1.9298245614035088,
|
| 161 |
+
"grad_norm": 19.83995819091797,
|
| 162 |
+
"learning_rate": 1.3686602870813398e-05,
|
| 163 |
+
"loss": 0.5928,
|
| 164 |
+
"step": 550
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 2.017543859649123,
|
| 168 |
+
"grad_norm": 6.053647994995117,
|
| 169 |
+
"learning_rate": 1.3626794258373206e-05,
|
| 170 |
+
"loss": 0.5816,
|
| 171 |
+
"step": 575
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 2.1052631578947367,
|
| 175 |
+
"grad_norm": 9.594230651855469,
|
| 176 |
+
"learning_rate": 1.3566985645933015e-05,
|
| 177 |
+
"loss": 0.488,
|
| 178 |
+
"step": 600
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 2.192982456140351,
|
| 182 |
+
"grad_norm": 13.84262752532959,
|
| 183 |
+
"learning_rate": 1.3507177033492823e-05,
|
| 184 |
+
"loss": 0.568,
|
| 185 |
+
"step": 625
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 2.280701754385965,
|
| 189 |
+
"grad_norm": 10.86662483215332,
|
| 190 |
+
"learning_rate": 1.3447368421052633e-05,
|
| 191 |
+
"loss": 0.4017,
|
| 192 |
+
"step": 650
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 2.3684210526315788,
|
| 196 |
+
"grad_norm": 16.06069564819336,
|
| 197 |
+
"learning_rate": 1.338755980861244e-05,
|
| 198 |
+
"loss": 0.5112,
|
| 199 |
+
"step": 675
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 2.456140350877193,
|
| 203 |
+
"grad_norm": 6.421328544616699,
|
| 204 |
+
"learning_rate": 1.3327751196172248e-05,
|
| 205 |
+
"loss": 0.3609,
|
| 206 |
+
"step": 700
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 2.543859649122807,
|
| 210 |
+
"grad_norm": 6.225348949432373,
|
| 211 |
+
"learning_rate": 1.3267942583732058e-05,
|
| 212 |
+
"loss": 0.5531,
|
| 213 |
+
"step": 725
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 2.6315789473684212,
|
| 217 |
+
"grad_norm": 14.942207336425781,
|
| 218 |
+
"learning_rate": 1.3208133971291866e-05,
|
| 219 |
+
"loss": 0.4272,
|
| 220 |
+
"step": 750
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 2.719298245614035,
|
| 224 |
+
"grad_norm": 8.975879669189453,
|
| 225 |
+
"learning_rate": 1.3148325358851675e-05,
|
| 226 |
+
"loss": 0.4509,
|
| 227 |
+
"step": 775
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 2.807017543859649,
|
| 231 |
+
"grad_norm": 21.482088088989258,
|
| 232 |
+
"learning_rate": 1.3088516746411483e-05,
|
| 233 |
+
"loss": 0.4233,
|
| 234 |
+
"step": 800
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 2.8947368421052633,
|
| 238 |
+
"grad_norm": 5.1405792236328125,
|
| 239 |
+
"learning_rate": 1.3028708133971293e-05,
|
| 240 |
+
"loss": 0.418,
|
| 241 |
+
"step": 825
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 2.982456140350877,
|
| 245 |
+
"grad_norm": 14.574483871459961,
|
| 246 |
+
"learning_rate": 1.29688995215311e-05,
|
| 247 |
+
"loss": 0.6778,
|
| 248 |
+
"step": 850
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 3.0701754385964914,
|
| 252 |
+
"grad_norm": 7.385740280151367,
|
| 253 |
+
"learning_rate": 1.290909090909091e-05,
|
| 254 |
+
"loss": 0.4127,
|
| 255 |
+
"step": 875
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 3.1578947368421053,
|
| 259 |
+
"grad_norm": 48.35734939575195,
|
| 260 |
+
"learning_rate": 1.2849282296650718e-05,
|
| 261 |
+
"loss": 0.4752,
|
| 262 |
+
"step": 900
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 3.245614035087719,
|
| 266 |
+
"grad_norm": 8.056925773620605,
|
| 267 |
+
"learning_rate": 1.2789473684210526e-05,
|
| 268 |
+
"loss": 0.411,
|
| 269 |
+
"step": 925
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 3.3333333333333335,
|
| 273 |
+
"grad_norm": 20.510982513427734,
|
| 274 |
+
"learning_rate": 1.2729665071770335e-05,
|
| 275 |
+
"loss": 0.495,
|
| 276 |
+
"step": 950
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 3.4210526315789473,
|
| 280 |
+
"grad_norm": 25.063684463500977,
|
| 281 |
+
"learning_rate": 1.2669856459330143e-05,
|
| 282 |
+
"loss": 0.2436,
|
| 283 |
+
"step": 975
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 3.5087719298245617,
|
| 287 |
+
"grad_norm": 15.523259162902832,
|
| 288 |
+
"learning_rate": 1.2610047846889953e-05,
|
| 289 |
+
"loss": 0.4612,
|
| 290 |
+
"step": 1000
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 3.5964912280701755,
|
| 294 |
+
"grad_norm": 1.9646857976913452,
|
| 295 |
+
"learning_rate": 1.255023923444976e-05,
|
| 296 |
+
"loss": 0.4426,
|
| 297 |
+
"step": 1025
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 3.6842105263157894,
|
| 301 |
+
"grad_norm": 30.676250457763672,
|
| 302 |
+
"learning_rate": 1.249043062200957e-05,
|
| 303 |
+
"loss": 0.3333,
|
| 304 |
+
"step": 1050
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 3.7719298245614032,
|
| 308 |
+
"grad_norm": 6.475268363952637,
|
| 309 |
+
"learning_rate": 1.2430622009569378e-05,
|
| 310 |
+
"loss": 0.2945,
|
| 311 |
+
"step": 1075
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 3.8596491228070176,
|
| 315 |
+
"grad_norm": 24.684438705444336,
|
| 316 |
+
"learning_rate": 1.2370813397129186e-05,
|
| 317 |
+
"loss": 0.4339,
|
| 318 |
+
"step": 1100
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 3.9473684210526314,
|
| 322 |
+
"grad_norm": 38.601226806640625,
|
| 323 |
+
"learning_rate": 1.2311004784688995e-05,
|
| 324 |
+
"loss": 0.3701,
|
| 325 |
+
"step": 1125
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 4.035087719298246,
|
| 329 |
+
"grad_norm": 2.8431146144866943,
|
| 330 |
+
"learning_rate": 1.2251196172248803e-05,
|
| 331 |
+
"loss": 0.2782,
|
| 332 |
+
"step": 1150
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 4.12280701754386,
|
| 336 |
+
"grad_norm": 23.384532928466797,
|
| 337 |
+
"learning_rate": 1.2191387559808613e-05,
|
| 338 |
+
"loss": 0.312,
|
| 339 |
+
"step": 1175
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 4.2105263157894735,
|
| 343 |
+
"grad_norm": 3.222729444503784,
|
| 344 |
+
"learning_rate": 1.213157894736842e-05,
|
| 345 |
+
"loss": 0.264,
|
| 346 |
+
"step": 1200
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 4.298245614035087,
|
| 350 |
+
"grad_norm": 3.5861947536468506,
|
| 351 |
+
"learning_rate": 1.207177033492823e-05,
|
| 352 |
+
"loss": 0.3104,
|
| 353 |
+
"step": 1225
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 4.385964912280702,
|
| 357 |
+
"grad_norm": 4.091048717498779,
|
| 358 |
+
"learning_rate": 1.2011961722488038e-05,
|
| 359 |
+
"loss": 0.3567,
|
| 360 |
+
"step": 1250
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 4.473684210526316,
|
| 364 |
+
"grad_norm": 1.3814598321914673,
|
| 365 |
+
"learning_rate": 1.1952153110047847e-05,
|
| 366 |
+
"loss": 0.3455,
|
| 367 |
+
"step": 1275
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 4.56140350877193,
|
| 371 |
+
"grad_norm": 2.480316162109375,
|
| 372 |
+
"learning_rate": 1.1892344497607655e-05,
|
| 373 |
+
"loss": 0.1769,
|
| 374 |
+
"step": 1300
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 4.649122807017544,
|
| 378 |
+
"grad_norm": 20.069162368774414,
|
| 379 |
+
"learning_rate": 1.1832535885167463e-05,
|
| 380 |
+
"loss": 0.2483,
|
| 381 |
+
"step": 1325
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 4.7368421052631575,
|
| 385 |
+
"grad_norm": 1.7885949611663818,
|
| 386 |
+
"learning_rate": 1.1772727272727272e-05,
|
| 387 |
+
"loss": 0.2283,
|
| 388 |
+
"step": 1350
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 4.824561403508772,
|
| 392 |
+
"grad_norm": 40.37774658203125,
|
| 393 |
+
"learning_rate": 1.171291866028708e-05,
|
| 394 |
+
"loss": 0.2277,
|
| 395 |
+
"step": 1375
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 4.912280701754386,
|
| 399 |
+
"grad_norm": 2.960739850997925,
|
| 400 |
+
"learning_rate": 1.1653110047846891e-05,
|
| 401 |
+
"loss": 0.2077,
|
| 402 |
+
"step": 1400
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 5.0,
|
| 406 |
+
"grad_norm": 14.10015869140625,
|
| 407 |
+
"learning_rate": 1.15933014354067e-05,
|
| 408 |
+
"loss": 0.3859,
|
| 409 |
+
"step": 1425
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 5.087719298245614,
|
| 413 |
+
"grad_norm": 4.440461158752441,
|
| 414 |
+
"learning_rate": 1.1533492822966509e-05,
|
| 415 |
+
"loss": 0.1354,
|
| 416 |
+
"step": 1450
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 5.175438596491228,
|
| 420 |
+
"grad_norm": 17.193849563598633,
|
| 421 |
+
"learning_rate": 1.1473684210526317e-05,
|
| 422 |
+
"loss": 0.2935,
|
| 423 |
+
"step": 1475
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 5.2631578947368425,
|
| 427 |
+
"grad_norm": 1.2828125953674316,
|
| 428 |
+
"learning_rate": 1.1413875598086125e-05,
|
| 429 |
+
"loss": 0.1394,
|
| 430 |
+
"step": 1500
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 5.350877192982456,
|
| 434 |
+
"grad_norm": 4.809077262878418,
|
| 435 |
+
"learning_rate": 1.1354066985645934e-05,
|
| 436 |
+
"loss": 0.3651,
|
| 437 |
+
"step": 1525
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 5.43859649122807,
|
| 441 |
+
"grad_norm": 43.53572463989258,
|
| 442 |
+
"learning_rate": 1.1294258373205742e-05,
|
| 443 |
+
"loss": 0.1187,
|
| 444 |
+
"step": 1550
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 5.526315789473684,
|
| 448 |
+
"grad_norm": 5.647329330444336,
|
| 449 |
+
"learning_rate": 1.1234449760765551e-05,
|
| 450 |
+
"loss": 0.2719,
|
| 451 |
+
"step": 1575
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 5.614035087719298,
|
| 455 |
+
"grad_norm": 25.37187957763672,
|
| 456 |
+
"learning_rate": 1.117464114832536e-05,
|
| 457 |
+
"loss": 0.2477,
|
| 458 |
+
"step": 1600
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 5.701754385964913,
|
| 462 |
+
"grad_norm": 4.228896617889404,
|
| 463 |
+
"learning_rate": 1.1114832535885169e-05,
|
| 464 |
+
"loss": 0.3608,
|
| 465 |
+
"step": 1625
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 5.7894736842105265,
|
| 469 |
+
"grad_norm": 2.1760356426239014,
|
| 470 |
+
"learning_rate": 1.1055023923444977e-05,
|
| 471 |
+
"loss": 0.1223,
|
| 472 |
+
"step": 1650
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 5.87719298245614,
|
| 476 |
+
"grad_norm": 28.078975677490234,
|
| 477 |
+
"learning_rate": 1.0995215311004786e-05,
|
| 478 |
+
"loss": 0.3207,
|
| 479 |
+
"step": 1675
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 5.964912280701754,
|
| 483 |
+
"grad_norm": 12.583624839782715,
|
| 484 |
+
"learning_rate": 1.0935406698564594e-05,
|
| 485 |
+
"loss": 0.1094,
|
| 486 |
+
"step": 1700
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 6.052631578947368,
|
| 490 |
+
"grad_norm": 0.4104742109775543,
|
| 491 |
+
"learning_rate": 1.0875598086124402e-05,
|
| 492 |
+
"loss": 0.1937,
|
| 493 |
+
"step": 1725
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 6.140350877192983,
|
| 497 |
+
"grad_norm": 1.5135095119476318,
|
| 498 |
+
"learning_rate": 1.0815789473684211e-05,
|
| 499 |
+
"loss": 0.3113,
|
| 500 |
+
"step": 1750
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 6.228070175438597,
|
| 504 |
+
"grad_norm": 1.9227218627929688,
|
| 505 |
+
"learning_rate": 1.0755980861244019e-05,
|
| 506 |
+
"loss": 0.1179,
|
| 507 |
+
"step": 1775
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 6.315789473684211,
|
| 511 |
+
"grad_norm": 7.322267055511475,
|
| 512 |
+
"learning_rate": 1.0696172248803829e-05,
|
| 513 |
+
"loss": 0.209,
|
| 514 |
+
"step": 1800
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 6.4035087719298245,
|
| 518 |
+
"grad_norm": 2.8994693756103516,
|
| 519 |
+
"learning_rate": 1.0636363636363636e-05,
|
| 520 |
+
"loss": 0.1743,
|
| 521 |
+
"step": 1825
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 6.491228070175438,
|
| 525 |
+
"grad_norm": 21.65730857849121,
|
| 526 |
+
"learning_rate": 1.0576555023923446e-05,
|
| 527 |
+
"loss": 0.2483,
|
| 528 |
+
"step": 1850
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 6.578947368421053,
|
| 532 |
+
"grad_norm": 24.210969924926758,
|
| 533 |
+
"learning_rate": 1.0516746411483254e-05,
|
| 534 |
+
"loss": 0.1348,
|
| 535 |
+
"step": 1875
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 6.666666666666667,
|
| 539 |
+
"grad_norm": 1.589753270149231,
|
| 540 |
+
"learning_rate": 1.0456937799043063e-05,
|
| 541 |
+
"loss": 0.229,
|
| 542 |
+
"step": 1900
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 6.754385964912281,
|
| 546 |
+
"grad_norm": 3.508387327194214,
|
| 547 |
+
"learning_rate": 1.0397129186602871e-05,
|
| 548 |
+
"loss": 0.0909,
|
| 549 |
+
"step": 1925
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 6.842105263157895,
|
| 553 |
+
"grad_norm": 1.89371919631958,
|
| 554 |
+
"learning_rate": 1.0337320574162679e-05,
|
| 555 |
+
"loss": 0.1859,
|
| 556 |
+
"step": 1950
|
| 557 |
+
},
|
| 558 |
+
{
|
| 559 |
+
"epoch": 6.9298245614035086,
|
| 560 |
+
"grad_norm": 0.35574591159820557,
|
| 561 |
+
"learning_rate": 1.0277511961722489e-05,
|
| 562 |
+
"loss": 0.1675,
|
| 563 |
+
"step": 1975
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 7.017543859649122,
|
| 567 |
+
"grad_norm": 0.621152937412262,
|
| 568 |
+
"learning_rate": 1.0217703349282296e-05,
|
| 569 |
+
"loss": 0.1082,
|
| 570 |
+
"step": 2000
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 7.105263157894737,
|
| 574 |
+
"grad_norm": 0.5694774985313416,
|
| 575 |
+
"learning_rate": 1.0157894736842106e-05,
|
| 576 |
+
"loss": 0.0886,
|
| 577 |
+
"step": 2025
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 7.192982456140351,
|
| 581 |
+
"grad_norm": 7.276403427124023,
|
| 582 |
+
"learning_rate": 1.0098086124401914e-05,
|
| 583 |
+
"loss": 0.2697,
|
| 584 |
+
"step": 2050
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 7.280701754385965,
|
| 588 |
+
"grad_norm": 0.4015137851238251,
|
| 589 |
+
"learning_rate": 1.0038277511961723e-05,
|
| 590 |
+
"loss": 0.1301,
|
| 591 |
+
"step": 2075
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 7.368421052631579,
|
| 595 |
+
"grad_norm": 0.27876588702201843,
|
| 596 |
+
"learning_rate": 9.978468899521531e-06,
|
| 597 |
+
"loss": 0.0687,
|
| 598 |
+
"step": 2100
|
| 599 |
+
},
|
| 600 |
+
{
|
| 601 |
+
"epoch": 7.456140350877193,
|
| 602 |
+
"grad_norm": 7.235930442810059,
|
| 603 |
+
"learning_rate": 9.918660287081339e-06,
|
| 604 |
+
"loss": 0.0734,
|
| 605 |
+
"step": 2125
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 7.543859649122807,
|
| 609 |
+
"grad_norm": 1.6751177310943604,
|
| 610 |
+
"learning_rate": 9.858851674641148e-06,
|
| 611 |
+
"loss": 0.1574,
|
| 612 |
+
"step": 2150
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 7.631578947368421,
|
| 616 |
+
"grad_norm": 0.8539127111434937,
|
| 617 |
+
"learning_rate": 9.799043062200956e-06,
|
| 618 |
+
"loss": 0.0483,
|
| 619 |
+
"step": 2175
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"epoch": 7.719298245614035,
|
| 623 |
+
"grad_norm": 0.26878660917282104,
|
| 624 |
+
"learning_rate": 9.739234449760766e-06,
|
| 625 |
+
"loss": 0.1632,
|
| 626 |
+
"step": 2200
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 7.807017543859649,
|
| 630 |
+
"grad_norm": 106.50234985351562,
|
| 631 |
+
"learning_rate": 9.679425837320574e-06,
|
| 632 |
+
"loss": 0.1156,
|
| 633 |
+
"step": 2225
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 7.894736842105263,
|
| 637 |
+
"grad_norm": 2.002487897872925,
|
| 638 |
+
"learning_rate": 9.619617224880383e-06,
|
| 639 |
+
"loss": 0.2036,
|
| 640 |
+
"step": 2250
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 7.982456140350877,
|
| 644 |
+
"grad_norm": 0.5302895307540894,
|
| 645 |
+
"learning_rate": 9.559808612440191e-06,
|
| 646 |
+
"loss": 0.1417,
|
| 647 |
+
"step": 2275
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 8.070175438596491,
|
| 651 |
+
"grad_norm": 0.314165323972702,
|
| 652 |
+
"learning_rate": 9.5e-06,
|
| 653 |
+
"loss": 0.0673,
|
| 654 |
+
"step": 2300
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 8.157894736842104,
|
| 658 |
+
"grad_norm": 0.5242946743965149,
|
| 659 |
+
"learning_rate": 9.440191387559808e-06,
|
| 660 |
+
"loss": 0.0689,
|
| 661 |
+
"step": 2325
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 8.24561403508772,
|
| 665 |
+
"grad_norm": 48.71388244628906,
|
| 666 |
+
"learning_rate": 9.380382775119616e-06,
|
| 667 |
+
"loss": 0.1863,
|
| 668 |
+
"step": 2350
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 8.333333333333334,
|
| 672 |
+
"grad_norm": 0.5670514106750488,
|
| 673 |
+
"learning_rate": 9.320574162679426e-06,
|
| 674 |
+
"loss": 0.0823,
|
| 675 |
+
"step": 2375
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 8.421052631578947,
|
| 679 |
+
"grad_norm": 0.6025238633155823,
|
| 680 |
+
"learning_rate": 9.260765550239234e-06,
|
| 681 |
+
"loss": 0.1118,
|
| 682 |
+
"step": 2400
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 8.508771929824562,
|
| 686 |
+
"grad_norm": 42.649696350097656,
|
| 687 |
+
"learning_rate": 9.200956937799043e-06,
|
| 688 |
+
"loss": 0.1333,
|
| 689 |
+
"step": 2425
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 8.596491228070175,
|
| 693 |
+
"grad_norm": 0.4079756438732147,
|
| 694 |
+
"learning_rate": 9.141148325358851e-06,
|
| 695 |
+
"loss": 0.156,
|
| 696 |
+
"step": 2450
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 8.68421052631579,
|
| 700 |
+
"grad_norm": 0.31388023495674133,
|
| 701 |
+
"learning_rate": 9.08133971291866e-06,
|
| 702 |
+
"loss": 0.2159,
|
| 703 |
+
"step": 2475
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 8.771929824561404,
|
| 707 |
+
"grad_norm": 0.2017977237701416,
|
| 708 |
+
"learning_rate": 9.021531100478468e-06,
|
| 709 |
+
"loss": 0.0936,
|
| 710 |
+
"step": 2500
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 8.859649122807017,
|
| 714 |
+
"grad_norm": 2.2358014583587646,
|
| 715 |
+
"learning_rate": 6.697368421052631e-06,
|
| 716 |
+
"loss": 0.1884,
|
| 717 |
+
"step": 2525
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 8.947368421052632,
|
| 721 |
+
"grad_norm": 0.5669077038764954,
|
| 722 |
+
"learning_rate": 6.615131578947369e-06,
|
| 723 |
+
"loss": 0.2183,
|
| 724 |
+
"step": 2550
|
| 725 |
+
},
|
| 726 |
+
{
|
| 727 |
+
"epoch": 9.035087719298245,
|
| 728 |
+
"grad_norm": 27.18621253967285,
|
| 729 |
+
"learning_rate": 6.532894736842106e-06,
|
| 730 |
+
"loss": 0.1232,
|
| 731 |
+
"step": 2575
|
| 732 |
+
},
|
| 733 |
+
{
|
| 734 |
+
"epoch": 9.12280701754386,
|
| 735 |
+
"grad_norm": 468.5591735839844,
|
| 736 |
+
"learning_rate": 6.450657894736842e-06,
|
| 737 |
+
"loss": 0.0516,
|
| 738 |
+
"step": 2600
|
| 739 |
+
},
|
| 740 |
+
{
|
| 741 |
+
"epoch": 9.210526315789474,
|
| 742 |
+
"grad_norm": 0.24007272720336914,
|
| 743 |
+
"learning_rate": 6.368421052631579e-06,
|
| 744 |
+
"loss": 0.0695,
|
| 745 |
+
"step": 2625
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 9.298245614035087,
|
| 749 |
+
"grad_norm": 99.38546752929688,
|
| 750 |
+
"learning_rate": 6.286184210526316e-06,
|
| 751 |
+
"loss": 0.0947,
|
| 752 |
+
"step": 2650
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 9.385964912280702,
|
| 756 |
+
"grad_norm": 3.8673267364501953,
|
| 757 |
+
"learning_rate": 6.203947368421053e-06,
|
| 758 |
+
"loss": 0.1062,
|
| 759 |
+
"step": 2675
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 9.473684210526315,
|
| 763 |
+
"grad_norm": 0.2559914290904999,
|
| 764 |
+
"learning_rate": 6.12171052631579e-06,
|
| 765 |
+
"loss": 0.0523,
|
| 766 |
+
"step": 2700
|
| 767 |
+
},
|
| 768 |
+
{
|
| 769 |
+
"epoch": 9.56140350877193,
|
| 770 |
+
"grad_norm": 0.12371934205293655,
|
| 771 |
+
"learning_rate": 6.039473684210526e-06,
|
| 772 |
+
"loss": 0.0495,
|
| 773 |
+
"step": 2725
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"epoch": 9.649122807017545,
|
| 777 |
+
"grad_norm": 2.5249898433685303,
|
| 778 |
+
"learning_rate": 5.957236842105263e-06,
|
| 779 |
+
"loss": 0.1678,
|
| 780 |
+
"step": 2750
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"epoch": 9.736842105263158,
|
| 784 |
+
"grad_norm": 0.2619442939758301,
|
| 785 |
+
"learning_rate": 5.8750000000000005e-06,
|
| 786 |
+
"loss": 0.0534,
|
| 787 |
+
"step": 2775
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"epoch": 9.824561403508772,
|
| 791 |
+
"grad_norm": 1.3705483675003052,
|
| 792 |
+
"learning_rate": 5.792763157894737e-06,
|
| 793 |
+
"loss": 0.1326,
|
| 794 |
+
"step": 2800
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 9.912280701754385,
|
| 798 |
+
"grad_norm": 74.48184967041016,
|
| 799 |
+
"learning_rate": 5.710526315789474e-06,
|
| 800 |
+
"loss": 0.1795,
|
| 801 |
+
"step": 2825
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 10.0,
|
| 805 |
+
"grad_norm": 79.78992462158203,
|
| 806 |
+
"learning_rate": 5.6282894736842106e-06,
|
| 807 |
+
"loss": 0.092,
|
| 808 |
+
"step": 2850
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"epoch": 10.087719298245615,
|
| 812 |
+
"grad_norm": 0.5286645889282227,
|
| 813 |
+
"learning_rate": 5.546052631578947e-06,
|
| 814 |
+
"loss": 0.0418,
|
| 815 |
+
"step": 2875
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"epoch": 10.175438596491228,
|
| 819 |
+
"grad_norm": 109.82157135009766,
|
| 820 |
+
"learning_rate": 5.463815789473684e-06,
|
| 821 |
+
"loss": 0.193,
|
| 822 |
+
"step": 2900
|
| 823 |
+
},
|
| 824 |
+
{
|
| 825 |
+
"epoch": 10.263157894736842,
|
| 826 |
+
"grad_norm": 3.963012933731079,
|
| 827 |
+
"learning_rate": 5.3815789473684215e-06,
|
| 828 |
+
"loss": 0.0897,
|
| 829 |
+
"step": 2925
|
| 830 |
+
},
|
| 831 |
+
{
|
| 832 |
+
"epoch": 10.350877192982455,
|
| 833 |
+
"grad_norm": 0.41540616750717163,
|
| 834 |
+
"learning_rate": 5.299342105263158e-06,
|
| 835 |
+
"loss": 0.1723,
|
| 836 |
+
"step": 2950
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 10.43859649122807,
|
| 840 |
+
"grad_norm": 5.900774002075195,
|
| 841 |
+
"learning_rate": 5.217105263157895e-06,
|
| 842 |
+
"loss": 0.0784,
|
| 843 |
+
"step": 2975
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 10.526315789473685,
|
| 847 |
+
"grad_norm": 12.152785301208496,
|
| 848 |
+
"learning_rate": 5.1348684210526315e-06,
|
| 849 |
+
"loss": 0.1186,
|
| 850 |
+
"step": 3000
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"epoch": 10.614035087719298,
|
| 854 |
+
"grad_norm": 0.21494676172733307,
|
| 855 |
+
"learning_rate": 3.6315789473684213e-06,
|
| 856 |
+
"loss": 0.0358,
|
| 857 |
+
"step": 3025
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 10.701754385964913,
|
| 861 |
+
"grad_norm": 90.48876953125,
|
| 862 |
+
"learning_rate": 3.537593984962406e-06,
|
| 863 |
+
"loss": 0.0471,
|
| 864 |
+
"step": 3050
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 10.789473684210526,
|
| 868 |
+
"grad_norm": 0.23261310160160065,
|
| 869 |
+
"learning_rate": 3.4436090225563912e-06,
|
| 870 |
+
"loss": 0.0276,
|
| 871 |
+
"step": 3075
|
| 872 |
+
},
|
| 873 |
+
{
|
| 874 |
+
"epoch": 10.87719298245614,
|
| 875 |
+
"grad_norm": 0.782564640045166,
|
| 876 |
+
"learning_rate": 3.349624060150376e-06,
|
| 877 |
+
"loss": 0.0718,
|
| 878 |
+
"step": 3100
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 10.964912280701755,
|
| 882 |
+
"grad_norm": 0.33198878169059753,
|
| 883 |
+
"learning_rate": 3.255639097744361e-06,
|
| 884 |
+
"loss": 0.0564,
|
| 885 |
+
"step": 3125
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 11.052631578947368,
|
| 889 |
+
"grad_norm": 0.14864841103553772,
|
| 890 |
+
"learning_rate": 3.161654135338346e-06,
|
| 891 |
+
"loss": 0.0508,
|
| 892 |
+
"step": 3150
|
| 893 |
+
},
|
| 894 |
+
{
|
| 895 |
+
"epoch": 11.140350877192983,
|
| 896 |
+
"grad_norm": 0.1588129848241806,
|
| 897 |
+
"learning_rate": 3.067669172932331e-06,
|
| 898 |
+
"loss": 0.1641,
|
| 899 |
+
"step": 3175
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 11.228070175438596,
|
| 903 |
+
"grad_norm": 0.6562497019767761,
|
| 904 |
+
"learning_rate": 2.973684210526316e-06,
|
| 905 |
+
"loss": 0.0544,
|
| 906 |
+
"step": 3200
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"epoch": 11.31578947368421,
|
| 910 |
+
"grad_norm": 0.3608649969100952,
|
| 911 |
+
"learning_rate": 2.879699248120301e-06,
|
| 912 |
+
"loss": 0.0391,
|
| 913 |
+
"step": 3225
|
| 914 |
+
},
|
| 915 |
+
{
|
| 916 |
+
"epoch": 11.403508771929825,
|
| 917 |
+
"grad_norm": 20.602819442749023,
|
| 918 |
+
"learning_rate": 2.785714285714286e-06,
|
| 919 |
+
"loss": 0.0338,
|
| 920 |
+
"step": 3250
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 11.491228070175438,
|
| 924 |
+
"grad_norm": 157.23171997070312,
|
| 925 |
+
"learning_rate": 2.691729323308271e-06,
|
| 926 |
+
"loss": 0.0771,
|
| 927 |
+
"step": 3275
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 11.578947368421053,
|
| 931 |
+
"grad_norm": 0.22338764369487762,
|
| 932 |
+
"learning_rate": 2.597744360902256e-06,
|
| 933 |
+
"loss": 0.1187,
|
| 934 |
+
"step": 3300
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 11.666666666666666,
|
| 938 |
+
"grad_norm": 57.78382110595703,
|
| 939 |
+
"learning_rate": 2.503759398496241e-06,
|
| 940 |
+
"loss": 0.0165,
|
| 941 |
+
"step": 3325
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 11.75438596491228,
|
| 945 |
+
"grad_norm": 7.400601863861084,
|
| 946 |
+
"learning_rate": 2.4097744360902257e-06,
|
| 947 |
+
"loss": 0.0092,
|
| 948 |
+
"step": 3350
|
| 949 |
+
},
|
| 950 |
+
{
|
| 951 |
+
"epoch": 11.842105263157894,
|
| 952 |
+
"grad_norm": 0.18190287053585052,
|
| 953 |
+
"learning_rate": 2.3157894736842105e-06,
|
| 954 |
+
"loss": 0.1059,
|
| 955 |
+
"step": 3375
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 11.929824561403509,
|
| 959 |
+
"grad_norm": 0.1269238144159317,
|
| 960 |
+
"learning_rate": 2.2218045112781957e-06,
|
| 961 |
+
"loss": 0.0571,
|
| 962 |
+
"step": 3400
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 12.017543859649123,
|
| 966 |
+
"grad_norm": 0.10843072831630707,
|
| 967 |
+
"learning_rate": 2.1278195488721805e-06,
|
| 968 |
+
"loss": 0.1341,
|
| 969 |
+
"step": 3425
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 12.105263157894736,
|
| 973 |
+
"grad_norm": 0.7199507355690002,
|
| 974 |
+
"learning_rate": 2.0338345864661656e-06,
|
| 975 |
+
"loss": 0.0063,
|
| 976 |
+
"step": 3450
|
| 977 |
+
},
|
| 978 |
+
{
|
| 979 |
+
"epoch": 12.192982456140351,
|
| 980 |
+
"grad_norm": 0.19009971618652344,
|
| 981 |
+
"learning_rate": 1.9398496240601504e-06,
|
| 982 |
+
"loss": 0.054,
|
| 983 |
+
"step": 3475
|
| 984 |
+
},
|
| 985 |
+
{
|
| 986 |
+
"epoch": 12.280701754385966,
|
| 987 |
+
"grad_norm": 0.18020951747894287,
|
| 988 |
+
"learning_rate": 1.8458646616541354e-06,
|
| 989 |
+
"loss": 0.0148,
|
| 990 |
+
"step": 3500
|
| 991 |
+
},
|
| 992 |
+
{
|
| 993 |
+
"epoch": 12.368421052631579,
|
| 994 |
+
"grad_norm": 0.2494085282087326,
|
| 995 |
+
"learning_rate": 6.569377990430622e-06,
|
| 996 |
+
"loss": 0.1168,
|
| 997 |
+
"step": 3525
|
| 998 |
+
},
|
| 999 |
+
{
|
| 1000 |
+
"epoch": 12.456140350877194,
|
| 1001 |
+
"grad_norm": 0.16571170091629028,
|
| 1002 |
+
"learning_rate": 6.509569377990431e-06,
|
| 1003 |
+
"loss": 0.0482,
|
| 1004 |
+
"step": 3550
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 12.543859649122806,
|
| 1008 |
+
"grad_norm": 0.10973577946424484,
|
| 1009 |
+
"learning_rate": 6.44976076555024e-06,
|
| 1010 |
+
"loss": 0.1372,
|
| 1011 |
+
"step": 3575
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 12.631578947368421,
|
| 1015 |
+
"grad_norm": 0.14920943975448608,
|
| 1016 |
+
"learning_rate": 6.389952153110048e-06,
|
| 1017 |
+
"loss": 0.0791,
|
| 1018 |
+
"step": 3600
|
| 1019 |
+
},
|
| 1020 |
+
{
|
| 1021 |
+
"epoch": 12.719298245614034,
|
| 1022 |
+
"grad_norm": 4.5087504386901855,
|
| 1023 |
+
"learning_rate": 6.330143540669857e-06,
|
| 1024 |
+
"loss": 0.1614,
|
| 1025 |
+
"step": 3625
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 12.807017543859649,
|
| 1029 |
+
"grad_norm": 0.11903402209281921,
|
| 1030 |
+
"learning_rate": 6.270334928229666e-06,
|
| 1031 |
+
"loss": 0.0082,
|
| 1032 |
+
"step": 3650
|
| 1033 |
+
},
|
| 1034 |
+
{
|
| 1035 |
+
"epoch": 12.894736842105264,
|
| 1036 |
+
"grad_norm": 0.24349439144134521,
|
| 1037 |
+
"learning_rate": 6.210526315789474e-06,
|
| 1038 |
+
"loss": 0.1936,
|
| 1039 |
+
"step": 3675
|
| 1040 |
+
},
|
| 1041 |
+
{
|
| 1042 |
+
"epoch": 12.982456140350877,
|
| 1043 |
+
"grad_norm": 146.2913360595703,
|
| 1044 |
+
"learning_rate": 6.150717703349282e-06,
|
| 1045 |
+
"loss": 0.1465,
|
| 1046 |
+
"step": 3700
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 13.070175438596491,
|
| 1050 |
+
"grad_norm": 88.66224670410156,
|
| 1051 |
+
"learning_rate": 6.090909090909091e-06,
|
| 1052 |
+
"loss": 0.0751,
|
| 1053 |
+
"step": 3725
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 13.157894736842104,
|
| 1057 |
+
"grad_norm": 0.11558452993631363,
|
| 1058 |
+
"learning_rate": 6.0311004784689e-06,
|
| 1059 |
+
"loss": 0.0652,
|
| 1060 |
+
"step": 3750
|
| 1061 |
+
},
|
| 1062 |
+
{
|
| 1063 |
+
"epoch": 13.24561403508772,
|
| 1064 |
+
"grad_norm": 3.1845107078552246,
|
| 1065 |
+
"learning_rate": 5.971291866028708e-06,
|
| 1066 |
+
"loss": 0.0969,
|
| 1067 |
+
"step": 3775
|
| 1068 |
+
},
|
| 1069 |
+
{
|
| 1070 |
+
"epoch": 13.333333333333334,
|
| 1071 |
+
"grad_norm": 0.5500511527061462,
|
| 1072 |
+
"learning_rate": 5.911483253588517e-06,
|
| 1073 |
+
"loss": 0.107,
|
| 1074 |
+
"step": 3800
|
| 1075 |
+
},
|
| 1076 |
+
{
|
| 1077 |
+
"epoch": 13.421052631578947,
|
| 1078 |
+
"grad_norm": 0.08417954295873642,
|
| 1079 |
+
"learning_rate": 5.851674641148326e-06,
|
| 1080 |
+
"loss": 0.0165,
|
| 1081 |
+
"step": 3825
|
| 1082 |
+
},
|
| 1083 |
+
{
|
| 1084 |
+
"epoch": 13.508771929824562,
|
| 1085 |
+
"grad_norm": 0.0910855233669281,
|
| 1086 |
+
"learning_rate": 5.791866028708134e-06,
|
| 1087 |
+
"loss": 0.1326,
|
| 1088 |
+
"step": 3850
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 13.596491228070175,
|
| 1092 |
+
"grad_norm": 65.31088256835938,
|
| 1093 |
+
"learning_rate": 5.732057416267942e-06,
|
| 1094 |
+
"loss": 0.0365,
|
| 1095 |
+
"step": 3875
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 13.68421052631579,
|
| 1099 |
+
"grad_norm": 0.09439278393983841,
|
| 1100 |
+
"learning_rate": 5.672248803827751e-06,
|
| 1101 |
+
"loss": 0.0784,
|
| 1102 |
+
"step": 3900
|
| 1103 |
+
},
|
| 1104 |
+
{
|
| 1105 |
+
"epoch": 13.771929824561404,
|
| 1106 |
+
"grad_norm": 0.17843316495418549,
|
| 1107 |
+
"learning_rate": 5.6124401913875595e-06,
|
| 1108 |
+
"loss": 0.0232,
|
| 1109 |
+
"step": 3925
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 13.859649122807017,
|
| 1113 |
+
"grad_norm": 0.16249841451644897,
|
| 1114 |
+
"learning_rate": 5.552631578947368e-06,
|
| 1115 |
+
"loss": 0.0107,
|
| 1116 |
+
"step": 3950
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 13.947368421052632,
|
| 1120 |
+
"grad_norm": 0.545814573764801,
|
| 1121 |
+
"learning_rate": 5.492822966507177e-06,
|
| 1122 |
+
"loss": 0.0947,
|
| 1123 |
+
"step": 3975
|
| 1124 |
+
},
|
| 1125 |
+
{
|
| 1126 |
+
"epoch": 14.035087719298245,
|
| 1127 |
+
"grad_norm": 0.7138233184814453,
|
| 1128 |
+
"learning_rate": 5.433014354066986e-06,
|
| 1129 |
+
"loss": 0.1812,
|
| 1130 |
+
"step": 4000
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 14.12280701754386,
|
| 1134 |
+
"grad_norm": 381.4528503417969,
|
| 1135 |
+
"learning_rate": 5.373205741626794e-06,
|
| 1136 |
+
"loss": 0.0399,
|
| 1137 |
+
"step": 4025
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 14.210526315789474,
|
| 1141 |
+
"grad_norm": 0.053043171763420105,
|
| 1142 |
+
"learning_rate": 5.313397129186603e-06,
|
| 1143 |
+
"loss": 0.1249,
|
| 1144 |
+
"step": 4050
|
| 1145 |
+
},
|
| 1146 |
+
{
|
| 1147 |
+
"epoch": 14.298245614035087,
|
| 1148 |
+
"grad_norm": 1.4563120603561401,
|
| 1149 |
+
"learning_rate": 5.253588516746412e-06,
|
| 1150 |
+
"loss": 0.0361,
|
| 1151 |
+
"step": 4075
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 14.385964912280702,
|
| 1155 |
+
"grad_norm": 0.06027642637491226,
|
| 1156 |
+
"learning_rate": 5.19377990430622e-06,
|
| 1157 |
+
"loss": 0.0799,
|
| 1158 |
+
"step": 4100
|
| 1159 |
+
},
|
| 1160 |
+
{
|
| 1161 |
+
"epoch": 14.473684210526315,
|
| 1162 |
+
"grad_norm": 0.11955183744430542,
|
| 1163 |
+
"learning_rate": 5.133971291866029e-06,
|
| 1164 |
+
"loss": 0.2006,
|
| 1165 |
+
"step": 4125
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 14.56140350877193,
|
| 1169 |
+
"grad_norm": 0.07738160341978073,
|
| 1170 |
+
"learning_rate": 5.074162679425838e-06,
|
| 1171 |
+
"loss": 0.16,
|
| 1172 |
+
"step": 4150
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 14.649122807017545,
|
| 1176 |
+
"grad_norm": 28.58780860900879,
|
| 1177 |
+
"learning_rate": 5.014354066985646e-06,
|
| 1178 |
+
"loss": 0.0675,
|
| 1179 |
+
"step": 4175
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 14.736842105263158,
|
| 1183 |
+
"grad_norm": 0.08654920011758804,
|
| 1184 |
+
"learning_rate": 4.954545454545455e-06,
|
| 1185 |
+
"loss": 0.0584,
|
| 1186 |
+
"step": 4200
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 14.824561403508772,
|
| 1190 |
+
"grad_norm": 0.1421671062707901,
|
| 1191 |
+
"learning_rate": 4.894736842105264e-06,
|
| 1192 |
+
"loss": 0.0049,
|
| 1193 |
+
"step": 4225
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 14.912280701754385,
|
| 1197 |
+
"grad_norm": 0.23447291553020477,
|
| 1198 |
+
"learning_rate": 4.834928229665072e-06,
|
| 1199 |
+
"loss": 0.0122,
|
| 1200 |
+
"step": 4250
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 15.0,
|
| 1204 |
+
"grad_norm": 54.224788665771484,
|
| 1205 |
+
"learning_rate": 4.775119617224881e-06,
|
| 1206 |
+
"loss": 0.0921,
|
| 1207 |
+
"step": 4275
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 15.087719298245615,
|
| 1211 |
+
"grad_norm": 0.08427131175994873,
|
| 1212 |
+
"learning_rate": 4.715311004784689e-06,
|
| 1213 |
+
"loss": 0.0061,
|
| 1214 |
+
"step": 4300
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 15.175438596491228,
|
| 1218 |
+
"grad_norm": 0.232700914144516,
|
| 1219 |
+
"learning_rate": 4.6555023923444975e-06,
|
| 1220 |
+
"loss": 0.1011,
|
| 1221 |
+
"step": 4325
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 15.263157894736842,
|
| 1225 |
+
"grad_norm": 0.09075941890478134,
|
| 1226 |
+
"learning_rate": 4.595693779904306e-06,
|
| 1227 |
+
"loss": 0.0419,
|
| 1228 |
+
"step": 4350
|
| 1229 |
+
},
|
| 1230 |
+
{
|
| 1231 |
+
"epoch": 15.350877192982455,
|
| 1232 |
+
"grad_norm": 0.20714852213859558,
|
| 1233 |
+
"learning_rate": 4.535885167464115e-06,
|
| 1234 |
+
"loss": 0.1786,
|
| 1235 |
+
"step": 4375
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 15.43859649122807,
|
| 1239 |
+
"grad_norm": 0.07586350291967392,
|
| 1240 |
+
"learning_rate": 4.476076555023924e-06,
|
| 1241 |
+
"loss": 0.0314,
|
| 1242 |
+
"step": 4400
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 15.526315789473685,
|
| 1246 |
+
"grad_norm": 0.29287371039390564,
|
| 1247 |
+
"learning_rate": 4.416267942583732e-06,
|
| 1248 |
+
"loss": 0.0635,
|
| 1249 |
+
"step": 4425
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 15.614035087719298,
|
| 1253 |
+
"grad_norm": 50.64339065551758,
|
| 1254 |
+
"learning_rate": 4.356459330143541e-06,
|
| 1255 |
+
"loss": 0.0854,
|
| 1256 |
+
"step": 4450
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 15.701754385964913,
|
| 1260 |
+
"grad_norm": 0.12661752104759216,
|
| 1261 |
+
"learning_rate": 4.29665071770335e-06,
|
| 1262 |
+
"loss": 0.1378,
|
| 1263 |
+
"step": 4475
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 15.789473684210526,
|
| 1267 |
+
"grad_norm": 2.418790340423584,
|
| 1268 |
+
"learning_rate": 4.2368421052631575e-06,
|
| 1269 |
+
"loss": 0.0325,
|
| 1270 |
+
"step": 4500
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 15.87719298245614,
|
| 1274 |
+
"grad_norm": 231.30796813964844,
|
| 1275 |
+
"learning_rate": 4.177033492822966e-06,
|
| 1276 |
+
"loss": 0.1398,
|
| 1277 |
+
"step": 4525
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 15.964912280701755,
|
| 1281 |
+
"grad_norm": 0.23073101043701172,
|
| 1282 |
+
"learning_rate": 4.117224880382775e-06,
|
| 1283 |
+
"loss": 0.0237,
|
| 1284 |
+
"step": 4550
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 16.05263157894737,
|
| 1288 |
+
"grad_norm": 0.044060900807380676,
|
| 1289 |
+
"learning_rate": 4.0574162679425835e-06,
|
| 1290 |
+
"loss": 0.0074,
|
| 1291 |
+
"step": 4575
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 16.140350877192983,
|
| 1295 |
+
"grad_norm": 1.4264813661575317,
|
| 1296 |
+
"learning_rate": 3.997607655502392e-06,
|
| 1297 |
+
"loss": 0.025,
|
| 1298 |
+
"step": 4600
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 16.228070175438596,
|
| 1302 |
+
"grad_norm": 0.41092243790626526,
|
| 1303 |
+
"learning_rate": 3.937799043062201e-06,
|
| 1304 |
+
"loss": 0.0049,
|
| 1305 |
+
"step": 4625
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 16.31578947368421,
|
| 1309 |
+
"grad_norm": 3.6420750617980957,
|
| 1310 |
+
"learning_rate": 3.8779904306220095e-06,
|
| 1311 |
+
"loss": 0.0491,
|
| 1312 |
+
"step": 4650
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"epoch": 16.403508771929825,
|
| 1316 |
+
"grad_norm": 0.08417033404111862,
|
| 1317 |
+
"learning_rate": 3.818181818181818e-06,
|
| 1318 |
+
"loss": 0.003,
|
| 1319 |
+
"step": 4675
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 16.49122807017544,
|
| 1323 |
+
"grad_norm": 0.41752633452415466,
|
| 1324 |
+
"learning_rate": 3.7583732057416273e-06,
|
| 1325 |
+
"loss": 0.0649,
|
| 1326 |
+
"step": 4700
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 16.57894736842105,
|
| 1330 |
+
"grad_norm": 0.04892294108867645,
|
| 1331 |
+
"learning_rate": 3.6985645933014356e-06,
|
| 1332 |
+
"loss": 0.1577,
|
| 1333 |
+
"step": 4725
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 16.666666666666668,
|
| 1337 |
+
"grad_norm": 0.07886885106563568,
|
| 1338 |
+
"learning_rate": 3.6387559808612442e-06,
|
| 1339 |
+
"loss": 0.1133,
|
| 1340 |
+
"step": 4750
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 16.75438596491228,
|
| 1344 |
+
"grad_norm": 202.08302307128906,
|
| 1345 |
+
"learning_rate": 3.5789473684210525e-06,
|
| 1346 |
+
"loss": 0.0151,
|
| 1347 |
+
"step": 4775
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 16.842105263157894,
|
| 1351 |
+
"grad_norm": 77.71251678466797,
|
| 1352 |
+
"learning_rate": 3.519138755980861e-06,
|
| 1353 |
+
"loss": 0.0644,
|
| 1354 |
+
"step": 4800
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 16.92982456140351,
|
| 1358 |
+
"grad_norm": 0.04525836929678917,
|
| 1359 |
+
"learning_rate": 3.45933014354067e-06,
|
| 1360 |
+
"loss": 0.0283,
|
| 1361 |
+
"step": 4825
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 17.017543859649123,
|
| 1365 |
+
"grad_norm": 0.035384029150009155,
|
| 1366 |
+
"learning_rate": 3.399521531100479e-06,
|
| 1367 |
+
"loss": 0.0431,
|
| 1368 |
+
"step": 4850
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 17.105263157894736,
|
| 1372 |
+
"grad_norm": 0.05048861354589462,
|
| 1373 |
+
"learning_rate": 3.339712918660287e-06,
|
| 1374 |
+
"loss": 0.0358,
|
| 1375 |
+
"step": 4875
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 17.19298245614035,
|
| 1379 |
+
"grad_norm": 0.1347443163394928,
|
| 1380 |
+
"learning_rate": 3.279904306220096e-06,
|
| 1381 |
+
"loss": 0.0206,
|
| 1382 |
+
"step": 4900
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 17.280701754385966,
|
| 1386 |
+
"grad_norm": 252.46458435058594,
|
| 1387 |
+
"learning_rate": 3.2200956937799046e-06,
|
| 1388 |
+
"loss": 0.0804,
|
| 1389 |
+
"step": 4925
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 17.36842105263158,
|
| 1393 |
+
"grad_norm": 0.05461292341351509,
|
| 1394 |
+
"learning_rate": 3.1602870813397132e-06,
|
| 1395 |
+
"loss": 0.0784,
|
| 1396 |
+
"step": 4950
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 17.45614035087719,
|
| 1400 |
+
"grad_norm": 2.0001866817474365,
|
| 1401 |
+
"learning_rate": 3.1004784688995215e-06,
|
| 1402 |
+
"loss": 0.0947,
|
| 1403 |
+
"step": 4975
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 17.54385964912281,
|
| 1407 |
+
"grad_norm": 0.04590539261698723,
|
| 1408 |
+
"learning_rate": 3.04066985645933e-06,
|
| 1409 |
+
"loss": 0.075,
|
| 1410 |
+
"step": 5000
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 17.63157894736842,
|
| 1414 |
+
"grad_norm": 0.04810748249292374,
|
| 1415 |
+
"learning_rate": 2.980861244019139e-06,
|
| 1416 |
+
"loss": 0.0342,
|
| 1417 |
+
"step": 5025
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 17.719298245614034,
|
| 1421 |
+
"grad_norm": 0.8750467896461487,
|
| 1422 |
+
"learning_rate": 2.9210526315789475e-06,
|
| 1423 |
+
"loss": 0.0618,
|
| 1424 |
+
"step": 5050
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 17.80701754385965,
|
| 1428 |
+
"grad_norm": 56.14554977416992,
|
| 1429 |
+
"learning_rate": 2.8612440191387558e-06,
|
| 1430 |
+
"loss": 0.026,
|
| 1431 |
+
"step": 5075
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 17.894736842105264,
|
| 1435 |
+
"grad_norm": 0.8380083441734314,
|
| 1436 |
+
"learning_rate": 2.8014354066985645e-06,
|
| 1437 |
+
"loss": 0.0135,
|
| 1438 |
+
"step": 5100
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 17.982456140350877,
|
| 1442 |
+
"grad_norm": 0.08012706786394119,
|
| 1443 |
+
"learning_rate": 2.741626794258373e-06,
|
| 1444 |
+
"loss": 0.0586,
|
| 1445 |
+
"step": 5125
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 18.07017543859649,
|
| 1449 |
+
"grad_norm": 1.2549360990524292,
|
| 1450 |
+
"learning_rate": 2.6818181818181822e-06,
|
| 1451 |
+
"loss": 0.0047,
|
| 1452 |
+
"step": 5150
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"epoch": 18.157894736842106,
|
| 1456 |
+
"grad_norm": 0.03993632644414902,
|
| 1457 |
+
"learning_rate": 2.6220095693779905e-06,
|
| 1458 |
+
"loss": 0.0869,
|
| 1459 |
+
"step": 5175
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 18.24561403508772,
|
| 1463 |
+
"grad_norm": 62.21980285644531,
|
| 1464 |
+
"learning_rate": 2.562200956937799e-06,
|
| 1465 |
+
"loss": 0.0987,
|
| 1466 |
+
"step": 5200
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 18.333333333333332,
|
| 1470 |
+
"grad_norm": 0.10092537105083466,
|
| 1471 |
+
"learning_rate": 2.502392344497608e-06,
|
| 1472 |
+
"loss": 0.1554,
|
| 1473 |
+
"step": 5225
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 18.42105263157895,
|
| 1477 |
+
"grad_norm": 0.7630422115325928,
|
| 1478 |
+
"learning_rate": 2.4425837320574165e-06,
|
| 1479 |
+
"loss": 0.0522,
|
| 1480 |
+
"step": 5250
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 18.50877192982456,
|
| 1484 |
+
"grad_norm": 62.42081069946289,
|
| 1485 |
+
"learning_rate": 2.382775119617225e-06,
|
| 1486 |
+
"loss": 0.0298,
|
| 1487 |
+
"step": 5275
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 18.596491228070175,
|
| 1491 |
+
"grad_norm": 0.046070296317338943,
|
| 1492 |
+
"learning_rate": 2.3229665071770335e-06,
|
| 1493 |
+
"loss": 0.1207,
|
| 1494 |
+
"step": 5300
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 18.68421052631579,
|
| 1498 |
+
"grad_norm": 0.04533836990594864,
|
| 1499 |
+
"learning_rate": 2.263157894736842e-06,
|
| 1500 |
+
"loss": 0.146,
|
| 1501 |
+
"step": 5325
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 18.771929824561404,
|
| 1505 |
+
"grad_norm": 0.03193026781082153,
|
| 1506 |
+
"learning_rate": 2.203349282296651e-06,
|
| 1507 |
+
"loss": 0.0106,
|
| 1508 |
+
"step": 5350
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 18.859649122807017,
|
| 1512 |
+
"grad_norm": 0.03594537451863289,
|
| 1513 |
+
"learning_rate": 2.143540669856459e-06,
|
| 1514 |
+
"loss": 0.0266,
|
| 1515 |
+
"step": 5375
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 18.94736842105263,
|
| 1519 |
+
"grad_norm": 0.07382319122552872,
|
| 1520 |
+
"learning_rate": 2.0837320574162678e-06,
|
| 1521 |
+
"loss": 0.0954,
|
| 1522 |
+
"step": 5400
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 19.035087719298247,
|
| 1526 |
+
"grad_norm": 0.0792032852768898,
|
| 1527 |
+
"learning_rate": 2.0239234449760764e-06,
|
| 1528 |
+
"loss": 0.0871,
|
| 1529 |
+
"step": 5425
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 19.12280701754386,
|
| 1533 |
+
"grad_norm": 0.06994396448135376,
|
| 1534 |
+
"learning_rate": 1.9641148325358855e-06,
|
| 1535 |
+
"loss": 0.0092,
|
| 1536 |
+
"step": 5450
|
| 1537 |
+
},
|
| 1538 |
+
{
|
| 1539 |
+
"epoch": 19.210526315789473,
|
| 1540 |
+
"grad_norm": 36.98837661743164,
|
| 1541 |
+
"learning_rate": 1.904306220095694e-06,
|
| 1542 |
+
"loss": 0.1442,
|
| 1543 |
+
"step": 5475
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"epoch": 19.29824561403509,
|
| 1547 |
+
"grad_norm": 0.11950813978910446,
|
| 1548 |
+
"learning_rate": 1.8444976076555023e-06,
|
| 1549 |
+
"loss": 0.0344,
|
| 1550 |
+
"step": 5500
|
| 1551 |
+
}
|
| 1552 |
+
],
|
| 1553 |
+
"logging_steps": 25,
|
| 1554 |
+
"max_steps": 6270,
|
| 1555 |
+
"num_input_tokens_seen": 0,
|
| 1556 |
+
"num_train_epochs": 22,
|
| 1557 |
+
"save_steps": 500,
|
| 1558 |
+
"stateful_callbacks": {
|
| 1559 |
+
"TrainerControl": {
|
| 1560 |
+
"args": {
|
| 1561 |
+
"should_epoch_stop": false,
|
| 1562 |
+
"should_evaluate": false,
|
| 1563 |
+
"should_log": false,
|
| 1564 |
+
"should_save": true,
|
| 1565 |
+
"should_training_stop": false
|
| 1566 |
+
},
|
| 1567 |
+
"attributes": {}
|
| 1568 |
+
}
|
| 1569 |
+
},
|
| 1570 |
+
"total_flos": 1.4836658875741053e+18,
|
| 1571 |
+
"train_batch_size": 4,
|
| 1572 |
+
"trial_name": null,
|
| 1573 |
+
"trial_params": null
|
| 1574 |
+
}
|
checkpoint-5500/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:21cda56f26874205e02c84553dd1454d4b3eefd72224b421d8d696952554b651
|
| 3 |
+
size 5841
|
checkpoint-6000/config.json
ADDED
|
@@ -0,0 +1,139 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"activation_dropout": 0.0,
|
| 3 |
+
"adapter_attn_dim": null,
|
| 4 |
+
"adapter_kernel_size": 3,
|
| 5 |
+
"adapter_stride": 2,
|
| 6 |
+
"add_adapter": false,
|
| 7 |
+
"apply_spec_augment": true,
|
| 8 |
+
"architectures": [
|
| 9 |
+
"Wav2Vec2ForSequenceClassification"
|
| 10 |
+
],
|
| 11 |
+
"attention_dropout": 0.1,
|
| 12 |
+
"bos_token_id": 1,
|
| 13 |
+
"classifier_proj_size": 256,
|
| 14 |
+
"codevector_dim": 256,
|
| 15 |
+
"contrastive_logits_temperature": 0.1,
|
| 16 |
+
"conv_bias": false,
|
| 17 |
+
"conv_dim": [
|
| 18 |
+
512,
|
| 19 |
+
512,
|
| 20 |
+
512,
|
| 21 |
+
512,
|
| 22 |
+
512,
|
| 23 |
+
512,
|
| 24 |
+
512
|
| 25 |
+
],
|
| 26 |
+
"conv_kernel": [
|
| 27 |
+
10,
|
| 28 |
+
3,
|
| 29 |
+
3,
|
| 30 |
+
3,
|
| 31 |
+
3,
|
| 32 |
+
2,
|
| 33 |
+
2
|
| 34 |
+
],
|
| 35 |
+
"conv_stride": [
|
| 36 |
+
5,
|
| 37 |
+
2,
|
| 38 |
+
2,
|
| 39 |
+
2,
|
| 40 |
+
2,
|
| 41 |
+
2,
|
| 42 |
+
2
|
| 43 |
+
],
|
| 44 |
+
"ctc_loss_reduction": "sum",
|
| 45 |
+
"ctc_zero_infinity": false,
|
| 46 |
+
"diversity_loss_weight": 0.1,
|
| 47 |
+
"do_stable_layer_norm": false,
|
| 48 |
+
"dtype": "float32",
|
| 49 |
+
"eos_token_id": 2,
|
| 50 |
+
"feat_extract_activation": "gelu",
|
| 51 |
+
"feat_extract_norm": "group",
|
| 52 |
+
"feat_proj_dropout": 0.1,
|
| 53 |
+
"feat_quantizer_dropout": 0.0,
|
| 54 |
+
"final_dropout": 0.0,
|
| 55 |
+
"finetuning_task": "wav2vec2_emotion",
|
| 56 |
+
"freeze_feat_extract_train": true,
|
| 57 |
+
"hidden_act": "gelu",
|
| 58 |
+
"hidden_dropout": 0.1,
|
| 59 |
+
"hidden_size": 768,
|
| 60 |
+
"id2label": {
|
| 61 |
+
"0": "neutral",
|
| 62 |
+
"1": "calm",
|
| 63 |
+
"2": "happy",
|
| 64 |
+
"3": "sad",
|
| 65 |
+
"4": "angry",
|
| 66 |
+
"5": "fearful",
|
| 67 |
+
"6": "disgust",
|
| 68 |
+
"7": "surprised"
|
| 69 |
+
},
|
| 70 |
+
"initializer_range": 0.02,
|
| 71 |
+
"intermediate_size": 3072,
|
| 72 |
+
"label2id": {
|
| 73 |
+
"angry": 4,
|
| 74 |
+
"calm": 1,
|
| 75 |
+
"disgust": 6,
|
| 76 |
+
"fearful": 5,
|
| 77 |
+
"happy": 2,
|
| 78 |
+
"neutral": 0,
|
| 79 |
+
"sad": 3,
|
| 80 |
+
"surprised": 7
|
| 81 |
+
},
|
| 82 |
+
"layer_norm_eps": 1e-05,
|
| 83 |
+
"layerdrop": 0.05,
|
| 84 |
+
"mask_channel_length": 10,
|
| 85 |
+
"mask_channel_min_space": 1,
|
| 86 |
+
"mask_channel_other": 0.0,
|
| 87 |
+
"mask_channel_prob": 0.0,
|
| 88 |
+
"mask_channel_selection": "static",
|
| 89 |
+
"mask_feature_length": 10,
|
| 90 |
+
"mask_feature_min_masks": 0,
|
| 91 |
+
"mask_feature_prob": 0.0,
|
| 92 |
+
"mask_time_length": 10,
|
| 93 |
+
"mask_time_min_masks": 2,
|
| 94 |
+
"mask_time_min_space": 1,
|
| 95 |
+
"mask_time_other": 0.0,
|
| 96 |
+
"mask_time_prob": 0.05,
|
| 97 |
+
"mask_time_selection": "static",
|
| 98 |
+
"model_type": "wav2vec2",
|
| 99 |
+
"no_mask_channel_overlap": false,
|
| 100 |
+
"no_mask_time_overlap": false,
|
| 101 |
+
"num_adapter_layers": 3,
|
| 102 |
+
"num_attention_heads": 12,
|
| 103 |
+
"num_codevector_groups": 2,
|
| 104 |
+
"num_codevectors_per_group": 320,
|
| 105 |
+
"num_conv_pos_embedding_groups": 16,
|
| 106 |
+
"num_conv_pos_embeddings": 128,
|
| 107 |
+
"num_feat_extract_layers": 7,
|
| 108 |
+
"num_hidden_layers": 12,
|
| 109 |
+
"num_negatives": 100,
|
| 110 |
+
"output_hidden_size": 768,
|
| 111 |
+
"pad_token_id": 0,
|
| 112 |
+
"proj_codevector_dim": 256,
|
| 113 |
+
"tdnn_dilation": [
|
| 114 |
+
1,
|
| 115 |
+
2,
|
| 116 |
+
3,
|
| 117 |
+
1,
|
| 118 |
+
1
|
| 119 |
+
],
|
| 120 |
+
"tdnn_dim": [
|
| 121 |
+
512,
|
| 122 |
+
512,
|
| 123 |
+
512,
|
| 124 |
+
512,
|
| 125 |
+
1500
|
| 126 |
+
],
|
| 127 |
+
"tdnn_kernel": [
|
| 128 |
+
5,
|
| 129 |
+
3,
|
| 130 |
+
3,
|
| 131 |
+
1,
|
| 132 |
+
1
|
| 133 |
+
],
|
| 134 |
+
"transformers_version": "4.57.1",
|
| 135 |
+
"use_cache": false,
|
| 136 |
+
"use_weighted_layer_sum": true,
|
| 137 |
+
"vocab_size": 32,
|
| 138 |
+
"xvector_output_dim": 512
|
| 139 |
+
}
|
checkpoint-6000/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2600128837f8b74885da533a0bcf1ef45d52254cf0dd27189add1b1e9fb9232d
|
| 3 |
+
size 378308660
|
checkpoint-6000/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:069ce2409b38d224f9f1c44b6441346f095b71bf488047de5d9378259f81012e
|
| 3 |
+
size 756740235
|
checkpoint-6000/preprocessor_config.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_normalize": false,
|
| 3 |
+
"feature_extractor_type": "Wav2Vec2FeatureExtractor",
|
| 4 |
+
"feature_size": 1,
|
| 5 |
+
"padding_side": "right",
|
| 6 |
+
"padding_value": 0.0,
|
| 7 |
+
"return_attention_mask": true,
|
| 8 |
+
"sampling_rate": 16000
|
| 9 |
+
}
|
checkpoint-6000/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bb17063f42a2fac2dcd80e95e1c31b027b16ae036530a4afc492a0e081337eba
|
| 3 |
+
size 14455
|
checkpoint-6000/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:392e86059c3103bc98ccb8e09eeed1120b322700161c67458aa573bdab39ed31
|
| 3 |
+
size 1465
|
checkpoint-6000/trainer_state.json
ADDED
|
@@ -0,0 +1,1714 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 21.05263157894737,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 6000,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.08771929824561403,
|
| 14 |
+
"grad_norm": 14.743780136108398,
|
| 15 |
+
"learning_rate": 1.4942583732057417e-05,
|
| 16 |
+
"loss": 0.9366,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.17543859649122806,
|
| 21 |
+
"grad_norm": 22.20017433166504,
|
| 22 |
+
"learning_rate": 1.4882775119617225e-05,
|
| 23 |
+
"loss": 1.2015,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.2631578947368421,
|
| 28 |
+
"grad_norm": 25.090787887573242,
|
| 29 |
+
"learning_rate": 1.4822966507177032e-05,
|
| 30 |
+
"loss": 1.0156,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.3508771929824561,
|
| 35 |
+
"grad_norm": 13.580609321594238,
|
| 36 |
+
"learning_rate": 1.4763157894736842e-05,
|
| 37 |
+
"loss": 1.117,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.43859649122807015,
|
| 42 |
+
"grad_norm": 12.062329292297363,
|
| 43 |
+
"learning_rate": 1.470334928229665e-05,
|
| 44 |
+
"loss": 0.8312,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.5263157894736842,
|
| 49 |
+
"grad_norm": 18.625316619873047,
|
| 50 |
+
"learning_rate": 1.464354066985646e-05,
|
| 51 |
+
"loss": 0.9938,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.6140350877192983,
|
| 56 |
+
"grad_norm": 11.628118515014648,
|
| 57 |
+
"learning_rate": 1.4583732057416269e-05,
|
| 58 |
+
"loss": 0.8922,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.7017543859649122,
|
| 63 |
+
"grad_norm": 12.121639251708984,
|
| 64 |
+
"learning_rate": 1.4523923444976078e-05,
|
| 65 |
+
"loss": 0.9787,
|
| 66 |
+
"step": 200
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 0.7894736842105263,
|
| 70 |
+
"grad_norm": 14.091375350952148,
|
| 71 |
+
"learning_rate": 1.4464114832535886e-05,
|
| 72 |
+
"loss": 0.806,
|
| 73 |
+
"step": 225
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.8771929824561403,
|
| 77 |
+
"grad_norm": 9.919096946716309,
|
| 78 |
+
"learning_rate": 1.4404306220095696e-05,
|
| 79 |
+
"loss": 0.9935,
|
| 80 |
+
"step": 250
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.9649122807017544,
|
| 84 |
+
"grad_norm": 13.80988597869873,
|
| 85 |
+
"learning_rate": 1.4344497607655503e-05,
|
| 86 |
+
"loss": 0.8116,
|
| 87 |
+
"step": 275
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 1.0526315789473684,
|
| 91 |
+
"grad_norm": 13.865740776062012,
|
| 92 |
+
"learning_rate": 1.4284688995215311e-05,
|
| 93 |
+
"loss": 0.7288,
|
| 94 |
+
"step": 300
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 1.1403508771929824,
|
| 98 |
+
"grad_norm": 9.76303768157959,
|
| 99 |
+
"learning_rate": 1.4224880382775121e-05,
|
| 100 |
+
"loss": 0.8283,
|
| 101 |
+
"step": 325
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 1.2280701754385965,
|
| 105 |
+
"grad_norm": 9.364381790161133,
|
| 106 |
+
"learning_rate": 1.4165071770334929e-05,
|
| 107 |
+
"loss": 0.6827,
|
| 108 |
+
"step": 350
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 1.3157894736842106,
|
| 112 |
+
"grad_norm": 13.79258918762207,
|
| 113 |
+
"learning_rate": 1.4105263157894738e-05,
|
| 114 |
+
"loss": 0.7745,
|
| 115 |
+
"step": 375
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 1.4035087719298245,
|
| 119 |
+
"grad_norm": 9.642690658569336,
|
| 120 |
+
"learning_rate": 1.4045454545454546e-05,
|
| 121 |
+
"loss": 0.7514,
|
| 122 |
+
"step": 400
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 1.4912280701754386,
|
| 126 |
+
"grad_norm": 14.490239143371582,
|
| 127 |
+
"learning_rate": 1.3985645933014356e-05,
|
| 128 |
+
"loss": 0.6594,
|
| 129 |
+
"step": 425
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 1.5789473684210527,
|
| 133 |
+
"grad_norm": 13.364012718200684,
|
| 134 |
+
"learning_rate": 1.3925837320574163e-05,
|
| 135 |
+
"loss": 0.6036,
|
| 136 |
+
"step": 450
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 1.6666666666666665,
|
| 140 |
+
"grad_norm": 6.50977087020874,
|
| 141 |
+
"learning_rate": 1.3866028708133971e-05,
|
| 142 |
+
"loss": 0.7616,
|
| 143 |
+
"step": 475
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 1.7543859649122808,
|
| 147 |
+
"grad_norm": 30.46417808532715,
|
| 148 |
+
"learning_rate": 1.380622009569378e-05,
|
| 149 |
+
"loss": 0.6735,
|
| 150 |
+
"step": 500
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 1.8421052631578947,
|
| 154 |
+
"grad_norm": 17.846593856811523,
|
| 155 |
+
"learning_rate": 1.3746411483253589e-05,
|
| 156 |
+
"loss": 0.5902,
|
| 157 |
+
"step": 525
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 1.9298245614035088,
|
| 161 |
+
"grad_norm": 19.83995819091797,
|
| 162 |
+
"learning_rate": 1.3686602870813398e-05,
|
| 163 |
+
"loss": 0.5928,
|
| 164 |
+
"step": 550
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 2.017543859649123,
|
| 168 |
+
"grad_norm": 6.053647994995117,
|
| 169 |
+
"learning_rate": 1.3626794258373206e-05,
|
| 170 |
+
"loss": 0.5816,
|
| 171 |
+
"step": 575
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 2.1052631578947367,
|
| 175 |
+
"grad_norm": 9.594230651855469,
|
| 176 |
+
"learning_rate": 1.3566985645933015e-05,
|
| 177 |
+
"loss": 0.488,
|
| 178 |
+
"step": 600
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 2.192982456140351,
|
| 182 |
+
"grad_norm": 13.84262752532959,
|
| 183 |
+
"learning_rate": 1.3507177033492823e-05,
|
| 184 |
+
"loss": 0.568,
|
| 185 |
+
"step": 625
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 2.280701754385965,
|
| 189 |
+
"grad_norm": 10.86662483215332,
|
| 190 |
+
"learning_rate": 1.3447368421052633e-05,
|
| 191 |
+
"loss": 0.4017,
|
| 192 |
+
"step": 650
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 2.3684210526315788,
|
| 196 |
+
"grad_norm": 16.06069564819336,
|
| 197 |
+
"learning_rate": 1.338755980861244e-05,
|
| 198 |
+
"loss": 0.5112,
|
| 199 |
+
"step": 675
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 2.456140350877193,
|
| 203 |
+
"grad_norm": 6.421328544616699,
|
| 204 |
+
"learning_rate": 1.3327751196172248e-05,
|
| 205 |
+
"loss": 0.3609,
|
| 206 |
+
"step": 700
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 2.543859649122807,
|
| 210 |
+
"grad_norm": 6.225348949432373,
|
| 211 |
+
"learning_rate": 1.3267942583732058e-05,
|
| 212 |
+
"loss": 0.5531,
|
| 213 |
+
"step": 725
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 2.6315789473684212,
|
| 217 |
+
"grad_norm": 14.942207336425781,
|
| 218 |
+
"learning_rate": 1.3208133971291866e-05,
|
| 219 |
+
"loss": 0.4272,
|
| 220 |
+
"step": 750
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 2.719298245614035,
|
| 224 |
+
"grad_norm": 8.975879669189453,
|
| 225 |
+
"learning_rate": 1.3148325358851675e-05,
|
| 226 |
+
"loss": 0.4509,
|
| 227 |
+
"step": 775
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 2.807017543859649,
|
| 231 |
+
"grad_norm": 21.482088088989258,
|
| 232 |
+
"learning_rate": 1.3088516746411483e-05,
|
| 233 |
+
"loss": 0.4233,
|
| 234 |
+
"step": 800
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 2.8947368421052633,
|
| 238 |
+
"grad_norm": 5.1405792236328125,
|
| 239 |
+
"learning_rate": 1.3028708133971293e-05,
|
| 240 |
+
"loss": 0.418,
|
| 241 |
+
"step": 825
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 2.982456140350877,
|
| 245 |
+
"grad_norm": 14.574483871459961,
|
| 246 |
+
"learning_rate": 1.29688995215311e-05,
|
| 247 |
+
"loss": 0.6778,
|
| 248 |
+
"step": 850
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 3.0701754385964914,
|
| 252 |
+
"grad_norm": 7.385740280151367,
|
| 253 |
+
"learning_rate": 1.290909090909091e-05,
|
| 254 |
+
"loss": 0.4127,
|
| 255 |
+
"step": 875
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 3.1578947368421053,
|
| 259 |
+
"grad_norm": 48.35734939575195,
|
| 260 |
+
"learning_rate": 1.2849282296650718e-05,
|
| 261 |
+
"loss": 0.4752,
|
| 262 |
+
"step": 900
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 3.245614035087719,
|
| 266 |
+
"grad_norm": 8.056925773620605,
|
| 267 |
+
"learning_rate": 1.2789473684210526e-05,
|
| 268 |
+
"loss": 0.411,
|
| 269 |
+
"step": 925
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 3.3333333333333335,
|
| 273 |
+
"grad_norm": 20.510982513427734,
|
| 274 |
+
"learning_rate": 1.2729665071770335e-05,
|
| 275 |
+
"loss": 0.495,
|
| 276 |
+
"step": 950
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 3.4210526315789473,
|
| 280 |
+
"grad_norm": 25.063684463500977,
|
| 281 |
+
"learning_rate": 1.2669856459330143e-05,
|
| 282 |
+
"loss": 0.2436,
|
| 283 |
+
"step": 975
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 3.5087719298245617,
|
| 287 |
+
"grad_norm": 15.523259162902832,
|
| 288 |
+
"learning_rate": 1.2610047846889953e-05,
|
| 289 |
+
"loss": 0.4612,
|
| 290 |
+
"step": 1000
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 3.5964912280701755,
|
| 294 |
+
"grad_norm": 1.9646857976913452,
|
| 295 |
+
"learning_rate": 1.255023923444976e-05,
|
| 296 |
+
"loss": 0.4426,
|
| 297 |
+
"step": 1025
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 3.6842105263157894,
|
| 301 |
+
"grad_norm": 30.676250457763672,
|
| 302 |
+
"learning_rate": 1.249043062200957e-05,
|
| 303 |
+
"loss": 0.3333,
|
| 304 |
+
"step": 1050
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 3.7719298245614032,
|
| 308 |
+
"grad_norm": 6.475268363952637,
|
| 309 |
+
"learning_rate": 1.2430622009569378e-05,
|
| 310 |
+
"loss": 0.2945,
|
| 311 |
+
"step": 1075
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 3.8596491228070176,
|
| 315 |
+
"grad_norm": 24.684438705444336,
|
| 316 |
+
"learning_rate": 1.2370813397129186e-05,
|
| 317 |
+
"loss": 0.4339,
|
| 318 |
+
"step": 1100
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 3.9473684210526314,
|
| 322 |
+
"grad_norm": 38.601226806640625,
|
| 323 |
+
"learning_rate": 1.2311004784688995e-05,
|
| 324 |
+
"loss": 0.3701,
|
| 325 |
+
"step": 1125
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 4.035087719298246,
|
| 329 |
+
"grad_norm": 2.8431146144866943,
|
| 330 |
+
"learning_rate": 1.2251196172248803e-05,
|
| 331 |
+
"loss": 0.2782,
|
| 332 |
+
"step": 1150
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 4.12280701754386,
|
| 336 |
+
"grad_norm": 23.384532928466797,
|
| 337 |
+
"learning_rate": 1.2191387559808613e-05,
|
| 338 |
+
"loss": 0.312,
|
| 339 |
+
"step": 1175
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 4.2105263157894735,
|
| 343 |
+
"grad_norm": 3.222729444503784,
|
| 344 |
+
"learning_rate": 1.213157894736842e-05,
|
| 345 |
+
"loss": 0.264,
|
| 346 |
+
"step": 1200
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 4.298245614035087,
|
| 350 |
+
"grad_norm": 3.5861947536468506,
|
| 351 |
+
"learning_rate": 1.207177033492823e-05,
|
| 352 |
+
"loss": 0.3104,
|
| 353 |
+
"step": 1225
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 4.385964912280702,
|
| 357 |
+
"grad_norm": 4.091048717498779,
|
| 358 |
+
"learning_rate": 1.2011961722488038e-05,
|
| 359 |
+
"loss": 0.3567,
|
| 360 |
+
"step": 1250
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 4.473684210526316,
|
| 364 |
+
"grad_norm": 1.3814598321914673,
|
| 365 |
+
"learning_rate": 1.1952153110047847e-05,
|
| 366 |
+
"loss": 0.3455,
|
| 367 |
+
"step": 1275
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 4.56140350877193,
|
| 371 |
+
"grad_norm": 2.480316162109375,
|
| 372 |
+
"learning_rate": 1.1892344497607655e-05,
|
| 373 |
+
"loss": 0.1769,
|
| 374 |
+
"step": 1300
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 4.649122807017544,
|
| 378 |
+
"grad_norm": 20.069162368774414,
|
| 379 |
+
"learning_rate": 1.1832535885167463e-05,
|
| 380 |
+
"loss": 0.2483,
|
| 381 |
+
"step": 1325
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 4.7368421052631575,
|
| 385 |
+
"grad_norm": 1.7885949611663818,
|
| 386 |
+
"learning_rate": 1.1772727272727272e-05,
|
| 387 |
+
"loss": 0.2283,
|
| 388 |
+
"step": 1350
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 4.824561403508772,
|
| 392 |
+
"grad_norm": 40.37774658203125,
|
| 393 |
+
"learning_rate": 1.171291866028708e-05,
|
| 394 |
+
"loss": 0.2277,
|
| 395 |
+
"step": 1375
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 4.912280701754386,
|
| 399 |
+
"grad_norm": 2.960739850997925,
|
| 400 |
+
"learning_rate": 1.1653110047846891e-05,
|
| 401 |
+
"loss": 0.2077,
|
| 402 |
+
"step": 1400
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 5.0,
|
| 406 |
+
"grad_norm": 14.10015869140625,
|
| 407 |
+
"learning_rate": 1.15933014354067e-05,
|
| 408 |
+
"loss": 0.3859,
|
| 409 |
+
"step": 1425
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 5.087719298245614,
|
| 413 |
+
"grad_norm": 4.440461158752441,
|
| 414 |
+
"learning_rate": 1.1533492822966509e-05,
|
| 415 |
+
"loss": 0.1354,
|
| 416 |
+
"step": 1450
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 5.175438596491228,
|
| 420 |
+
"grad_norm": 17.193849563598633,
|
| 421 |
+
"learning_rate": 1.1473684210526317e-05,
|
| 422 |
+
"loss": 0.2935,
|
| 423 |
+
"step": 1475
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 5.2631578947368425,
|
| 427 |
+
"grad_norm": 1.2828125953674316,
|
| 428 |
+
"learning_rate": 1.1413875598086125e-05,
|
| 429 |
+
"loss": 0.1394,
|
| 430 |
+
"step": 1500
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 5.350877192982456,
|
| 434 |
+
"grad_norm": 4.809077262878418,
|
| 435 |
+
"learning_rate": 1.1354066985645934e-05,
|
| 436 |
+
"loss": 0.3651,
|
| 437 |
+
"step": 1525
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 5.43859649122807,
|
| 441 |
+
"grad_norm": 43.53572463989258,
|
| 442 |
+
"learning_rate": 1.1294258373205742e-05,
|
| 443 |
+
"loss": 0.1187,
|
| 444 |
+
"step": 1550
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 5.526315789473684,
|
| 448 |
+
"grad_norm": 5.647329330444336,
|
| 449 |
+
"learning_rate": 1.1234449760765551e-05,
|
| 450 |
+
"loss": 0.2719,
|
| 451 |
+
"step": 1575
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 5.614035087719298,
|
| 455 |
+
"grad_norm": 25.37187957763672,
|
| 456 |
+
"learning_rate": 1.117464114832536e-05,
|
| 457 |
+
"loss": 0.2477,
|
| 458 |
+
"step": 1600
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 5.701754385964913,
|
| 462 |
+
"grad_norm": 4.228896617889404,
|
| 463 |
+
"learning_rate": 1.1114832535885169e-05,
|
| 464 |
+
"loss": 0.3608,
|
| 465 |
+
"step": 1625
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 5.7894736842105265,
|
| 469 |
+
"grad_norm": 2.1760356426239014,
|
| 470 |
+
"learning_rate": 1.1055023923444977e-05,
|
| 471 |
+
"loss": 0.1223,
|
| 472 |
+
"step": 1650
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 5.87719298245614,
|
| 476 |
+
"grad_norm": 28.078975677490234,
|
| 477 |
+
"learning_rate": 1.0995215311004786e-05,
|
| 478 |
+
"loss": 0.3207,
|
| 479 |
+
"step": 1675
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 5.964912280701754,
|
| 483 |
+
"grad_norm": 12.583624839782715,
|
| 484 |
+
"learning_rate": 1.0935406698564594e-05,
|
| 485 |
+
"loss": 0.1094,
|
| 486 |
+
"step": 1700
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 6.052631578947368,
|
| 490 |
+
"grad_norm": 0.4104742109775543,
|
| 491 |
+
"learning_rate": 1.0875598086124402e-05,
|
| 492 |
+
"loss": 0.1937,
|
| 493 |
+
"step": 1725
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 6.140350877192983,
|
| 497 |
+
"grad_norm": 1.5135095119476318,
|
| 498 |
+
"learning_rate": 1.0815789473684211e-05,
|
| 499 |
+
"loss": 0.3113,
|
| 500 |
+
"step": 1750
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 6.228070175438597,
|
| 504 |
+
"grad_norm": 1.9227218627929688,
|
| 505 |
+
"learning_rate": 1.0755980861244019e-05,
|
| 506 |
+
"loss": 0.1179,
|
| 507 |
+
"step": 1775
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 6.315789473684211,
|
| 511 |
+
"grad_norm": 7.322267055511475,
|
| 512 |
+
"learning_rate": 1.0696172248803829e-05,
|
| 513 |
+
"loss": 0.209,
|
| 514 |
+
"step": 1800
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 6.4035087719298245,
|
| 518 |
+
"grad_norm": 2.8994693756103516,
|
| 519 |
+
"learning_rate": 1.0636363636363636e-05,
|
| 520 |
+
"loss": 0.1743,
|
| 521 |
+
"step": 1825
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 6.491228070175438,
|
| 525 |
+
"grad_norm": 21.65730857849121,
|
| 526 |
+
"learning_rate": 1.0576555023923446e-05,
|
| 527 |
+
"loss": 0.2483,
|
| 528 |
+
"step": 1850
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 6.578947368421053,
|
| 532 |
+
"grad_norm": 24.210969924926758,
|
| 533 |
+
"learning_rate": 1.0516746411483254e-05,
|
| 534 |
+
"loss": 0.1348,
|
| 535 |
+
"step": 1875
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 6.666666666666667,
|
| 539 |
+
"grad_norm": 1.589753270149231,
|
| 540 |
+
"learning_rate": 1.0456937799043063e-05,
|
| 541 |
+
"loss": 0.229,
|
| 542 |
+
"step": 1900
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 6.754385964912281,
|
| 546 |
+
"grad_norm": 3.508387327194214,
|
| 547 |
+
"learning_rate": 1.0397129186602871e-05,
|
| 548 |
+
"loss": 0.0909,
|
| 549 |
+
"step": 1925
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 6.842105263157895,
|
| 553 |
+
"grad_norm": 1.89371919631958,
|
| 554 |
+
"learning_rate": 1.0337320574162679e-05,
|
| 555 |
+
"loss": 0.1859,
|
| 556 |
+
"step": 1950
|
| 557 |
+
},
|
| 558 |
+
{
|
| 559 |
+
"epoch": 6.9298245614035086,
|
| 560 |
+
"grad_norm": 0.35574591159820557,
|
| 561 |
+
"learning_rate": 1.0277511961722489e-05,
|
| 562 |
+
"loss": 0.1675,
|
| 563 |
+
"step": 1975
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 7.017543859649122,
|
| 567 |
+
"grad_norm": 0.621152937412262,
|
| 568 |
+
"learning_rate": 1.0217703349282296e-05,
|
| 569 |
+
"loss": 0.1082,
|
| 570 |
+
"step": 2000
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 7.105263157894737,
|
| 574 |
+
"grad_norm": 0.5694774985313416,
|
| 575 |
+
"learning_rate": 1.0157894736842106e-05,
|
| 576 |
+
"loss": 0.0886,
|
| 577 |
+
"step": 2025
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 7.192982456140351,
|
| 581 |
+
"grad_norm": 7.276403427124023,
|
| 582 |
+
"learning_rate": 1.0098086124401914e-05,
|
| 583 |
+
"loss": 0.2697,
|
| 584 |
+
"step": 2050
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 7.280701754385965,
|
| 588 |
+
"grad_norm": 0.4015137851238251,
|
| 589 |
+
"learning_rate": 1.0038277511961723e-05,
|
| 590 |
+
"loss": 0.1301,
|
| 591 |
+
"step": 2075
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 7.368421052631579,
|
| 595 |
+
"grad_norm": 0.27876588702201843,
|
| 596 |
+
"learning_rate": 9.978468899521531e-06,
|
| 597 |
+
"loss": 0.0687,
|
| 598 |
+
"step": 2100
|
| 599 |
+
},
|
| 600 |
+
{
|
| 601 |
+
"epoch": 7.456140350877193,
|
| 602 |
+
"grad_norm": 7.235930442810059,
|
| 603 |
+
"learning_rate": 9.918660287081339e-06,
|
| 604 |
+
"loss": 0.0734,
|
| 605 |
+
"step": 2125
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 7.543859649122807,
|
| 609 |
+
"grad_norm": 1.6751177310943604,
|
| 610 |
+
"learning_rate": 9.858851674641148e-06,
|
| 611 |
+
"loss": 0.1574,
|
| 612 |
+
"step": 2150
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 7.631578947368421,
|
| 616 |
+
"grad_norm": 0.8539127111434937,
|
| 617 |
+
"learning_rate": 9.799043062200956e-06,
|
| 618 |
+
"loss": 0.0483,
|
| 619 |
+
"step": 2175
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"epoch": 7.719298245614035,
|
| 623 |
+
"grad_norm": 0.26878660917282104,
|
| 624 |
+
"learning_rate": 9.739234449760766e-06,
|
| 625 |
+
"loss": 0.1632,
|
| 626 |
+
"step": 2200
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 7.807017543859649,
|
| 630 |
+
"grad_norm": 106.50234985351562,
|
| 631 |
+
"learning_rate": 9.679425837320574e-06,
|
| 632 |
+
"loss": 0.1156,
|
| 633 |
+
"step": 2225
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 7.894736842105263,
|
| 637 |
+
"grad_norm": 2.002487897872925,
|
| 638 |
+
"learning_rate": 9.619617224880383e-06,
|
| 639 |
+
"loss": 0.2036,
|
| 640 |
+
"step": 2250
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 7.982456140350877,
|
| 644 |
+
"grad_norm": 0.5302895307540894,
|
| 645 |
+
"learning_rate": 9.559808612440191e-06,
|
| 646 |
+
"loss": 0.1417,
|
| 647 |
+
"step": 2275
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 8.070175438596491,
|
| 651 |
+
"grad_norm": 0.314165323972702,
|
| 652 |
+
"learning_rate": 9.5e-06,
|
| 653 |
+
"loss": 0.0673,
|
| 654 |
+
"step": 2300
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 8.157894736842104,
|
| 658 |
+
"grad_norm": 0.5242946743965149,
|
| 659 |
+
"learning_rate": 9.440191387559808e-06,
|
| 660 |
+
"loss": 0.0689,
|
| 661 |
+
"step": 2325
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 8.24561403508772,
|
| 665 |
+
"grad_norm": 48.71388244628906,
|
| 666 |
+
"learning_rate": 9.380382775119616e-06,
|
| 667 |
+
"loss": 0.1863,
|
| 668 |
+
"step": 2350
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 8.333333333333334,
|
| 672 |
+
"grad_norm": 0.5670514106750488,
|
| 673 |
+
"learning_rate": 9.320574162679426e-06,
|
| 674 |
+
"loss": 0.0823,
|
| 675 |
+
"step": 2375
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 8.421052631578947,
|
| 679 |
+
"grad_norm": 0.6025238633155823,
|
| 680 |
+
"learning_rate": 9.260765550239234e-06,
|
| 681 |
+
"loss": 0.1118,
|
| 682 |
+
"step": 2400
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 8.508771929824562,
|
| 686 |
+
"grad_norm": 42.649696350097656,
|
| 687 |
+
"learning_rate": 9.200956937799043e-06,
|
| 688 |
+
"loss": 0.1333,
|
| 689 |
+
"step": 2425
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 8.596491228070175,
|
| 693 |
+
"grad_norm": 0.4079756438732147,
|
| 694 |
+
"learning_rate": 9.141148325358851e-06,
|
| 695 |
+
"loss": 0.156,
|
| 696 |
+
"step": 2450
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 8.68421052631579,
|
| 700 |
+
"grad_norm": 0.31388023495674133,
|
| 701 |
+
"learning_rate": 9.08133971291866e-06,
|
| 702 |
+
"loss": 0.2159,
|
| 703 |
+
"step": 2475
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 8.771929824561404,
|
| 707 |
+
"grad_norm": 0.2017977237701416,
|
| 708 |
+
"learning_rate": 9.021531100478468e-06,
|
| 709 |
+
"loss": 0.0936,
|
| 710 |
+
"step": 2500
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 8.859649122807017,
|
| 714 |
+
"grad_norm": 2.2358014583587646,
|
| 715 |
+
"learning_rate": 6.697368421052631e-06,
|
| 716 |
+
"loss": 0.1884,
|
| 717 |
+
"step": 2525
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 8.947368421052632,
|
| 721 |
+
"grad_norm": 0.5669077038764954,
|
| 722 |
+
"learning_rate": 6.615131578947369e-06,
|
| 723 |
+
"loss": 0.2183,
|
| 724 |
+
"step": 2550
|
| 725 |
+
},
|
| 726 |
+
{
|
| 727 |
+
"epoch": 9.035087719298245,
|
| 728 |
+
"grad_norm": 27.18621253967285,
|
| 729 |
+
"learning_rate": 6.532894736842106e-06,
|
| 730 |
+
"loss": 0.1232,
|
| 731 |
+
"step": 2575
|
| 732 |
+
},
|
| 733 |
+
{
|
| 734 |
+
"epoch": 9.12280701754386,
|
| 735 |
+
"grad_norm": 468.5591735839844,
|
| 736 |
+
"learning_rate": 6.450657894736842e-06,
|
| 737 |
+
"loss": 0.0516,
|
| 738 |
+
"step": 2600
|
| 739 |
+
},
|
| 740 |
+
{
|
| 741 |
+
"epoch": 9.210526315789474,
|
| 742 |
+
"grad_norm": 0.24007272720336914,
|
| 743 |
+
"learning_rate": 6.368421052631579e-06,
|
| 744 |
+
"loss": 0.0695,
|
| 745 |
+
"step": 2625
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 9.298245614035087,
|
| 749 |
+
"grad_norm": 99.38546752929688,
|
| 750 |
+
"learning_rate": 6.286184210526316e-06,
|
| 751 |
+
"loss": 0.0947,
|
| 752 |
+
"step": 2650
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 9.385964912280702,
|
| 756 |
+
"grad_norm": 3.8673267364501953,
|
| 757 |
+
"learning_rate": 6.203947368421053e-06,
|
| 758 |
+
"loss": 0.1062,
|
| 759 |
+
"step": 2675
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 9.473684210526315,
|
| 763 |
+
"grad_norm": 0.2559914290904999,
|
| 764 |
+
"learning_rate": 6.12171052631579e-06,
|
| 765 |
+
"loss": 0.0523,
|
| 766 |
+
"step": 2700
|
| 767 |
+
},
|
| 768 |
+
{
|
| 769 |
+
"epoch": 9.56140350877193,
|
| 770 |
+
"grad_norm": 0.12371934205293655,
|
| 771 |
+
"learning_rate": 6.039473684210526e-06,
|
| 772 |
+
"loss": 0.0495,
|
| 773 |
+
"step": 2725
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"epoch": 9.649122807017545,
|
| 777 |
+
"grad_norm": 2.5249898433685303,
|
| 778 |
+
"learning_rate": 5.957236842105263e-06,
|
| 779 |
+
"loss": 0.1678,
|
| 780 |
+
"step": 2750
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"epoch": 9.736842105263158,
|
| 784 |
+
"grad_norm": 0.2619442939758301,
|
| 785 |
+
"learning_rate": 5.8750000000000005e-06,
|
| 786 |
+
"loss": 0.0534,
|
| 787 |
+
"step": 2775
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"epoch": 9.824561403508772,
|
| 791 |
+
"grad_norm": 1.3705483675003052,
|
| 792 |
+
"learning_rate": 5.792763157894737e-06,
|
| 793 |
+
"loss": 0.1326,
|
| 794 |
+
"step": 2800
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 9.912280701754385,
|
| 798 |
+
"grad_norm": 74.48184967041016,
|
| 799 |
+
"learning_rate": 5.710526315789474e-06,
|
| 800 |
+
"loss": 0.1795,
|
| 801 |
+
"step": 2825
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 10.0,
|
| 805 |
+
"grad_norm": 79.78992462158203,
|
| 806 |
+
"learning_rate": 5.6282894736842106e-06,
|
| 807 |
+
"loss": 0.092,
|
| 808 |
+
"step": 2850
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"epoch": 10.087719298245615,
|
| 812 |
+
"grad_norm": 0.5286645889282227,
|
| 813 |
+
"learning_rate": 5.546052631578947e-06,
|
| 814 |
+
"loss": 0.0418,
|
| 815 |
+
"step": 2875
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"epoch": 10.175438596491228,
|
| 819 |
+
"grad_norm": 109.82157135009766,
|
| 820 |
+
"learning_rate": 5.463815789473684e-06,
|
| 821 |
+
"loss": 0.193,
|
| 822 |
+
"step": 2900
|
| 823 |
+
},
|
| 824 |
+
{
|
| 825 |
+
"epoch": 10.263157894736842,
|
| 826 |
+
"grad_norm": 3.963012933731079,
|
| 827 |
+
"learning_rate": 5.3815789473684215e-06,
|
| 828 |
+
"loss": 0.0897,
|
| 829 |
+
"step": 2925
|
| 830 |
+
},
|
| 831 |
+
{
|
| 832 |
+
"epoch": 10.350877192982455,
|
| 833 |
+
"grad_norm": 0.41540616750717163,
|
| 834 |
+
"learning_rate": 5.299342105263158e-06,
|
| 835 |
+
"loss": 0.1723,
|
| 836 |
+
"step": 2950
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 10.43859649122807,
|
| 840 |
+
"grad_norm": 5.900774002075195,
|
| 841 |
+
"learning_rate": 5.217105263157895e-06,
|
| 842 |
+
"loss": 0.0784,
|
| 843 |
+
"step": 2975
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 10.526315789473685,
|
| 847 |
+
"grad_norm": 12.152785301208496,
|
| 848 |
+
"learning_rate": 5.1348684210526315e-06,
|
| 849 |
+
"loss": 0.1186,
|
| 850 |
+
"step": 3000
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"epoch": 10.614035087719298,
|
| 854 |
+
"grad_norm": 0.21494676172733307,
|
| 855 |
+
"learning_rate": 3.6315789473684213e-06,
|
| 856 |
+
"loss": 0.0358,
|
| 857 |
+
"step": 3025
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 10.701754385964913,
|
| 861 |
+
"grad_norm": 90.48876953125,
|
| 862 |
+
"learning_rate": 3.537593984962406e-06,
|
| 863 |
+
"loss": 0.0471,
|
| 864 |
+
"step": 3050
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 10.789473684210526,
|
| 868 |
+
"grad_norm": 0.23261310160160065,
|
| 869 |
+
"learning_rate": 3.4436090225563912e-06,
|
| 870 |
+
"loss": 0.0276,
|
| 871 |
+
"step": 3075
|
| 872 |
+
},
|
| 873 |
+
{
|
| 874 |
+
"epoch": 10.87719298245614,
|
| 875 |
+
"grad_norm": 0.782564640045166,
|
| 876 |
+
"learning_rate": 3.349624060150376e-06,
|
| 877 |
+
"loss": 0.0718,
|
| 878 |
+
"step": 3100
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 10.964912280701755,
|
| 882 |
+
"grad_norm": 0.33198878169059753,
|
| 883 |
+
"learning_rate": 3.255639097744361e-06,
|
| 884 |
+
"loss": 0.0564,
|
| 885 |
+
"step": 3125
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 11.052631578947368,
|
| 889 |
+
"grad_norm": 0.14864841103553772,
|
| 890 |
+
"learning_rate": 3.161654135338346e-06,
|
| 891 |
+
"loss": 0.0508,
|
| 892 |
+
"step": 3150
|
| 893 |
+
},
|
| 894 |
+
{
|
| 895 |
+
"epoch": 11.140350877192983,
|
| 896 |
+
"grad_norm": 0.1588129848241806,
|
| 897 |
+
"learning_rate": 3.067669172932331e-06,
|
| 898 |
+
"loss": 0.1641,
|
| 899 |
+
"step": 3175
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 11.228070175438596,
|
| 903 |
+
"grad_norm": 0.6562497019767761,
|
| 904 |
+
"learning_rate": 2.973684210526316e-06,
|
| 905 |
+
"loss": 0.0544,
|
| 906 |
+
"step": 3200
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"epoch": 11.31578947368421,
|
| 910 |
+
"grad_norm": 0.3608649969100952,
|
| 911 |
+
"learning_rate": 2.879699248120301e-06,
|
| 912 |
+
"loss": 0.0391,
|
| 913 |
+
"step": 3225
|
| 914 |
+
},
|
| 915 |
+
{
|
| 916 |
+
"epoch": 11.403508771929825,
|
| 917 |
+
"grad_norm": 20.602819442749023,
|
| 918 |
+
"learning_rate": 2.785714285714286e-06,
|
| 919 |
+
"loss": 0.0338,
|
| 920 |
+
"step": 3250
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 11.491228070175438,
|
| 924 |
+
"grad_norm": 157.23171997070312,
|
| 925 |
+
"learning_rate": 2.691729323308271e-06,
|
| 926 |
+
"loss": 0.0771,
|
| 927 |
+
"step": 3275
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 11.578947368421053,
|
| 931 |
+
"grad_norm": 0.22338764369487762,
|
| 932 |
+
"learning_rate": 2.597744360902256e-06,
|
| 933 |
+
"loss": 0.1187,
|
| 934 |
+
"step": 3300
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 11.666666666666666,
|
| 938 |
+
"grad_norm": 57.78382110595703,
|
| 939 |
+
"learning_rate": 2.503759398496241e-06,
|
| 940 |
+
"loss": 0.0165,
|
| 941 |
+
"step": 3325
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 11.75438596491228,
|
| 945 |
+
"grad_norm": 7.400601863861084,
|
| 946 |
+
"learning_rate": 2.4097744360902257e-06,
|
| 947 |
+
"loss": 0.0092,
|
| 948 |
+
"step": 3350
|
| 949 |
+
},
|
| 950 |
+
{
|
| 951 |
+
"epoch": 11.842105263157894,
|
| 952 |
+
"grad_norm": 0.18190287053585052,
|
| 953 |
+
"learning_rate": 2.3157894736842105e-06,
|
| 954 |
+
"loss": 0.1059,
|
| 955 |
+
"step": 3375
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 11.929824561403509,
|
| 959 |
+
"grad_norm": 0.1269238144159317,
|
| 960 |
+
"learning_rate": 2.2218045112781957e-06,
|
| 961 |
+
"loss": 0.0571,
|
| 962 |
+
"step": 3400
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 12.017543859649123,
|
| 966 |
+
"grad_norm": 0.10843072831630707,
|
| 967 |
+
"learning_rate": 2.1278195488721805e-06,
|
| 968 |
+
"loss": 0.1341,
|
| 969 |
+
"step": 3425
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 12.105263157894736,
|
| 973 |
+
"grad_norm": 0.7199507355690002,
|
| 974 |
+
"learning_rate": 2.0338345864661656e-06,
|
| 975 |
+
"loss": 0.0063,
|
| 976 |
+
"step": 3450
|
| 977 |
+
},
|
| 978 |
+
{
|
| 979 |
+
"epoch": 12.192982456140351,
|
| 980 |
+
"grad_norm": 0.19009971618652344,
|
| 981 |
+
"learning_rate": 1.9398496240601504e-06,
|
| 982 |
+
"loss": 0.054,
|
| 983 |
+
"step": 3475
|
| 984 |
+
},
|
| 985 |
+
{
|
| 986 |
+
"epoch": 12.280701754385966,
|
| 987 |
+
"grad_norm": 0.18020951747894287,
|
| 988 |
+
"learning_rate": 1.8458646616541354e-06,
|
| 989 |
+
"loss": 0.0148,
|
| 990 |
+
"step": 3500
|
| 991 |
+
},
|
| 992 |
+
{
|
| 993 |
+
"epoch": 12.368421052631579,
|
| 994 |
+
"grad_norm": 0.2494085282087326,
|
| 995 |
+
"learning_rate": 6.569377990430622e-06,
|
| 996 |
+
"loss": 0.1168,
|
| 997 |
+
"step": 3525
|
| 998 |
+
},
|
| 999 |
+
{
|
| 1000 |
+
"epoch": 12.456140350877194,
|
| 1001 |
+
"grad_norm": 0.16571170091629028,
|
| 1002 |
+
"learning_rate": 6.509569377990431e-06,
|
| 1003 |
+
"loss": 0.0482,
|
| 1004 |
+
"step": 3550
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 12.543859649122806,
|
| 1008 |
+
"grad_norm": 0.10973577946424484,
|
| 1009 |
+
"learning_rate": 6.44976076555024e-06,
|
| 1010 |
+
"loss": 0.1372,
|
| 1011 |
+
"step": 3575
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 12.631578947368421,
|
| 1015 |
+
"grad_norm": 0.14920943975448608,
|
| 1016 |
+
"learning_rate": 6.389952153110048e-06,
|
| 1017 |
+
"loss": 0.0791,
|
| 1018 |
+
"step": 3600
|
| 1019 |
+
},
|
| 1020 |
+
{
|
| 1021 |
+
"epoch": 12.719298245614034,
|
| 1022 |
+
"grad_norm": 4.5087504386901855,
|
| 1023 |
+
"learning_rate": 6.330143540669857e-06,
|
| 1024 |
+
"loss": 0.1614,
|
| 1025 |
+
"step": 3625
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 12.807017543859649,
|
| 1029 |
+
"grad_norm": 0.11903402209281921,
|
| 1030 |
+
"learning_rate": 6.270334928229666e-06,
|
| 1031 |
+
"loss": 0.0082,
|
| 1032 |
+
"step": 3650
|
| 1033 |
+
},
|
| 1034 |
+
{
|
| 1035 |
+
"epoch": 12.894736842105264,
|
| 1036 |
+
"grad_norm": 0.24349439144134521,
|
| 1037 |
+
"learning_rate": 6.210526315789474e-06,
|
| 1038 |
+
"loss": 0.1936,
|
| 1039 |
+
"step": 3675
|
| 1040 |
+
},
|
| 1041 |
+
{
|
| 1042 |
+
"epoch": 12.982456140350877,
|
| 1043 |
+
"grad_norm": 146.2913360595703,
|
| 1044 |
+
"learning_rate": 6.150717703349282e-06,
|
| 1045 |
+
"loss": 0.1465,
|
| 1046 |
+
"step": 3700
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 13.070175438596491,
|
| 1050 |
+
"grad_norm": 88.66224670410156,
|
| 1051 |
+
"learning_rate": 6.090909090909091e-06,
|
| 1052 |
+
"loss": 0.0751,
|
| 1053 |
+
"step": 3725
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 13.157894736842104,
|
| 1057 |
+
"grad_norm": 0.11558452993631363,
|
| 1058 |
+
"learning_rate": 6.0311004784689e-06,
|
| 1059 |
+
"loss": 0.0652,
|
| 1060 |
+
"step": 3750
|
| 1061 |
+
},
|
| 1062 |
+
{
|
| 1063 |
+
"epoch": 13.24561403508772,
|
| 1064 |
+
"grad_norm": 3.1845107078552246,
|
| 1065 |
+
"learning_rate": 5.971291866028708e-06,
|
| 1066 |
+
"loss": 0.0969,
|
| 1067 |
+
"step": 3775
|
| 1068 |
+
},
|
| 1069 |
+
{
|
| 1070 |
+
"epoch": 13.333333333333334,
|
| 1071 |
+
"grad_norm": 0.5500511527061462,
|
| 1072 |
+
"learning_rate": 5.911483253588517e-06,
|
| 1073 |
+
"loss": 0.107,
|
| 1074 |
+
"step": 3800
|
| 1075 |
+
},
|
| 1076 |
+
{
|
| 1077 |
+
"epoch": 13.421052631578947,
|
| 1078 |
+
"grad_norm": 0.08417954295873642,
|
| 1079 |
+
"learning_rate": 5.851674641148326e-06,
|
| 1080 |
+
"loss": 0.0165,
|
| 1081 |
+
"step": 3825
|
| 1082 |
+
},
|
| 1083 |
+
{
|
| 1084 |
+
"epoch": 13.508771929824562,
|
| 1085 |
+
"grad_norm": 0.0910855233669281,
|
| 1086 |
+
"learning_rate": 5.791866028708134e-06,
|
| 1087 |
+
"loss": 0.1326,
|
| 1088 |
+
"step": 3850
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 13.596491228070175,
|
| 1092 |
+
"grad_norm": 65.31088256835938,
|
| 1093 |
+
"learning_rate": 5.732057416267942e-06,
|
| 1094 |
+
"loss": 0.0365,
|
| 1095 |
+
"step": 3875
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 13.68421052631579,
|
| 1099 |
+
"grad_norm": 0.09439278393983841,
|
| 1100 |
+
"learning_rate": 5.672248803827751e-06,
|
| 1101 |
+
"loss": 0.0784,
|
| 1102 |
+
"step": 3900
|
| 1103 |
+
},
|
| 1104 |
+
{
|
| 1105 |
+
"epoch": 13.771929824561404,
|
| 1106 |
+
"grad_norm": 0.17843316495418549,
|
| 1107 |
+
"learning_rate": 5.6124401913875595e-06,
|
| 1108 |
+
"loss": 0.0232,
|
| 1109 |
+
"step": 3925
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 13.859649122807017,
|
| 1113 |
+
"grad_norm": 0.16249841451644897,
|
| 1114 |
+
"learning_rate": 5.552631578947368e-06,
|
| 1115 |
+
"loss": 0.0107,
|
| 1116 |
+
"step": 3950
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 13.947368421052632,
|
| 1120 |
+
"grad_norm": 0.545814573764801,
|
| 1121 |
+
"learning_rate": 5.492822966507177e-06,
|
| 1122 |
+
"loss": 0.0947,
|
| 1123 |
+
"step": 3975
|
| 1124 |
+
},
|
| 1125 |
+
{
|
| 1126 |
+
"epoch": 14.035087719298245,
|
| 1127 |
+
"grad_norm": 0.7138233184814453,
|
| 1128 |
+
"learning_rate": 5.433014354066986e-06,
|
| 1129 |
+
"loss": 0.1812,
|
| 1130 |
+
"step": 4000
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 14.12280701754386,
|
| 1134 |
+
"grad_norm": 381.4528503417969,
|
| 1135 |
+
"learning_rate": 5.373205741626794e-06,
|
| 1136 |
+
"loss": 0.0399,
|
| 1137 |
+
"step": 4025
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 14.210526315789474,
|
| 1141 |
+
"grad_norm": 0.053043171763420105,
|
| 1142 |
+
"learning_rate": 5.313397129186603e-06,
|
| 1143 |
+
"loss": 0.1249,
|
| 1144 |
+
"step": 4050
|
| 1145 |
+
},
|
| 1146 |
+
{
|
| 1147 |
+
"epoch": 14.298245614035087,
|
| 1148 |
+
"grad_norm": 1.4563120603561401,
|
| 1149 |
+
"learning_rate": 5.253588516746412e-06,
|
| 1150 |
+
"loss": 0.0361,
|
| 1151 |
+
"step": 4075
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 14.385964912280702,
|
| 1155 |
+
"grad_norm": 0.06027642637491226,
|
| 1156 |
+
"learning_rate": 5.19377990430622e-06,
|
| 1157 |
+
"loss": 0.0799,
|
| 1158 |
+
"step": 4100
|
| 1159 |
+
},
|
| 1160 |
+
{
|
| 1161 |
+
"epoch": 14.473684210526315,
|
| 1162 |
+
"grad_norm": 0.11955183744430542,
|
| 1163 |
+
"learning_rate": 5.133971291866029e-06,
|
| 1164 |
+
"loss": 0.2006,
|
| 1165 |
+
"step": 4125
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 14.56140350877193,
|
| 1169 |
+
"grad_norm": 0.07738160341978073,
|
| 1170 |
+
"learning_rate": 5.074162679425838e-06,
|
| 1171 |
+
"loss": 0.16,
|
| 1172 |
+
"step": 4150
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 14.649122807017545,
|
| 1176 |
+
"grad_norm": 28.58780860900879,
|
| 1177 |
+
"learning_rate": 5.014354066985646e-06,
|
| 1178 |
+
"loss": 0.0675,
|
| 1179 |
+
"step": 4175
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 14.736842105263158,
|
| 1183 |
+
"grad_norm": 0.08654920011758804,
|
| 1184 |
+
"learning_rate": 4.954545454545455e-06,
|
| 1185 |
+
"loss": 0.0584,
|
| 1186 |
+
"step": 4200
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 14.824561403508772,
|
| 1190 |
+
"grad_norm": 0.1421671062707901,
|
| 1191 |
+
"learning_rate": 4.894736842105264e-06,
|
| 1192 |
+
"loss": 0.0049,
|
| 1193 |
+
"step": 4225
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 14.912280701754385,
|
| 1197 |
+
"grad_norm": 0.23447291553020477,
|
| 1198 |
+
"learning_rate": 4.834928229665072e-06,
|
| 1199 |
+
"loss": 0.0122,
|
| 1200 |
+
"step": 4250
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 15.0,
|
| 1204 |
+
"grad_norm": 54.224788665771484,
|
| 1205 |
+
"learning_rate": 4.775119617224881e-06,
|
| 1206 |
+
"loss": 0.0921,
|
| 1207 |
+
"step": 4275
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 15.087719298245615,
|
| 1211 |
+
"grad_norm": 0.08427131175994873,
|
| 1212 |
+
"learning_rate": 4.715311004784689e-06,
|
| 1213 |
+
"loss": 0.0061,
|
| 1214 |
+
"step": 4300
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 15.175438596491228,
|
| 1218 |
+
"grad_norm": 0.232700914144516,
|
| 1219 |
+
"learning_rate": 4.6555023923444975e-06,
|
| 1220 |
+
"loss": 0.1011,
|
| 1221 |
+
"step": 4325
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 15.263157894736842,
|
| 1225 |
+
"grad_norm": 0.09075941890478134,
|
| 1226 |
+
"learning_rate": 4.595693779904306e-06,
|
| 1227 |
+
"loss": 0.0419,
|
| 1228 |
+
"step": 4350
|
| 1229 |
+
},
|
| 1230 |
+
{
|
| 1231 |
+
"epoch": 15.350877192982455,
|
| 1232 |
+
"grad_norm": 0.20714852213859558,
|
| 1233 |
+
"learning_rate": 4.535885167464115e-06,
|
| 1234 |
+
"loss": 0.1786,
|
| 1235 |
+
"step": 4375
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 15.43859649122807,
|
| 1239 |
+
"grad_norm": 0.07586350291967392,
|
| 1240 |
+
"learning_rate": 4.476076555023924e-06,
|
| 1241 |
+
"loss": 0.0314,
|
| 1242 |
+
"step": 4400
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 15.526315789473685,
|
| 1246 |
+
"grad_norm": 0.29287371039390564,
|
| 1247 |
+
"learning_rate": 4.416267942583732e-06,
|
| 1248 |
+
"loss": 0.0635,
|
| 1249 |
+
"step": 4425
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 15.614035087719298,
|
| 1253 |
+
"grad_norm": 50.64339065551758,
|
| 1254 |
+
"learning_rate": 4.356459330143541e-06,
|
| 1255 |
+
"loss": 0.0854,
|
| 1256 |
+
"step": 4450
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 15.701754385964913,
|
| 1260 |
+
"grad_norm": 0.12661752104759216,
|
| 1261 |
+
"learning_rate": 4.29665071770335e-06,
|
| 1262 |
+
"loss": 0.1378,
|
| 1263 |
+
"step": 4475
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 15.789473684210526,
|
| 1267 |
+
"grad_norm": 2.418790340423584,
|
| 1268 |
+
"learning_rate": 4.2368421052631575e-06,
|
| 1269 |
+
"loss": 0.0325,
|
| 1270 |
+
"step": 4500
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 15.87719298245614,
|
| 1274 |
+
"grad_norm": 231.30796813964844,
|
| 1275 |
+
"learning_rate": 4.177033492822966e-06,
|
| 1276 |
+
"loss": 0.1398,
|
| 1277 |
+
"step": 4525
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 15.964912280701755,
|
| 1281 |
+
"grad_norm": 0.23073101043701172,
|
| 1282 |
+
"learning_rate": 4.117224880382775e-06,
|
| 1283 |
+
"loss": 0.0237,
|
| 1284 |
+
"step": 4550
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 16.05263157894737,
|
| 1288 |
+
"grad_norm": 0.044060900807380676,
|
| 1289 |
+
"learning_rate": 4.0574162679425835e-06,
|
| 1290 |
+
"loss": 0.0074,
|
| 1291 |
+
"step": 4575
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 16.140350877192983,
|
| 1295 |
+
"grad_norm": 1.4264813661575317,
|
| 1296 |
+
"learning_rate": 3.997607655502392e-06,
|
| 1297 |
+
"loss": 0.025,
|
| 1298 |
+
"step": 4600
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 16.228070175438596,
|
| 1302 |
+
"grad_norm": 0.41092243790626526,
|
| 1303 |
+
"learning_rate": 3.937799043062201e-06,
|
| 1304 |
+
"loss": 0.0049,
|
| 1305 |
+
"step": 4625
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 16.31578947368421,
|
| 1309 |
+
"grad_norm": 3.6420750617980957,
|
| 1310 |
+
"learning_rate": 3.8779904306220095e-06,
|
| 1311 |
+
"loss": 0.0491,
|
| 1312 |
+
"step": 4650
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"epoch": 16.403508771929825,
|
| 1316 |
+
"grad_norm": 0.08417033404111862,
|
| 1317 |
+
"learning_rate": 3.818181818181818e-06,
|
| 1318 |
+
"loss": 0.003,
|
| 1319 |
+
"step": 4675
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 16.49122807017544,
|
| 1323 |
+
"grad_norm": 0.41752633452415466,
|
| 1324 |
+
"learning_rate": 3.7583732057416273e-06,
|
| 1325 |
+
"loss": 0.0649,
|
| 1326 |
+
"step": 4700
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 16.57894736842105,
|
| 1330 |
+
"grad_norm": 0.04892294108867645,
|
| 1331 |
+
"learning_rate": 3.6985645933014356e-06,
|
| 1332 |
+
"loss": 0.1577,
|
| 1333 |
+
"step": 4725
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 16.666666666666668,
|
| 1337 |
+
"grad_norm": 0.07886885106563568,
|
| 1338 |
+
"learning_rate": 3.6387559808612442e-06,
|
| 1339 |
+
"loss": 0.1133,
|
| 1340 |
+
"step": 4750
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 16.75438596491228,
|
| 1344 |
+
"grad_norm": 202.08302307128906,
|
| 1345 |
+
"learning_rate": 3.5789473684210525e-06,
|
| 1346 |
+
"loss": 0.0151,
|
| 1347 |
+
"step": 4775
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 16.842105263157894,
|
| 1351 |
+
"grad_norm": 77.71251678466797,
|
| 1352 |
+
"learning_rate": 3.519138755980861e-06,
|
| 1353 |
+
"loss": 0.0644,
|
| 1354 |
+
"step": 4800
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 16.92982456140351,
|
| 1358 |
+
"grad_norm": 0.04525836929678917,
|
| 1359 |
+
"learning_rate": 3.45933014354067e-06,
|
| 1360 |
+
"loss": 0.0283,
|
| 1361 |
+
"step": 4825
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 17.017543859649123,
|
| 1365 |
+
"grad_norm": 0.035384029150009155,
|
| 1366 |
+
"learning_rate": 3.399521531100479e-06,
|
| 1367 |
+
"loss": 0.0431,
|
| 1368 |
+
"step": 4850
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 17.105263157894736,
|
| 1372 |
+
"grad_norm": 0.05048861354589462,
|
| 1373 |
+
"learning_rate": 3.339712918660287e-06,
|
| 1374 |
+
"loss": 0.0358,
|
| 1375 |
+
"step": 4875
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 17.19298245614035,
|
| 1379 |
+
"grad_norm": 0.1347443163394928,
|
| 1380 |
+
"learning_rate": 3.279904306220096e-06,
|
| 1381 |
+
"loss": 0.0206,
|
| 1382 |
+
"step": 4900
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 17.280701754385966,
|
| 1386 |
+
"grad_norm": 252.46458435058594,
|
| 1387 |
+
"learning_rate": 3.2200956937799046e-06,
|
| 1388 |
+
"loss": 0.0804,
|
| 1389 |
+
"step": 4925
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 17.36842105263158,
|
| 1393 |
+
"grad_norm": 0.05461292341351509,
|
| 1394 |
+
"learning_rate": 3.1602870813397132e-06,
|
| 1395 |
+
"loss": 0.0784,
|
| 1396 |
+
"step": 4950
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 17.45614035087719,
|
| 1400 |
+
"grad_norm": 2.0001866817474365,
|
| 1401 |
+
"learning_rate": 3.1004784688995215e-06,
|
| 1402 |
+
"loss": 0.0947,
|
| 1403 |
+
"step": 4975
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 17.54385964912281,
|
| 1407 |
+
"grad_norm": 0.04590539261698723,
|
| 1408 |
+
"learning_rate": 3.04066985645933e-06,
|
| 1409 |
+
"loss": 0.075,
|
| 1410 |
+
"step": 5000
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 17.63157894736842,
|
| 1414 |
+
"grad_norm": 0.04810748249292374,
|
| 1415 |
+
"learning_rate": 2.980861244019139e-06,
|
| 1416 |
+
"loss": 0.0342,
|
| 1417 |
+
"step": 5025
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 17.719298245614034,
|
| 1421 |
+
"grad_norm": 0.8750467896461487,
|
| 1422 |
+
"learning_rate": 2.9210526315789475e-06,
|
| 1423 |
+
"loss": 0.0618,
|
| 1424 |
+
"step": 5050
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 17.80701754385965,
|
| 1428 |
+
"grad_norm": 56.14554977416992,
|
| 1429 |
+
"learning_rate": 2.8612440191387558e-06,
|
| 1430 |
+
"loss": 0.026,
|
| 1431 |
+
"step": 5075
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 17.894736842105264,
|
| 1435 |
+
"grad_norm": 0.8380083441734314,
|
| 1436 |
+
"learning_rate": 2.8014354066985645e-06,
|
| 1437 |
+
"loss": 0.0135,
|
| 1438 |
+
"step": 5100
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 17.982456140350877,
|
| 1442 |
+
"grad_norm": 0.08012706786394119,
|
| 1443 |
+
"learning_rate": 2.741626794258373e-06,
|
| 1444 |
+
"loss": 0.0586,
|
| 1445 |
+
"step": 5125
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 18.07017543859649,
|
| 1449 |
+
"grad_norm": 1.2549360990524292,
|
| 1450 |
+
"learning_rate": 2.6818181818181822e-06,
|
| 1451 |
+
"loss": 0.0047,
|
| 1452 |
+
"step": 5150
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"epoch": 18.157894736842106,
|
| 1456 |
+
"grad_norm": 0.03993632644414902,
|
| 1457 |
+
"learning_rate": 2.6220095693779905e-06,
|
| 1458 |
+
"loss": 0.0869,
|
| 1459 |
+
"step": 5175
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 18.24561403508772,
|
| 1463 |
+
"grad_norm": 62.21980285644531,
|
| 1464 |
+
"learning_rate": 2.562200956937799e-06,
|
| 1465 |
+
"loss": 0.0987,
|
| 1466 |
+
"step": 5200
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 18.333333333333332,
|
| 1470 |
+
"grad_norm": 0.10092537105083466,
|
| 1471 |
+
"learning_rate": 2.502392344497608e-06,
|
| 1472 |
+
"loss": 0.1554,
|
| 1473 |
+
"step": 5225
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 18.42105263157895,
|
| 1477 |
+
"grad_norm": 0.7630422115325928,
|
| 1478 |
+
"learning_rate": 2.4425837320574165e-06,
|
| 1479 |
+
"loss": 0.0522,
|
| 1480 |
+
"step": 5250
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 18.50877192982456,
|
| 1484 |
+
"grad_norm": 62.42081069946289,
|
| 1485 |
+
"learning_rate": 2.382775119617225e-06,
|
| 1486 |
+
"loss": 0.0298,
|
| 1487 |
+
"step": 5275
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 18.596491228070175,
|
| 1491 |
+
"grad_norm": 0.046070296317338943,
|
| 1492 |
+
"learning_rate": 2.3229665071770335e-06,
|
| 1493 |
+
"loss": 0.1207,
|
| 1494 |
+
"step": 5300
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 18.68421052631579,
|
| 1498 |
+
"grad_norm": 0.04533836990594864,
|
| 1499 |
+
"learning_rate": 2.263157894736842e-06,
|
| 1500 |
+
"loss": 0.146,
|
| 1501 |
+
"step": 5325
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 18.771929824561404,
|
| 1505 |
+
"grad_norm": 0.03193026781082153,
|
| 1506 |
+
"learning_rate": 2.203349282296651e-06,
|
| 1507 |
+
"loss": 0.0106,
|
| 1508 |
+
"step": 5350
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 18.859649122807017,
|
| 1512 |
+
"grad_norm": 0.03594537451863289,
|
| 1513 |
+
"learning_rate": 2.143540669856459e-06,
|
| 1514 |
+
"loss": 0.0266,
|
| 1515 |
+
"step": 5375
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 18.94736842105263,
|
| 1519 |
+
"grad_norm": 0.07382319122552872,
|
| 1520 |
+
"learning_rate": 2.0837320574162678e-06,
|
| 1521 |
+
"loss": 0.0954,
|
| 1522 |
+
"step": 5400
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 19.035087719298247,
|
| 1526 |
+
"grad_norm": 0.0792032852768898,
|
| 1527 |
+
"learning_rate": 2.0239234449760764e-06,
|
| 1528 |
+
"loss": 0.0871,
|
| 1529 |
+
"step": 5425
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 19.12280701754386,
|
| 1533 |
+
"grad_norm": 0.06994396448135376,
|
| 1534 |
+
"learning_rate": 1.9641148325358855e-06,
|
| 1535 |
+
"loss": 0.0092,
|
| 1536 |
+
"step": 5450
|
| 1537 |
+
},
|
| 1538 |
+
{
|
| 1539 |
+
"epoch": 19.210526315789473,
|
| 1540 |
+
"grad_norm": 36.98837661743164,
|
| 1541 |
+
"learning_rate": 1.904306220095694e-06,
|
| 1542 |
+
"loss": 0.1442,
|
| 1543 |
+
"step": 5475
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"epoch": 19.29824561403509,
|
| 1547 |
+
"grad_norm": 0.11950813978910446,
|
| 1548 |
+
"learning_rate": 1.8444976076555023e-06,
|
| 1549 |
+
"loss": 0.0344,
|
| 1550 |
+
"step": 5500
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 19.385964912280702,
|
| 1554 |
+
"grad_norm": 0.053549814969301224,
|
| 1555 |
+
"learning_rate": 1.7846889952153112e-06,
|
| 1556 |
+
"loss": 0.0609,
|
| 1557 |
+
"step": 5525
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 19.473684210526315,
|
| 1561 |
+
"grad_norm": 0.05915123224258423,
|
| 1562 |
+
"learning_rate": 1.7248803827751196e-06,
|
| 1563 |
+
"loss": 0.0212,
|
| 1564 |
+
"step": 5550
|
| 1565 |
+
},
|
| 1566 |
+
{
|
| 1567 |
+
"epoch": 19.56140350877193,
|
| 1568 |
+
"grad_norm": 0.03509140759706497,
|
| 1569 |
+
"learning_rate": 1.6650717703349283e-06,
|
| 1570 |
+
"loss": 0.0019,
|
| 1571 |
+
"step": 5575
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 19.649122807017545,
|
| 1575 |
+
"grad_norm": 0.04774395003914833,
|
| 1576 |
+
"learning_rate": 1.605263157894737e-06,
|
| 1577 |
+
"loss": 0.032,
|
| 1578 |
+
"step": 5600
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 19.736842105263158,
|
| 1582 |
+
"grad_norm": 0.1909104436635971,
|
| 1583 |
+
"learning_rate": 1.5454545454545454e-06,
|
| 1584 |
+
"loss": 0.0351,
|
| 1585 |
+
"step": 5625
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"epoch": 19.82456140350877,
|
| 1589 |
+
"grad_norm": 0.11199568212032318,
|
| 1590 |
+
"learning_rate": 1.4856459330143541e-06,
|
| 1591 |
+
"loss": 0.0268,
|
| 1592 |
+
"step": 5650
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 19.912280701754387,
|
| 1596 |
+
"grad_norm": 17.852022171020508,
|
| 1597 |
+
"learning_rate": 1.4258373205741628e-06,
|
| 1598 |
+
"loss": 0.0397,
|
| 1599 |
+
"step": 5675
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 20.0,
|
| 1603 |
+
"grad_norm": 0.18822583556175232,
|
| 1604 |
+
"learning_rate": 1.3660287081339715e-06,
|
| 1605 |
+
"loss": 0.0541,
|
| 1606 |
+
"step": 5700
|
| 1607 |
+
},
|
| 1608 |
+
{
|
| 1609 |
+
"epoch": 20.087719298245613,
|
| 1610 |
+
"grad_norm": 50.338890075683594,
|
| 1611 |
+
"learning_rate": 1.30622009569378e-06,
|
| 1612 |
+
"loss": 0.016,
|
| 1613 |
+
"step": 5725
|
| 1614 |
+
},
|
| 1615 |
+
{
|
| 1616 |
+
"epoch": 20.17543859649123,
|
| 1617 |
+
"grad_norm": 0.05340347811579704,
|
| 1618 |
+
"learning_rate": 1.2464114832535886e-06,
|
| 1619 |
+
"loss": 0.0084,
|
| 1620 |
+
"step": 5750
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 20.263157894736842,
|
| 1624 |
+
"grad_norm": 0.04629586637020111,
|
| 1625 |
+
"learning_rate": 1.186602870813397e-06,
|
| 1626 |
+
"loss": 0.0356,
|
| 1627 |
+
"step": 5775
|
| 1628 |
+
},
|
| 1629 |
+
{
|
| 1630 |
+
"epoch": 20.350877192982455,
|
| 1631 |
+
"grad_norm": 0.9073031544685364,
|
| 1632 |
+
"learning_rate": 1.1267942583732058e-06,
|
| 1633 |
+
"loss": 0.0786,
|
| 1634 |
+
"step": 5800
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 20.43859649122807,
|
| 1638 |
+
"grad_norm": 10.095131874084473,
|
| 1639 |
+
"learning_rate": 1.0669856459330144e-06,
|
| 1640 |
+
"loss": 0.0419,
|
| 1641 |
+
"step": 5825
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 20.526315789473685,
|
| 1645 |
+
"grad_norm": 0.16011740267276764,
|
| 1646 |
+
"learning_rate": 1.0071770334928231e-06,
|
| 1647 |
+
"loss": 0.0653,
|
| 1648 |
+
"step": 5850
|
| 1649 |
+
},
|
| 1650 |
+
{
|
| 1651 |
+
"epoch": 20.614035087719298,
|
| 1652 |
+
"grad_norm": 0.18243364989757538,
|
| 1653 |
+
"learning_rate": 9.473684210526316e-07,
|
| 1654 |
+
"loss": 0.0015,
|
| 1655 |
+
"step": 5875
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 20.70175438596491,
|
| 1659 |
+
"grad_norm": 0.14108963310718536,
|
| 1660 |
+
"learning_rate": 8.875598086124402e-07,
|
| 1661 |
+
"loss": 0.0817,
|
| 1662 |
+
"step": 5900
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 20.789473684210527,
|
| 1666 |
+
"grad_norm": 0.03629032149910927,
|
| 1667 |
+
"learning_rate": 8.277511961722487e-07,
|
| 1668 |
+
"loss": 0.002,
|
| 1669 |
+
"step": 5925
|
| 1670 |
+
},
|
| 1671 |
+
{
|
| 1672 |
+
"epoch": 20.87719298245614,
|
| 1673 |
+
"grad_norm": 21.195932388305664,
|
| 1674 |
+
"learning_rate": 7.679425837320574e-07,
|
| 1675 |
+
"loss": 0.1265,
|
| 1676 |
+
"step": 5950
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 20.964912280701753,
|
| 1680 |
+
"grad_norm": 0.036704275757074356,
|
| 1681 |
+
"learning_rate": 7.081339712918661e-07,
|
| 1682 |
+
"loss": 0.0369,
|
| 1683 |
+
"step": 5975
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 21.05263157894737,
|
| 1687 |
+
"grad_norm": 327.310791015625,
|
| 1688 |
+
"learning_rate": 6.483253588516747e-07,
|
| 1689 |
+
"loss": 0.0288,
|
| 1690 |
+
"step": 6000
|
| 1691 |
+
}
|
| 1692 |
+
],
|
| 1693 |
+
"logging_steps": 25,
|
| 1694 |
+
"max_steps": 6270,
|
| 1695 |
+
"num_input_tokens_seen": 0,
|
| 1696 |
+
"num_train_epochs": 22,
|
| 1697 |
+
"save_steps": 500,
|
| 1698 |
+
"stateful_callbacks": {
|
| 1699 |
+
"TrainerControl": {
|
| 1700 |
+
"args": {
|
| 1701 |
+
"should_epoch_stop": false,
|
| 1702 |
+
"should_evaluate": false,
|
| 1703 |
+
"should_log": false,
|
| 1704 |
+
"should_save": true,
|
| 1705 |
+
"should_training_stop": false
|
| 1706 |
+
},
|
| 1707 |
+
"attributes": {}
|
| 1708 |
+
}
|
| 1709 |
+
},
|
| 1710 |
+
"total_flos": 1.6186141004321981e+18,
|
| 1711 |
+
"train_batch_size": 4,
|
| 1712 |
+
"trial_name": null,
|
| 1713 |
+
"trial_params": null
|
| 1714 |
+
}
|
checkpoint-6000/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:21cda56f26874205e02c84553dd1454d4b3eefd72224b421d8d696952554b651
|
| 3 |
+
size 5841
|
checkpoint-6270/config.json
ADDED
|
@@ -0,0 +1,139 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"activation_dropout": 0.0,
|
| 3 |
+
"adapter_attn_dim": null,
|
| 4 |
+
"adapter_kernel_size": 3,
|
| 5 |
+
"adapter_stride": 2,
|
| 6 |
+
"add_adapter": false,
|
| 7 |
+
"apply_spec_augment": true,
|
| 8 |
+
"architectures": [
|
| 9 |
+
"Wav2Vec2ForSequenceClassification"
|
| 10 |
+
],
|
| 11 |
+
"attention_dropout": 0.1,
|
| 12 |
+
"bos_token_id": 1,
|
| 13 |
+
"classifier_proj_size": 256,
|
| 14 |
+
"codevector_dim": 256,
|
| 15 |
+
"contrastive_logits_temperature": 0.1,
|
| 16 |
+
"conv_bias": false,
|
| 17 |
+
"conv_dim": [
|
| 18 |
+
512,
|
| 19 |
+
512,
|
| 20 |
+
512,
|
| 21 |
+
512,
|
| 22 |
+
512,
|
| 23 |
+
512,
|
| 24 |
+
512
|
| 25 |
+
],
|
| 26 |
+
"conv_kernel": [
|
| 27 |
+
10,
|
| 28 |
+
3,
|
| 29 |
+
3,
|
| 30 |
+
3,
|
| 31 |
+
3,
|
| 32 |
+
2,
|
| 33 |
+
2
|
| 34 |
+
],
|
| 35 |
+
"conv_stride": [
|
| 36 |
+
5,
|
| 37 |
+
2,
|
| 38 |
+
2,
|
| 39 |
+
2,
|
| 40 |
+
2,
|
| 41 |
+
2,
|
| 42 |
+
2
|
| 43 |
+
],
|
| 44 |
+
"ctc_loss_reduction": "sum",
|
| 45 |
+
"ctc_zero_infinity": false,
|
| 46 |
+
"diversity_loss_weight": 0.1,
|
| 47 |
+
"do_stable_layer_norm": false,
|
| 48 |
+
"dtype": "float32",
|
| 49 |
+
"eos_token_id": 2,
|
| 50 |
+
"feat_extract_activation": "gelu",
|
| 51 |
+
"feat_extract_norm": "group",
|
| 52 |
+
"feat_proj_dropout": 0.1,
|
| 53 |
+
"feat_quantizer_dropout": 0.0,
|
| 54 |
+
"final_dropout": 0.0,
|
| 55 |
+
"finetuning_task": "wav2vec2_emotion",
|
| 56 |
+
"freeze_feat_extract_train": true,
|
| 57 |
+
"hidden_act": "gelu",
|
| 58 |
+
"hidden_dropout": 0.1,
|
| 59 |
+
"hidden_size": 768,
|
| 60 |
+
"id2label": {
|
| 61 |
+
"0": "neutral",
|
| 62 |
+
"1": "calm",
|
| 63 |
+
"2": "happy",
|
| 64 |
+
"3": "sad",
|
| 65 |
+
"4": "angry",
|
| 66 |
+
"5": "fearful",
|
| 67 |
+
"6": "disgust",
|
| 68 |
+
"7": "surprised"
|
| 69 |
+
},
|
| 70 |
+
"initializer_range": 0.02,
|
| 71 |
+
"intermediate_size": 3072,
|
| 72 |
+
"label2id": {
|
| 73 |
+
"angry": 4,
|
| 74 |
+
"calm": 1,
|
| 75 |
+
"disgust": 6,
|
| 76 |
+
"fearful": 5,
|
| 77 |
+
"happy": 2,
|
| 78 |
+
"neutral": 0,
|
| 79 |
+
"sad": 3,
|
| 80 |
+
"surprised": 7
|
| 81 |
+
},
|
| 82 |
+
"layer_norm_eps": 1e-05,
|
| 83 |
+
"layerdrop": 0.05,
|
| 84 |
+
"mask_channel_length": 10,
|
| 85 |
+
"mask_channel_min_space": 1,
|
| 86 |
+
"mask_channel_other": 0.0,
|
| 87 |
+
"mask_channel_prob": 0.0,
|
| 88 |
+
"mask_channel_selection": "static",
|
| 89 |
+
"mask_feature_length": 10,
|
| 90 |
+
"mask_feature_min_masks": 0,
|
| 91 |
+
"mask_feature_prob": 0.0,
|
| 92 |
+
"mask_time_length": 10,
|
| 93 |
+
"mask_time_min_masks": 2,
|
| 94 |
+
"mask_time_min_space": 1,
|
| 95 |
+
"mask_time_other": 0.0,
|
| 96 |
+
"mask_time_prob": 0.05,
|
| 97 |
+
"mask_time_selection": "static",
|
| 98 |
+
"model_type": "wav2vec2",
|
| 99 |
+
"no_mask_channel_overlap": false,
|
| 100 |
+
"no_mask_time_overlap": false,
|
| 101 |
+
"num_adapter_layers": 3,
|
| 102 |
+
"num_attention_heads": 12,
|
| 103 |
+
"num_codevector_groups": 2,
|
| 104 |
+
"num_codevectors_per_group": 320,
|
| 105 |
+
"num_conv_pos_embedding_groups": 16,
|
| 106 |
+
"num_conv_pos_embeddings": 128,
|
| 107 |
+
"num_feat_extract_layers": 7,
|
| 108 |
+
"num_hidden_layers": 12,
|
| 109 |
+
"num_negatives": 100,
|
| 110 |
+
"output_hidden_size": 768,
|
| 111 |
+
"pad_token_id": 0,
|
| 112 |
+
"proj_codevector_dim": 256,
|
| 113 |
+
"tdnn_dilation": [
|
| 114 |
+
1,
|
| 115 |
+
2,
|
| 116 |
+
3,
|
| 117 |
+
1,
|
| 118 |
+
1
|
| 119 |
+
],
|
| 120 |
+
"tdnn_dim": [
|
| 121 |
+
512,
|
| 122 |
+
512,
|
| 123 |
+
512,
|
| 124 |
+
512,
|
| 125 |
+
1500
|
| 126 |
+
],
|
| 127 |
+
"tdnn_kernel": [
|
| 128 |
+
5,
|
| 129 |
+
3,
|
| 130 |
+
3,
|
| 131 |
+
1,
|
| 132 |
+
1
|
| 133 |
+
],
|
| 134 |
+
"transformers_version": "4.57.1",
|
| 135 |
+
"use_cache": false,
|
| 136 |
+
"use_weighted_layer_sum": true,
|
| 137 |
+
"vocab_size": 32,
|
| 138 |
+
"xvector_output_dim": 512
|
| 139 |
+
}
|
checkpoint-6270/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d01726ed5d89e30cdf84b5dbff1273c611514bc8962ba0dfa1601980938c63a6
|
| 3 |
+
size 378308660
|
checkpoint-6270/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1fccffb0c49119d9ccafaf6c505e414b75c0deb8bcbfa344d3db218479db4817
|
| 3 |
+
size 756740235
|
checkpoint-6270/preprocessor_config.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_normalize": false,
|
| 3 |
+
"feature_extractor_type": "Wav2Vec2FeatureExtractor",
|
| 4 |
+
"feature_size": 1,
|
| 5 |
+
"padding_side": "right",
|
| 6 |
+
"padding_value": 0.0,
|
| 7 |
+
"return_attention_mask": true,
|
| 8 |
+
"sampling_rate": 16000
|
| 9 |
+
}
|
checkpoint-6270/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:12c1d73d2b023c9b8c6bcb57163168ee23e9cf5b28d17af3d6ee6e2709094aed
|
| 3 |
+
size 14455
|
checkpoint-6270/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d8ffea8fc20da6bae69a62553fa99ac517931836e4e15c9007f0ae2118d30bb0
|
| 3 |
+
size 1465
|
checkpoint-6270/trainer_state.json
ADDED
|
@@ -0,0 +1,1784 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 22.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 6270,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.08771929824561403,
|
| 14 |
+
"grad_norm": 14.743780136108398,
|
| 15 |
+
"learning_rate": 1.4942583732057417e-05,
|
| 16 |
+
"loss": 0.9366,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.17543859649122806,
|
| 21 |
+
"grad_norm": 22.20017433166504,
|
| 22 |
+
"learning_rate": 1.4882775119617225e-05,
|
| 23 |
+
"loss": 1.2015,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.2631578947368421,
|
| 28 |
+
"grad_norm": 25.090787887573242,
|
| 29 |
+
"learning_rate": 1.4822966507177032e-05,
|
| 30 |
+
"loss": 1.0156,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.3508771929824561,
|
| 35 |
+
"grad_norm": 13.580609321594238,
|
| 36 |
+
"learning_rate": 1.4763157894736842e-05,
|
| 37 |
+
"loss": 1.117,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.43859649122807015,
|
| 42 |
+
"grad_norm": 12.062329292297363,
|
| 43 |
+
"learning_rate": 1.470334928229665e-05,
|
| 44 |
+
"loss": 0.8312,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.5263157894736842,
|
| 49 |
+
"grad_norm": 18.625316619873047,
|
| 50 |
+
"learning_rate": 1.464354066985646e-05,
|
| 51 |
+
"loss": 0.9938,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.6140350877192983,
|
| 56 |
+
"grad_norm": 11.628118515014648,
|
| 57 |
+
"learning_rate": 1.4583732057416269e-05,
|
| 58 |
+
"loss": 0.8922,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.7017543859649122,
|
| 63 |
+
"grad_norm": 12.121639251708984,
|
| 64 |
+
"learning_rate": 1.4523923444976078e-05,
|
| 65 |
+
"loss": 0.9787,
|
| 66 |
+
"step": 200
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 0.7894736842105263,
|
| 70 |
+
"grad_norm": 14.091375350952148,
|
| 71 |
+
"learning_rate": 1.4464114832535886e-05,
|
| 72 |
+
"loss": 0.806,
|
| 73 |
+
"step": 225
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.8771929824561403,
|
| 77 |
+
"grad_norm": 9.919096946716309,
|
| 78 |
+
"learning_rate": 1.4404306220095696e-05,
|
| 79 |
+
"loss": 0.9935,
|
| 80 |
+
"step": 250
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.9649122807017544,
|
| 84 |
+
"grad_norm": 13.80988597869873,
|
| 85 |
+
"learning_rate": 1.4344497607655503e-05,
|
| 86 |
+
"loss": 0.8116,
|
| 87 |
+
"step": 275
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 1.0526315789473684,
|
| 91 |
+
"grad_norm": 13.865740776062012,
|
| 92 |
+
"learning_rate": 1.4284688995215311e-05,
|
| 93 |
+
"loss": 0.7288,
|
| 94 |
+
"step": 300
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 1.1403508771929824,
|
| 98 |
+
"grad_norm": 9.76303768157959,
|
| 99 |
+
"learning_rate": 1.4224880382775121e-05,
|
| 100 |
+
"loss": 0.8283,
|
| 101 |
+
"step": 325
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 1.2280701754385965,
|
| 105 |
+
"grad_norm": 9.364381790161133,
|
| 106 |
+
"learning_rate": 1.4165071770334929e-05,
|
| 107 |
+
"loss": 0.6827,
|
| 108 |
+
"step": 350
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 1.3157894736842106,
|
| 112 |
+
"grad_norm": 13.79258918762207,
|
| 113 |
+
"learning_rate": 1.4105263157894738e-05,
|
| 114 |
+
"loss": 0.7745,
|
| 115 |
+
"step": 375
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 1.4035087719298245,
|
| 119 |
+
"grad_norm": 9.642690658569336,
|
| 120 |
+
"learning_rate": 1.4045454545454546e-05,
|
| 121 |
+
"loss": 0.7514,
|
| 122 |
+
"step": 400
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 1.4912280701754386,
|
| 126 |
+
"grad_norm": 14.490239143371582,
|
| 127 |
+
"learning_rate": 1.3985645933014356e-05,
|
| 128 |
+
"loss": 0.6594,
|
| 129 |
+
"step": 425
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 1.5789473684210527,
|
| 133 |
+
"grad_norm": 13.364012718200684,
|
| 134 |
+
"learning_rate": 1.3925837320574163e-05,
|
| 135 |
+
"loss": 0.6036,
|
| 136 |
+
"step": 450
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 1.6666666666666665,
|
| 140 |
+
"grad_norm": 6.50977087020874,
|
| 141 |
+
"learning_rate": 1.3866028708133971e-05,
|
| 142 |
+
"loss": 0.7616,
|
| 143 |
+
"step": 475
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 1.7543859649122808,
|
| 147 |
+
"grad_norm": 30.46417808532715,
|
| 148 |
+
"learning_rate": 1.380622009569378e-05,
|
| 149 |
+
"loss": 0.6735,
|
| 150 |
+
"step": 500
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 1.8421052631578947,
|
| 154 |
+
"grad_norm": 17.846593856811523,
|
| 155 |
+
"learning_rate": 1.3746411483253589e-05,
|
| 156 |
+
"loss": 0.5902,
|
| 157 |
+
"step": 525
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 1.9298245614035088,
|
| 161 |
+
"grad_norm": 19.83995819091797,
|
| 162 |
+
"learning_rate": 1.3686602870813398e-05,
|
| 163 |
+
"loss": 0.5928,
|
| 164 |
+
"step": 550
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 2.017543859649123,
|
| 168 |
+
"grad_norm": 6.053647994995117,
|
| 169 |
+
"learning_rate": 1.3626794258373206e-05,
|
| 170 |
+
"loss": 0.5816,
|
| 171 |
+
"step": 575
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 2.1052631578947367,
|
| 175 |
+
"grad_norm": 9.594230651855469,
|
| 176 |
+
"learning_rate": 1.3566985645933015e-05,
|
| 177 |
+
"loss": 0.488,
|
| 178 |
+
"step": 600
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 2.192982456140351,
|
| 182 |
+
"grad_norm": 13.84262752532959,
|
| 183 |
+
"learning_rate": 1.3507177033492823e-05,
|
| 184 |
+
"loss": 0.568,
|
| 185 |
+
"step": 625
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 2.280701754385965,
|
| 189 |
+
"grad_norm": 10.86662483215332,
|
| 190 |
+
"learning_rate": 1.3447368421052633e-05,
|
| 191 |
+
"loss": 0.4017,
|
| 192 |
+
"step": 650
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 2.3684210526315788,
|
| 196 |
+
"grad_norm": 16.06069564819336,
|
| 197 |
+
"learning_rate": 1.338755980861244e-05,
|
| 198 |
+
"loss": 0.5112,
|
| 199 |
+
"step": 675
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 2.456140350877193,
|
| 203 |
+
"grad_norm": 6.421328544616699,
|
| 204 |
+
"learning_rate": 1.3327751196172248e-05,
|
| 205 |
+
"loss": 0.3609,
|
| 206 |
+
"step": 700
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 2.543859649122807,
|
| 210 |
+
"grad_norm": 6.225348949432373,
|
| 211 |
+
"learning_rate": 1.3267942583732058e-05,
|
| 212 |
+
"loss": 0.5531,
|
| 213 |
+
"step": 725
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 2.6315789473684212,
|
| 217 |
+
"grad_norm": 14.942207336425781,
|
| 218 |
+
"learning_rate": 1.3208133971291866e-05,
|
| 219 |
+
"loss": 0.4272,
|
| 220 |
+
"step": 750
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 2.719298245614035,
|
| 224 |
+
"grad_norm": 8.975879669189453,
|
| 225 |
+
"learning_rate": 1.3148325358851675e-05,
|
| 226 |
+
"loss": 0.4509,
|
| 227 |
+
"step": 775
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 2.807017543859649,
|
| 231 |
+
"grad_norm": 21.482088088989258,
|
| 232 |
+
"learning_rate": 1.3088516746411483e-05,
|
| 233 |
+
"loss": 0.4233,
|
| 234 |
+
"step": 800
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 2.8947368421052633,
|
| 238 |
+
"grad_norm": 5.1405792236328125,
|
| 239 |
+
"learning_rate": 1.3028708133971293e-05,
|
| 240 |
+
"loss": 0.418,
|
| 241 |
+
"step": 825
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 2.982456140350877,
|
| 245 |
+
"grad_norm": 14.574483871459961,
|
| 246 |
+
"learning_rate": 1.29688995215311e-05,
|
| 247 |
+
"loss": 0.6778,
|
| 248 |
+
"step": 850
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 3.0701754385964914,
|
| 252 |
+
"grad_norm": 7.385740280151367,
|
| 253 |
+
"learning_rate": 1.290909090909091e-05,
|
| 254 |
+
"loss": 0.4127,
|
| 255 |
+
"step": 875
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 3.1578947368421053,
|
| 259 |
+
"grad_norm": 48.35734939575195,
|
| 260 |
+
"learning_rate": 1.2849282296650718e-05,
|
| 261 |
+
"loss": 0.4752,
|
| 262 |
+
"step": 900
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 3.245614035087719,
|
| 266 |
+
"grad_norm": 8.056925773620605,
|
| 267 |
+
"learning_rate": 1.2789473684210526e-05,
|
| 268 |
+
"loss": 0.411,
|
| 269 |
+
"step": 925
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 3.3333333333333335,
|
| 273 |
+
"grad_norm": 20.510982513427734,
|
| 274 |
+
"learning_rate": 1.2729665071770335e-05,
|
| 275 |
+
"loss": 0.495,
|
| 276 |
+
"step": 950
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 3.4210526315789473,
|
| 280 |
+
"grad_norm": 25.063684463500977,
|
| 281 |
+
"learning_rate": 1.2669856459330143e-05,
|
| 282 |
+
"loss": 0.2436,
|
| 283 |
+
"step": 975
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 3.5087719298245617,
|
| 287 |
+
"grad_norm": 15.523259162902832,
|
| 288 |
+
"learning_rate": 1.2610047846889953e-05,
|
| 289 |
+
"loss": 0.4612,
|
| 290 |
+
"step": 1000
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 3.5964912280701755,
|
| 294 |
+
"grad_norm": 1.9646857976913452,
|
| 295 |
+
"learning_rate": 1.255023923444976e-05,
|
| 296 |
+
"loss": 0.4426,
|
| 297 |
+
"step": 1025
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 3.6842105263157894,
|
| 301 |
+
"grad_norm": 30.676250457763672,
|
| 302 |
+
"learning_rate": 1.249043062200957e-05,
|
| 303 |
+
"loss": 0.3333,
|
| 304 |
+
"step": 1050
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 3.7719298245614032,
|
| 308 |
+
"grad_norm": 6.475268363952637,
|
| 309 |
+
"learning_rate": 1.2430622009569378e-05,
|
| 310 |
+
"loss": 0.2945,
|
| 311 |
+
"step": 1075
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 3.8596491228070176,
|
| 315 |
+
"grad_norm": 24.684438705444336,
|
| 316 |
+
"learning_rate": 1.2370813397129186e-05,
|
| 317 |
+
"loss": 0.4339,
|
| 318 |
+
"step": 1100
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 3.9473684210526314,
|
| 322 |
+
"grad_norm": 38.601226806640625,
|
| 323 |
+
"learning_rate": 1.2311004784688995e-05,
|
| 324 |
+
"loss": 0.3701,
|
| 325 |
+
"step": 1125
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 4.035087719298246,
|
| 329 |
+
"grad_norm": 2.8431146144866943,
|
| 330 |
+
"learning_rate": 1.2251196172248803e-05,
|
| 331 |
+
"loss": 0.2782,
|
| 332 |
+
"step": 1150
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 4.12280701754386,
|
| 336 |
+
"grad_norm": 23.384532928466797,
|
| 337 |
+
"learning_rate": 1.2191387559808613e-05,
|
| 338 |
+
"loss": 0.312,
|
| 339 |
+
"step": 1175
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 4.2105263157894735,
|
| 343 |
+
"grad_norm": 3.222729444503784,
|
| 344 |
+
"learning_rate": 1.213157894736842e-05,
|
| 345 |
+
"loss": 0.264,
|
| 346 |
+
"step": 1200
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 4.298245614035087,
|
| 350 |
+
"grad_norm": 3.5861947536468506,
|
| 351 |
+
"learning_rate": 1.207177033492823e-05,
|
| 352 |
+
"loss": 0.3104,
|
| 353 |
+
"step": 1225
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 4.385964912280702,
|
| 357 |
+
"grad_norm": 4.091048717498779,
|
| 358 |
+
"learning_rate": 1.2011961722488038e-05,
|
| 359 |
+
"loss": 0.3567,
|
| 360 |
+
"step": 1250
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 4.473684210526316,
|
| 364 |
+
"grad_norm": 1.3814598321914673,
|
| 365 |
+
"learning_rate": 1.1952153110047847e-05,
|
| 366 |
+
"loss": 0.3455,
|
| 367 |
+
"step": 1275
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 4.56140350877193,
|
| 371 |
+
"grad_norm": 2.480316162109375,
|
| 372 |
+
"learning_rate": 1.1892344497607655e-05,
|
| 373 |
+
"loss": 0.1769,
|
| 374 |
+
"step": 1300
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 4.649122807017544,
|
| 378 |
+
"grad_norm": 20.069162368774414,
|
| 379 |
+
"learning_rate": 1.1832535885167463e-05,
|
| 380 |
+
"loss": 0.2483,
|
| 381 |
+
"step": 1325
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 4.7368421052631575,
|
| 385 |
+
"grad_norm": 1.7885949611663818,
|
| 386 |
+
"learning_rate": 1.1772727272727272e-05,
|
| 387 |
+
"loss": 0.2283,
|
| 388 |
+
"step": 1350
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 4.824561403508772,
|
| 392 |
+
"grad_norm": 40.37774658203125,
|
| 393 |
+
"learning_rate": 1.171291866028708e-05,
|
| 394 |
+
"loss": 0.2277,
|
| 395 |
+
"step": 1375
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 4.912280701754386,
|
| 399 |
+
"grad_norm": 2.960739850997925,
|
| 400 |
+
"learning_rate": 1.1653110047846891e-05,
|
| 401 |
+
"loss": 0.2077,
|
| 402 |
+
"step": 1400
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 5.0,
|
| 406 |
+
"grad_norm": 14.10015869140625,
|
| 407 |
+
"learning_rate": 1.15933014354067e-05,
|
| 408 |
+
"loss": 0.3859,
|
| 409 |
+
"step": 1425
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 5.087719298245614,
|
| 413 |
+
"grad_norm": 4.440461158752441,
|
| 414 |
+
"learning_rate": 1.1533492822966509e-05,
|
| 415 |
+
"loss": 0.1354,
|
| 416 |
+
"step": 1450
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 5.175438596491228,
|
| 420 |
+
"grad_norm": 17.193849563598633,
|
| 421 |
+
"learning_rate": 1.1473684210526317e-05,
|
| 422 |
+
"loss": 0.2935,
|
| 423 |
+
"step": 1475
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 5.2631578947368425,
|
| 427 |
+
"grad_norm": 1.2828125953674316,
|
| 428 |
+
"learning_rate": 1.1413875598086125e-05,
|
| 429 |
+
"loss": 0.1394,
|
| 430 |
+
"step": 1500
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 5.350877192982456,
|
| 434 |
+
"grad_norm": 4.809077262878418,
|
| 435 |
+
"learning_rate": 1.1354066985645934e-05,
|
| 436 |
+
"loss": 0.3651,
|
| 437 |
+
"step": 1525
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 5.43859649122807,
|
| 441 |
+
"grad_norm": 43.53572463989258,
|
| 442 |
+
"learning_rate": 1.1294258373205742e-05,
|
| 443 |
+
"loss": 0.1187,
|
| 444 |
+
"step": 1550
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 5.526315789473684,
|
| 448 |
+
"grad_norm": 5.647329330444336,
|
| 449 |
+
"learning_rate": 1.1234449760765551e-05,
|
| 450 |
+
"loss": 0.2719,
|
| 451 |
+
"step": 1575
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 5.614035087719298,
|
| 455 |
+
"grad_norm": 25.37187957763672,
|
| 456 |
+
"learning_rate": 1.117464114832536e-05,
|
| 457 |
+
"loss": 0.2477,
|
| 458 |
+
"step": 1600
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 5.701754385964913,
|
| 462 |
+
"grad_norm": 4.228896617889404,
|
| 463 |
+
"learning_rate": 1.1114832535885169e-05,
|
| 464 |
+
"loss": 0.3608,
|
| 465 |
+
"step": 1625
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 5.7894736842105265,
|
| 469 |
+
"grad_norm": 2.1760356426239014,
|
| 470 |
+
"learning_rate": 1.1055023923444977e-05,
|
| 471 |
+
"loss": 0.1223,
|
| 472 |
+
"step": 1650
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 5.87719298245614,
|
| 476 |
+
"grad_norm": 28.078975677490234,
|
| 477 |
+
"learning_rate": 1.0995215311004786e-05,
|
| 478 |
+
"loss": 0.3207,
|
| 479 |
+
"step": 1675
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 5.964912280701754,
|
| 483 |
+
"grad_norm": 12.583624839782715,
|
| 484 |
+
"learning_rate": 1.0935406698564594e-05,
|
| 485 |
+
"loss": 0.1094,
|
| 486 |
+
"step": 1700
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 6.052631578947368,
|
| 490 |
+
"grad_norm": 0.4104742109775543,
|
| 491 |
+
"learning_rate": 1.0875598086124402e-05,
|
| 492 |
+
"loss": 0.1937,
|
| 493 |
+
"step": 1725
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 6.140350877192983,
|
| 497 |
+
"grad_norm": 1.5135095119476318,
|
| 498 |
+
"learning_rate": 1.0815789473684211e-05,
|
| 499 |
+
"loss": 0.3113,
|
| 500 |
+
"step": 1750
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 6.228070175438597,
|
| 504 |
+
"grad_norm": 1.9227218627929688,
|
| 505 |
+
"learning_rate": 1.0755980861244019e-05,
|
| 506 |
+
"loss": 0.1179,
|
| 507 |
+
"step": 1775
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 6.315789473684211,
|
| 511 |
+
"grad_norm": 7.322267055511475,
|
| 512 |
+
"learning_rate": 1.0696172248803829e-05,
|
| 513 |
+
"loss": 0.209,
|
| 514 |
+
"step": 1800
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 6.4035087719298245,
|
| 518 |
+
"grad_norm": 2.8994693756103516,
|
| 519 |
+
"learning_rate": 1.0636363636363636e-05,
|
| 520 |
+
"loss": 0.1743,
|
| 521 |
+
"step": 1825
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 6.491228070175438,
|
| 525 |
+
"grad_norm": 21.65730857849121,
|
| 526 |
+
"learning_rate": 1.0576555023923446e-05,
|
| 527 |
+
"loss": 0.2483,
|
| 528 |
+
"step": 1850
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 6.578947368421053,
|
| 532 |
+
"grad_norm": 24.210969924926758,
|
| 533 |
+
"learning_rate": 1.0516746411483254e-05,
|
| 534 |
+
"loss": 0.1348,
|
| 535 |
+
"step": 1875
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 6.666666666666667,
|
| 539 |
+
"grad_norm": 1.589753270149231,
|
| 540 |
+
"learning_rate": 1.0456937799043063e-05,
|
| 541 |
+
"loss": 0.229,
|
| 542 |
+
"step": 1900
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 6.754385964912281,
|
| 546 |
+
"grad_norm": 3.508387327194214,
|
| 547 |
+
"learning_rate": 1.0397129186602871e-05,
|
| 548 |
+
"loss": 0.0909,
|
| 549 |
+
"step": 1925
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 6.842105263157895,
|
| 553 |
+
"grad_norm": 1.89371919631958,
|
| 554 |
+
"learning_rate": 1.0337320574162679e-05,
|
| 555 |
+
"loss": 0.1859,
|
| 556 |
+
"step": 1950
|
| 557 |
+
},
|
| 558 |
+
{
|
| 559 |
+
"epoch": 6.9298245614035086,
|
| 560 |
+
"grad_norm": 0.35574591159820557,
|
| 561 |
+
"learning_rate": 1.0277511961722489e-05,
|
| 562 |
+
"loss": 0.1675,
|
| 563 |
+
"step": 1975
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 7.017543859649122,
|
| 567 |
+
"grad_norm": 0.621152937412262,
|
| 568 |
+
"learning_rate": 1.0217703349282296e-05,
|
| 569 |
+
"loss": 0.1082,
|
| 570 |
+
"step": 2000
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 7.105263157894737,
|
| 574 |
+
"grad_norm": 0.5694774985313416,
|
| 575 |
+
"learning_rate": 1.0157894736842106e-05,
|
| 576 |
+
"loss": 0.0886,
|
| 577 |
+
"step": 2025
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 7.192982456140351,
|
| 581 |
+
"grad_norm": 7.276403427124023,
|
| 582 |
+
"learning_rate": 1.0098086124401914e-05,
|
| 583 |
+
"loss": 0.2697,
|
| 584 |
+
"step": 2050
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 7.280701754385965,
|
| 588 |
+
"grad_norm": 0.4015137851238251,
|
| 589 |
+
"learning_rate": 1.0038277511961723e-05,
|
| 590 |
+
"loss": 0.1301,
|
| 591 |
+
"step": 2075
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 7.368421052631579,
|
| 595 |
+
"grad_norm": 0.27876588702201843,
|
| 596 |
+
"learning_rate": 9.978468899521531e-06,
|
| 597 |
+
"loss": 0.0687,
|
| 598 |
+
"step": 2100
|
| 599 |
+
},
|
| 600 |
+
{
|
| 601 |
+
"epoch": 7.456140350877193,
|
| 602 |
+
"grad_norm": 7.235930442810059,
|
| 603 |
+
"learning_rate": 9.918660287081339e-06,
|
| 604 |
+
"loss": 0.0734,
|
| 605 |
+
"step": 2125
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 7.543859649122807,
|
| 609 |
+
"grad_norm": 1.6751177310943604,
|
| 610 |
+
"learning_rate": 9.858851674641148e-06,
|
| 611 |
+
"loss": 0.1574,
|
| 612 |
+
"step": 2150
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 7.631578947368421,
|
| 616 |
+
"grad_norm": 0.8539127111434937,
|
| 617 |
+
"learning_rate": 9.799043062200956e-06,
|
| 618 |
+
"loss": 0.0483,
|
| 619 |
+
"step": 2175
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"epoch": 7.719298245614035,
|
| 623 |
+
"grad_norm": 0.26878660917282104,
|
| 624 |
+
"learning_rate": 9.739234449760766e-06,
|
| 625 |
+
"loss": 0.1632,
|
| 626 |
+
"step": 2200
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 7.807017543859649,
|
| 630 |
+
"grad_norm": 106.50234985351562,
|
| 631 |
+
"learning_rate": 9.679425837320574e-06,
|
| 632 |
+
"loss": 0.1156,
|
| 633 |
+
"step": 2225
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 7.894736842105263,
|
| 637 |
+
"grad_norm": 2.002487897872925,
|
| 638 |
+
"learning_rate": 9.619617224880383e-06,
|
| 639 |
+
"loss": 0.2036,
|
| 640 |
+
"step": 2250
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 7.982456140350877,
|
| 644 |
+
"grad_norm": 0.5302895307540894,
|
| 645 |
+
"learning_rate": 9.559808612440191e-06,
|
| 646 |
+
"loss": 0.1417,
|
| 647 |
+
"step": 2275
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 8.070175438596491,
|
| 651 |
+
"grad_norm": 0.314165323972702,
|
| 652 |
+
"learning_rate": 9.5e-06,
|
| 653 |
+
"loss": 0.0673,
|
| 654 |
+
"step": 2300
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 8.157894736842104,
|
| 658 |
+
"grad_norm": 0.5242946743965149,
|
| 659 |
+
"learning_rate": 9.440191387559808e-06,
|
| 660 |
+
"loss": 0.0689,
|
| 661 |
+
"step": 2325
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 8.24561403508772,
|
| 665 |
+
"grad_norm": 48.71388244628906,
|
| 666 |
+
"learning_rate": 9.380382775119616e-06,
|
| 667 |
+
"loss": 0.1863,
|
| 668 |
+
"step": 2350
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 8.333333333333334,
|
| 672 |
+
"grad_norm": 0.5670514106750488,
|
| 673 |
+
"learning_rate": 9.320574162679426e-06,
|
| 674 |
+
"loss": 0.0823,
|
| 675 |
+
"step": 2375
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 8.421052631578947,
|
| 679 |
+
"grad_norm": 0.6025238633155823,
|
| 680 |
+
"learning_rate": 9.260765550239234e-06,
|
| 681 |
+
"loss": 0.1118,
|
| 682 |
+
"step": 2400
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 8.508771929824562,
|
| 686 |
+
"grad_norm": 42.649696350097656,
|
| 687 |
+
"learning_rate": 9.200956937799043e-06,
|
| 688 |
+
"loss": 0.1333,
|
| 689 |
+
"step": 2425
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 8.596491228070175,
|
| 693 |
+
"grad_norm": 0.4079756438732147,
|
| 694 |
+
"learning_rate": 9.141148325358851e-06,
|
| 695 |
+
"loss": 0.156,
|
| 696 |
+
"step": 2450
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 8.68421052631579,
|
| 700 |
+
"grad_norm": 0.31388023495674133,
|
| 701 |
+
"learning_rate": 9.08133971291866e-06,
|
| 702 |
+
"loss": 0.2159,
|
| 703 |
+
"step": 2475
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 8.771929824561404,
|
| 707 |
+
"grad_norm": 0.2017977237701416,
|
| 708 |
+
"learning_rate": 9.021531100478468e-06,
|
| 709 |
+
"loss": 0.0936,
|
| 710 |
+
"step": 2500
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 8.859649122807017,
|
| 714 |
+
"grad_norm": 2.2358014583587646,
|
| 715 |
+
"learning_rate": 6.697368421052631e-06,
|
| 716 |
+
"loss": 0.1884,
|
| 717 |
+
"step": 2525
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 8.947368421052632,
|
| 721 |
+
"grad_norm": 0.5669077038764954,
|
| 722 |
+
"learning_rate": 6.615131578947369e-06,
|
| 723 |
+
"loss": 0.2183,
|
| 724 |
+
"step": 2550
|
| 725 |
+
},
|
| 726 |
+
{
|
| 727 |
+
"epoch": 9.035087719298245,
|
| 728 |
+
"grad_norm": 27.18621253967285,
|
| 729 |
+
"learning_rate": 6.532894736842106e-06,
|
| 730 |
+
"loss": 0.1232,
|
| 731 |
+
"step": 2575
|
| 732 |
+
},
|
| 733 |
+
{
|
| 734 |
+
"epoch": 9.12280701754386,
|
| 735 |
+
"grad_norm": 468.5591735839844,
|
| 736 |
+
"learning_rate": 6.450657894736842e-06,
|
| 737 |
+
"loss": 0.0516,
|
| 738 |
+
"step": 2600
|
| 739 |
+
},
|
| 740 |
+
{
|
| 741 |
+
"epoch": 9.210526315789474,
|
| 742 |
+
"grad_norm": 0.24007272720336914,
|
| 743 |
+
"learning_rate": 6.368421052631579e-06,
|
| 744 |
+
"loss": 0.0695,
|
| 745 |
+
"step": 2625
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 9.298245614035087,
|
| 749 |
+
"grad_norm": 99.38546752929688,
|
| 750 |
+
"learning_rate": 6.286184210526316e-06,
|
| 751 |
+
"loss": 0.0947,
|
| 752 |
+
"step": 2650
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 9.385964912280702,
|
| 756 |
+
"grad_norm": 3.8673267364501953,
|
| 757 |
+
"learning_rate": 6.203947368421053e-06,
|
| 758 |
+
"loss": 0.1062,
|
| 759 |
+
"step": 2675
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 9.473684210526315,
|
| 763 |
+
"grad_norm": 0.2559914290904999,
|
| 764 |
+
"learning_rate": 6.12171052631579e-06,
|
| 765 |
+
"loss": 0.0523,
|
| 766 |
+
"step": 2700
|
| 767 |
+
},
|
| 768 |
+
{
|
| 769 |
+
"epoch": 9.56140350877193,
|
| 770 |
+
"grad_norm": 0.12371934205293655,
|
| 771 |
+
"learning_rate": 6.039473684210526e-06,
|
| 772 |
+
"loss": 0.0495,
|
| 773 |
+
"step": 2725
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"epoch": 9.649122807017545,
|
| 777 |
+
"grad_norm": 2.5249898433685303,
|
| 778 |
+
"learning_rate": 5.957236842105263e-06,
|
| 779 |
+
"loss": 0.1678,
|
| 780 |
+
"step": 2750
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"epoch": 9.736842105263158,
|
| 784 |
+
"grad_norm": 0.2619442939758301,
|
| 785 |
+
"learning_rate": 5.8750000000000005e-06,
|
| 786 |
+
"loss": 0.0534,
|
| 787 |
+
"step": 2775
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"epoch": 9.824561403508772,
|
| 791 |
+
"grad_norm": 1.3705483675003052,
|
| 792 |
+
"learning_rate": 5.792763157894737e-06,
|
| 793 |
+
"loss": 0.1326,
|
| 794 |
+
"step": 2800
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 9.912280701754385,
|
| 798 |
+
"grad_norm": 74.48184967041016,
|
| 799 |
+
"learning_rate": 5.710526315789474e-06,
|
| 800 |
+
"loss": 0.1795,
|
| 801 |
+
"step": 2825
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 10.0,
|
| 805 |
+
"grad_norm": 79.78992462158203,
|
| 806 |
+
"learning_rate": 5.6282894736842106e-06,
|
| 807 |
+
"loss": 0.092,
|
| 808 |
+
"step": 2850
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"epoch": 10.087719298245615,
|
| 812 |
+
"grad_norm": 0.5286645889282227,
|
| 813 |
+
"learning_rate": 5.546052631578947e-06,
|
| 814 |
+
"loss": 0.0418,
|
| 815 |
+
"step": 2875
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"epoch": 10.175438596491228,
|
| 819 |
+
"grad_norm": 109.82157135009766,
|
| 820 |
+
"learning_rate": 5.463815789473684e-06,
|
| 821 |
+
"loss": 0.193,
|
| 822 |
+
"step": 2900
|
| 823 |
+
},
|
| 824 |
+
{
|
| 825 |
+
"epoch": 10.263157894736842,
|
| 826 |
+
"grad_norm": 3.963012933731079,
|
| 827 |
+
"learning_rate": 5.3815789473684215e-06,
|
| 828 |
+
"loss": 0.0897,
|
| 829 |
+
"step": 2925
|
| 830 |
+
},
|
| 831 |
+
{
|
| 832 |
+
"epoch": 10.350877192982455,
|
| 833 |
+
"grad_norm": 0.41540616750717163,
|
| 834 |
+
"learning_rate": 5.299342105263158e-06,
|
| 835 |
+
"loss": 0.1723,
|
| 836 |
+
"step": 2950
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 10.43859649122807,
|
| 840 |
+
"grad_norm": 5.900774002075195,
|
| 841 |
+
"learning_rate": 5.217105263157895e-06,
|
| 842 |
+
"loss": 0.0784,
|
| 843 |
+
"step": 2975
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 10.526315789473685,
|
| 847 |
+
"grad_norm": 12.152785301208496,
|
| 848 |
+
"learning_rate": 5.1348684210526315e-06,
|
| 849 |
+
"loss": 0.1186,
|
| 850 |
+
"step": 3000
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"epoch": 10.614035087719298,
|
| 854 |
+
"grad_norm": 0.21494676172733307,
|
| 855 |
+
"learning_rate": 3.6315789473684213e-06,
|
| 856 |
+
"loss": 0.0358,
|
| 857 |
+
"step": 3025
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 10.701754385964913,
|
| 861 |
+
"grad_norm": 90.48876953125,
|
| 862 |
+
"learning_rate": 3.537593984962406e-06,
|
| 863 |
+
"loss": 0.0471,
|
| 864 |
+
"step": 3050
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 10.789473684210526,
|
| 868 |
+
"grad_norm": 0.23261310160160065,
|
| 869 |
+
"learning_rate": 3.4436090225563912e-06,
|
| 870 |
+
"loss": 0.0276,
|
| 871 |
+
"step": 3075
|
| 872 |
+
},
|
| 873 |
+
{
|
| 874 |
+
"epoch": 10.87719298245614,
|
| 875 |
+
"grad_norm": 0.782564640045166,
|
| 876 |
+
"learning_rate": 3.349624060150376e-06,
|
| 877 |
+
"loss": 0.0718,
|
| 878 |
+
"step": 3100
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 10.964912280701755,
|
| 882 |
+
"grad_norm": 0.33198878169059753,
|
| 883 |
+
"learning_rate": 3.255639097744361e-06,
|
| 884 |
+
"loss": 0.0564,
|
| 885 |
+
"step": 3125
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 11.052631578947368,
|
| 889 |
+
"grad_norm": 0.14864841103553772,
|
| 890 |
+
"learning_rate": 3.161654135338346e-06,
|
| 891 |
+
"loss": 0.0508,
|
| 892 |
+
"step": 3150
|
| 893 |
+
},
|
| 894 |
+
{
|
| 895 |
+
"epoch": 11.140350877192983,
|
| 896 |
+
"grad_norm": 0.1588129848241806,
|
| 897 |
+
"learning_rate": 3.067669172932331e-06,
|
| 898 |
+
"loss": 0.1641,
|
| 899 |
+
"step": 3175
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 11.228070175438596,
|
| 903 |
+
"grad_norm": 0.6562497019767761,
|
| 904 |
+
"learning_rate": 2.973684210526316e-06,
|
| 905 |
+
"loss": 0.0544,
|
| 906 |
+
"step": 3200
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"epoch": 11.31578947368421,
|
| 910 |
+
"grad_norm": 0.3608649969100952,
|
| 911 |
+
"learning_rate": 2.879699248120301e-06,
|
| 912 |
+
"loss": 0.0391,
|
| 913 |
+
"step": 3225
|
| 914 |
+
},
|
| 915 |
+
{
|
| 916 |
+
"epoch": 11.403508771929825,
|
| 917 |
+
"grad_norm": 20.602819442749023,
|
| 918 |
+
"learning_rate": 2.785714285714286e-06,
|
| 919 |
+
"loss": 0.0338,
|
| 920 |
+
"step": 3250
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 11.491228070175438,
|
| 924 |
+
"grad_norm": 157.23171997070312,
|
| 925 |
+
"learning_rate": 2.691729323308271e-06,
|
| 926 |
+
"loss": 0.0771,
|
| 927 |
+
"step": 3275
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 11.578947368421053,
|
| 931 |
+
"grad_norm": 0.22338764369487762,
|
| 932 |
+
"learning_rate": 2.597744360902256e-06,
|
| 933 |
+
"loss": 0.1187,
|
| 934 |
+
"step": 3300
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 11.666666666666666,
|
| 938 |
+
"grad_norm": 57.78382110595703,
|
| 939 |
+
"learning_rate": 2.503759398496241e-06,
|
| 940 |
+
"loss": 0.0165,
|
| 941 |
+
"step": 3325
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 11.75438596491228,
|
| 945 |
+
"grad_norm": 7.400601863861084,
|
| 946 |
+
"learning_rate": 2.4097744360902257e-06,
|
| 947 |
+
"loss": 0.0092,
|
| 948 |
+
"step": 3350
|
| 949 |
+
},
|
| 950 |
+
{
|
| 951 |
+
"epoch": 11.842105263157894,
|
| 952 |
+
"grad_norm": 0.18190287053585052,
|
| 953 |
+
"learning_rate": 2.3157894736842105e-06,
|
| 954 |
+
"loss": 0.1059,
|
| 955 |
+
"step": 3375
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 11.929824561403509,
|
| 959 |
+
"grad_norm": 0.1269238144159317,
|
| 960 |
+
"learning_rate": 2.2218045112781957e-06,
|
| 961 |
+
"loss": 0.0571,
|
| 962 |
+
"step": 3400
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 12.017543859649123,
|
| 966 |
+
"grad_norm": 0.10843072831630707,
|
| 967 |
+
"learning_rate": 2.1278195488721805e-06,
|
| 968 |
+
"loss": 0.1341,
|
| 969 |
+
"step": 3425
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 12.105263157894736,
|
| 973 |
+
"grad_norm": 0.7199507355690002,
|
| 974 |
+
"learning_rate": 2.0338345864661656e-06,
|
| 975 |
+
"loss": 0.0063,
|
| 976 |
+
"step": 3450
|
| 977 |
+
},
|
| 978 |
+
{
|
| 979 |
+
"epoch": 12.192982456140351,
|
| 980 |
+
"grad_norm": 0.19009971618652344,
|
| 981 |
+
"learning_rate": 1.9398496240601504e-06,
|
| 982 |
+
"loss": 0.054,
|
| 983 |
+
"step": 3475
|
| 984 |
+
},
|
| 985 |
+
{
|
| 986 |
+
"epoch": 12.280701754385966,
|
| 987 |
+
"grad_norm": 0.18020951747894287,
|
| 988 |
+
"learning_rate": 1.8458646616541354e-06,
|
| 989 |
+
"loss": 0.0148,
|
| 990 |
+
"step": 3500
|
| 991 |
+
},
|
| 992 |
+
{
|
| 993 |
+
"epoch": 12.368421052631579,
|
| 994 |
+
"grad_norm": 0.2494085282087326,
|
| 995 |
+
"learning_rate": 6.569377990430622e-06,
|
| 996 |
+
"loss": 0.1168,
|
| 997 |
+
"step": 3525
|
| 998 |
+
},
|
| 999 |
+
{
|
| 1000 |
+
"epoch": 12.456140350877194,
|
| 1001 |
+
"grad_norm": 0.16571170091629028,
|
| 1002 |
+
"learning_rate": 6.509569377990431e-06,
|
| 1003 |
+
"loss": 0.0482,
|
| 1004 |
+
"step": 3550
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 12.543859649122806,
|
| 1008 |
+
"grad_norm": 0.10973577946424484,
|
| 1009 |
+
"learning_rate": 6.44976076555024e-06,
|
| 1010 |
+
"loss": 0.1372,
|
| 1011 |
+
"step": 3575
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 12.631578947368421,
|
| 1015 |
+
"grad_norm": 0.14920943975448608,
|
| 1016 |
+
"learning_rate": 6.389952153110048e-06,
|
| 1017 |
+
"loss": 0.0791,
|
| 1018 |
+
"step": 3600
|
| 1019 |
+
},
|
| 1020 |
+
{
|
| 1021 |
+
"epoch": 12.719298245614034,
|
| 1022 |
+
"grad_norm": 4.5087504386901855,
|
| 1023 |
+
"learning_rate": 6.330143540669857e-06,
|
| 1024 |
+
"loss": 0.1614,
|
| 1025 |
+
"step": 3625
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 12.807017543859649,
|
| 1029 |
+
"grad_norm": 0.11903402209281921,
|
| 1030 |
+
"learning_rate": 6.270334928229666e-06,
|
| 1031 |
+
"loss": 0.0082,
|
| 1032 |
+
"step": 3650
|
| 1033 |
+
},
|
| 1034 |
+
{
|
| 1035 |
+
"epoch": 12.894736842105264,
|
| 1036 |
+
"grad_norm": 0.24349439144134521,
|
| 1037 |
+
"learning_rate": 6.210526315789474e-06,
|
| 1038 |
+
"loss": 0.1936,
|
| 1039 |
+
"step": 3675
|
| 1040 |
+
},
|
| 1041 |
+
{
|
| 1042 |
+
"epoch": 12.982456140350877,
|
| 1043 |
+
"grad_norm": 146.2913360595703,
|
| 1044 |
+
"learning_rate": 6.150717703349282e-06,
|
| 1045 |
+
"loss": 0.1465,
|
| 1046 |
+
"step": 3700
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 13.070175438596491,
|
| 1050 |
+
"grad_norm": 88.66224670410156,
|
| 1051 |
+
"learning_rate": 6.090909090909091e-06,
|
| 1052 |
+
"loss": 0.0751,
|
| 1053 |
+
"step": 3725
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 13.157894736842104,
|
| 1057 |
+
"grad_norm": 0.11558452993631363,
|
| 1058 |
+
"learning_rate": 6.0311004784689e-06,
|
| 1059 |
+
"loss": 0.0652,
|
| 1060 |
+
"step": 3750
|
| 1061 |
+
},
|
| 1062 |
+
{
|
| 1063 |
+
"epoch": 13.24561403508772,
|
| 1064 |
+
"grad_norm": 3.1845107078552246,
|
| 1065 |
+
"learning_rate": 5.971291866028708e-06,
|
| 1066 |
+
"loss": 0.0969,
|
| 1067 |
+
"step": 3775
|
| 1068 |
+
},
|
| 1069 |
+
{
|
| 1070 |
+
"epoch": 13.333333333333334,
|
| 1071 |
+
"grad_norm": 0.5500511527061462,
|
| 1072 |
+
"learning_rate": 5.911483253588517e-06,
|
| 1073 |
+
"loss": 0.107,
|
| 1074 |
+
"step": 3800
|
| 1075 |
+
},
|
| 1076 |
+
{
|
| 1077 |
+
"epoch": 13.421052631578947,
|
| 1078 |
+
"grad_norm": 0.08417954295873642,
|
| 1079 |
+
"learning_rate": 5.851674641148326e-06,
|
| 1080 |
+
"loss": 0.0165,
|
| 1081 |
+
"step": 3825
|
| 1082 |
+
},
|
| 1083 |
+
{
|
| 1084 |
+
"epoch": 13.508771929824562,
|
| 1085 |
+
"grad_norm": 0.0910855233669281,
|
| 1086 |
+
"learning_rate": 5.791866028708134e-06,
|
| 1087 |
+
"loss": 0.1326,
|
| 1088 |
+
"step": 3850
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 13.596491228070175,
|
| 1092 |
+
"grad_norm": 65.31088256835938,
|
| 1093 |
+
"learning_rate": 5.732057416267942e-06,
|
| 1094 |
+
"loss": 0.0365,
|
| 1095 |
+
"step": 3875
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 13.68421052631579,
|
| 1099 |
+
"grad_norm": 0.09439278393983841,
|
| 1100 |
+
"learning_rate": 5.672248803827751e-06,
|
| 1101 |
+
"loss": 0.0784,
|
| 1102 |
+
"step": 3900
|
| 1103 |
+
},
|
| 1104 |
+
{
|
| 1105 |
+
"epoch": 13.771929824561404,
|
| 1106 |
+
"grad_norm": 0.17843316495418549,
|
| 1107 |
+
"learning_rate": 5.6124401913875595e-06,
|
| 1108 |
+
"loss": 0.0232,
|
| 1109 |
+
"step": 3925
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 13.859649122807017,
|
| 1113 |
+
"grad_norm": 0.16249841451644897,
|
| 1114 |
+
"learning_rate": 5.552631578947368e-06,
|
| 1115 |
+
"loss": 0.0107,
|
| 1116 |
+
"step": 3950
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 13.947368421052632,
|
| 1120 |
+
"grad_norm": 0.545814573764801,
|
| 1121 |
+
"learning_rate": 5.492822966507177e-06,
|
| 1122 |
+
"loss": 0.0947,
|
| 1123 |
+
"step": 3975
|
| 1124 |
+
},
|
| 1125 |
+
{
|
| 1126 |
+
"epoch": 14.035087719298245,
|
| 1127 |
+
"grad_norm": 0.7138233184814453,
|
| 1128 |
+
"learning_rate": 5.433014354066986e-06,
|
| 1129 |
+
"loss": 0.1812,
|
| 1130 |
+
"step": 4000
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 14.12280701754386,
|
| 1134 |
+
"grad_norm": 381.4528503417969,
|
| 1135 |
+
"learning_rate": 5.373205741626794e-06,
|
| 1136 |
+
"loss": 0.0399,
|
| 1137 |
+
"step": 4025
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 14.210526315789474,
|
| 1141 |
+
"grad_norm": 0.053043171763420105,
|
| 1142 |
+
"learning_rate": 5.313397129186603e-06,
|
| 1143 |
+
"loss": 0.1249,
|
| 1144 |
+
"step": 4050
|
| 1145 |
+
},
|
| 1146 |
+
{
|
| 1147 |
+
"epoch": 14.298245614035087,
|
| 1148 |
+
"grad_norm": 1.4563120603561401,
|
| 1149 |
+
"learning_rate": 5.253588516746412e-06,
|
| 1150 |
+
"loss": 0.0361,
|
| 1151 |
+
"step": 4075
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 14.385964912280702,
|
| 1155 |
+
"grad_norm": 0.06027642637491226,
|
| 1156 |
+
"learning_rate": 5.19377990430622e-06,
|
| 1157 |
+
"loss": 0.0799,
|
| 1158 |
+
"step": 4100
|
| 1159 |
+
},
|
| 1160 |
+
{
|
| 1161 |
+
"epoch": 14.473684210526315,
|
| 1162 |
+
"grad_norm": 0.11955183744430542,
|
| 1163 |
+
"learning_rate": 5.133971291866029e-06,
|
| 1164 |
+
"loss": 0.2006,
|
| 1165 |
+
"step": 4125
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 14.56140350877193,
|
| 1169 |
+
"grad_norm": 0.07738160341978073,
|
| 1170 |
+
"learning_rate": 5.074162679425838e-06,
|
| 1171 |
+
"loss": 0.16,
|
| 1172 |
+
"step": 4150
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 14.649122807017545,
|
| 1176 |
+
"grad_norm": 28.58780860900879,
|
| 1177 |
+
"learning_rate": 5.014354066985646e-06,
|
| 1178 |
+
"loss": 0.0675,
|
| 1179 |
+
"step": 4175
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 14.736842105263158,
|
| 1183 |
+
"grad_norm": 0.08654920011758804,
|
| 1184 |
+
"learning_rate": 4.954545454545455e-06,
|
| 1185 |
+
"loss": 0.0584,
|
| 1186 |
+
"step": 4200
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 14.824561403508772,
|
| 1190 |
+
"grad_norm": 0.1421671062707901,
|
| 1191 |
+
"learning_rate": 4.894736842105264e-06,
|
| 1192 |
+
"loss": 0.0049,
|
| 1193 |
+
"step": 4225
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 14.912280701754385,
|
| 1197 |
+
"grad_norm": 0.23447291553020477,
|
| 1198 |
+
"learning_rate": 4.834928229665072e-06,
|
| 1199 |
+
"loss": 0.0122,
|
| 1200 |
+
"step": 4250
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 15.0,
|
| 1204 |
+
"grad_norm": 54.224788665771484,
|
| 1205 |
+
"learning_rate": 4.775119617224881e-06,
|
| 1206 |
+
"loss": 0.0921,
|
| 1207 |
+
"step": 4275
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 15.087719298245615,
|
| 1211 |
+
"grad_norm": 0.08427131175994873,
|
| 1212 |
+
"learning_rate": 4.715311004784689e-06,
|
| 1213 |
+
"loss": 0.0061,
|
| 1214 |
+
"step": 4300
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 15.175438596491228,
|
| 1218 |
+
"grad_norm": 0.232700914144516,
|
| 1219 |
+
"learning_rate": 4.6555023923444975e-06,
|
| 1220 |
+
"loss": 0.1011,
|
| 1221 |
+
"step": 4325
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 15.263157894736842,
|
| 1225 |
+
"grad_norm": 0.09075941890478134,
|
| 1226 |
+
"learning_rate": 4.595693779904306e-06,
|
| 1227 |
+
"loss": 0.0419,
|
| 1228 |
+
"step": 4350
|
| 1229 |
+
},
|
| 1230 |
+
{
|
| 1231 |
+
"epoch": 15.350877192982455,
|
| 1232 |
+
"grad_norm": 0.20714852213859558,
|
| 1233 |
+
"learning_rate": 4.535885167464115e-06,
|
| 1234 |
+
"loss": 0.1786,
|
| 1235 |
+
"step": 4375
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 15.43859649122807,
|
| 1239 |
+
"grad_norm": 0.07586350291967392,
|
| 1240 |
+
"learning_rate": 4.476076555023924e-06,
|
| 1241 |
+
"loss": 0.0314,
|
| 1242 |
+
"step": 4400
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 15.526315789473685,
|
| 1246 |
+
"grad_norm": 0.29287371039390564,
|
| 1247 |
+
"learning_rate": 4.416267942583732e-06,
|
| 1248 |
+
"loss": 0.0635,
|
| 1249 |
+
"step": 4425
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 15.614035087719298,
|
| 1253 |
+
"grad_norm": 50.64339065551758,
|
| 1254 |
+
"learning_rate": 4.356459330143541e-06,
|
| 1255 |
+
"loss": 0.0854,
|
| 1256 |
+
"step": 4450
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 15.701754385964913,
|
| 1260 |
+
"grad_norm": 0.12661752104759216,
|
| 1261 |
+
"learning_rate": 4.29665071770335e-06,
|
| 1262 |
+
"loss": 0.1378,
|
| 1263 |
+
"step": 4475
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 15.789473684210526,
|
| 1267 |
+
"grad_norm": 2.418790340423584,
|
| 1268 |
+
"learning_rate": 4.2368421052631575e-06,
|
| 1269 |
+
"loss": 0.0325,
|
| 1270 |
+
"step": 4500
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 15.87719298245614,
|
| 1274 |
+
"grad_norm": 231.30796813964844,
|
| 1275 |
+
"learning_rate": 4.177033492822966e-06,
|
| 1276 |
+
"loss": 0.1398,
|
| 1277 |
+
"step": 4525
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 15.964912280701755,
|
| 1281 |
+
"grad_norm": 0.23073101043701172,
|
| 1282 |
+
"learning_rate": 4.117224880382775e-06,
|
| 1283 |
+
"loss": 0.0237,
|
| 1284 |
+
"step": 4550
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 16.05263157894737,
|
| 1288 |
+
"grad_norm": 0.044060900807380676,
|
| 1289 |
+
"learning_rate": 4.0574162679425835e-06,
|
| 1290 |
+
"loss": 0.0074,
|
| 1291 |
+
"step": 4575
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 16.140350877192983,
|
| 1295 |
+
"grad_norm": 1.4264813661575317,
|
| 1296 |
+
"learning_rate": 3.997607655502392e-06,
|
| 1297 |
+
"loss": 0.025,
|
| 1298 |
+
"step": 4600
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 16.228070175438596,
|
| 1302 |
+
"grad_norm": 0.41092243790626526,
|
| 1303 |
+
"learning_rate": 3.937799043062201e-06,
|
| 1304 |
+
"loss": 0.0049,
|
| 1305 |
+
"step": 4625
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 16.31578947368421,
|
| 1309 |
+
"grad_norm": 3.6420750617980957,
|
| 1310 |
+
"learning_rate": 3.8779904306220095e-06,
|
| 1311 |
+
"loss": 0.0491,
|
| 1312 |
+
"step": 4650
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"epoch": 16.403508771929825,
|
| 1316 |
+
"grad_norm": 0.08417033404111862,
|
| 1317 |
+
"learning_rate": 3.818181818181818e-06,
|
| 1318 |
+
"loss": 0.003,
|
| 1319 |
+
"step": 4675
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 16.49122807017544,
|
| 1323 |
+
"grad_norm": 0.41752633452415466,
|
| 1324 |
+
"learning_rate": 3.7583732057416273e-06,
|
| 1325 |
+
"loss": 0.0649,
|
| 1326 |
+
"step": 4700
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 16.57894736842105,
|
| 1330 |
+
"grad_norm": 0.04892294108867645,
|
| 1331 |
+
"learning_rate": 3.6985645933014356e-06,
|
| 1332 |
+
"loss": 0.1577,
|
| 1333 |
+
"step": 4725
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 16.666666666666668,
|
| 1337 |
+
"grad_norm": 0.07886885106563568,
|
| 1338 |
+
"learning_rate": 3.6387559808612442e-06,
|
| 1339 |
+
"loss": 0.1133,
|
| 1340 |
+
"step": 4750
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 16.75438596491228,
|
| 1344 |
+
"grad_norm": 202.08302307128906,
|
| 1345 |
+
"learning_rate": 3.5789473684210525e-06,
|
| 1346 |
+
"loss": 0.0151,
|
| 1347 |
+
"step": 4775
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 16.842105263157894,
|
| 1351 |
+
"grad_norm": 77.71251678466797,
|
| 1352 |
+
"learning_rate": 3.519138755980861e-06,
|
| 1353 |
+
"loss": 0.0644,
|
| 1354 |
+
"step": 4800
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 16.92982456140351,
|
| 1358 |
+
"grad_norm": 0.04525836929678917,
|
| 1359 |
+
"learning_rate": 3.45933014354067e-06,
|
| 1360 |
+
"loss": 0.0283,
|
| 1361 |
+
"step": 4825
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 17.017543859649123,
|
| 1365 |
+
"grad_norm": 0.035384029150009155,
|
| 1366 |
+
"learning_rate": 3.399521531100479e-06,
|
| 1367 |
+
"loss": 0.0431,
|
| 1368 |
+
"step": 4850
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 17.105263157894736,
|
| 1372 |
+
"grad_norm": 0.05048861354589462,
|
| 1373 |
+
"learning_rate": 3.339712918660287e-06,
|
| 1374 |
+
"loss": 0.0358,
|
| 1375 |
+
"step": 4875
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 17.19298245614035,
|
| 1379 |
+
"grad_norm": 0.1347443163394928,
|
| 1380 |
+
"learning_rate": 3.279904306220096e-06,
|
| 1381 |
+
"loss": 0.0206,
|
| 1382 |
+
"step": 4900
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 17.280701754385966,
|
| 1386 |
+
"grad_norm": 252.46458435058594,
|
| 1387 |
+
"learning_rate": 3.2200956937799046e-06,
|
| 1388 |
+
"loss": 0.0804,
|
| 1389 |
+
"step": 4925
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 17.36842105263158,
|
| 1393 |
+
"grad_norm": 0.05461292341351509,
|
| 1394 |
+
"learning_rate": 3.1602870813397132e-06,
|
| 1395 |
+
"loss": 0.0784,
|
| 1396 |
+
"step": 4950
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 17.45614035087719,
|
| 1400 |
+
"grad_norm": 2.0001866817474365,
|
| 1401 |
+
"learning_rate": 3.1004784688995215e-06,
|
| 1402 |
+
"loss": 0.0947,
|
| 1403 |
+
"step": 4975
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 17.54385964912281,
|
| 1407 |
+
"grad_norm": 0.04590539261698723,
|
| 1408 |
+
"learning_rate": 3.04066985645933e-06,
|
| 1409 |
+
"loss": 0.075,
|
| 1410 |
+
"step": 5000
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 17.63157894736842,
|
| 1414 |
+
"grad_norm": 0.04810748249292374,
|
| 1415 |
+
"learning_rate": 2.980861244019139e-06,
|
| 1416 |
+
"loss": 0.0342,
|
| 1417 |
+
"step": 5025
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 17.719298245614034,
|
| 1421 |
+
"grad_norm": 0.8750467896461487,
|
| 1422 |
+
"learning_rate": 2.9210526315789475e-06,
|
| 1423 |
+
"loss": 0.0618,
|
| 1424 |
+
"step": 5050
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 17.80701754385965,
|
| 1428 |
+
"grad_norm": 56.14554977416992,
|
| 1429 |
+
"learning_rate": 2.8612440191387558e-06,
|
| 1430 |
+
"loss": 0.026,
|
| 1431 |
+
"step": 5075
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 17.894736842105264,
|
| 1435 |
+
"grad_norm": 0.8380083441734314,
|
| 1436 |
+
"learning_rate": 2.8014354066985645e-06,
|
| 1437 |
+
"loss": 0.0135,
|
| 1438 |
+
"step": 5100
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 17.982456140350877,
|
| 1442 |
+
"grad_norm": 0.08012706786394119,
|
| 1443 |
+
"learning_rate": 2.741626794258373e-06,
|
| 1444 |
+
"loss": 0.0586,
|
| 1445 |
+
"step": 5125
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 18.07017543859649,
|
| 1449 |
+
"grad_norm": 1.2549360990524292,
|
| 1450 |
+
"learning_rate": 2.6818181818181822e-06,
|
| 1451 |
+
"loss": 0.0047,
|
| 1452 |
+
"step": 5150
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"epoch": 18.157894736842106,
|
| 1456 |
+
"grad_norm": 0.03993632644414902,
|
| 1457 |
+
"learning_rate": 2.6220095693779905e-06,
|
| 1458 |
+
"loss": 0.0869,
|
| 1459 |
+
"step": 5175
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 18.24561403508772,
|
| 1463 |
+
"grad_norm": 62.21980285644531,
|
| 1464 |
+
"learning_rate": 2.562200956937799e-06,
|
| 1465 |
+
"loss": 0.0987,
|
| 1466 |
+
"step": 5200
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 18.333333333333332,
|
| 1470 |
+
"grad_norm": 0.10092537105083466,
|
| 1471 |
+
"learning_rate": 2.502392344497608e-06,
|
| 1472 |
+
"loss": 0.1554,
|
| 1473 |
+
"step": 5225
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 18.42105263157895,
|
| 1477 |
+
"grad_norm": 0.7630422115325928,
|
| 1478 |
+
"learning_rate": 2.4425837320574165e-06,
|
| 1479 |
+
"loss": 0.0522,
|
| 1480 |
+
"step": 5250
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 18.50877192982456,
|
| 1484 |
+
"grad_norm": 62.42081069946289,
|
| 1485 |
+
"learning_rate": 2.382775119617225e-06,
|
| 1486 |
+
"loss": 0.0298,
|
| 1487 |
+
"step": 5275
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 18.596491228070175,
|
| 1491 |
+
"grad_norm": 0.046070296317338943,
|
| 1492 |
+
"learning_rate": 2.3229665071770335e-06,
|
| 1493 |
+
"loss": 0.1207,
|
| 1494 |
+
"step": 5300
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 18.68421052631579,
|
| 1498 |
+
"grad_norm": 0.04533836990594864,
|
| 1499 |
+
"learning_rate": 2.263157894736842e-06,
|
| 1500 |
+
"loss": 0.146,
|
| 1501 |
+
"step": 5325
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 18.771929824561404,
|
| 1505 |
+
"grad_norm": 0.03193026781082153,
|
| 1506 |
+
"learning_rate": 2.203349282296651e-06,
|
| 1507 |
+
"loss": 0.0106,
|
| 1508 |
+
"step": 5350
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 18.859649122807017,
|
| 1512 |
+
"grad_norm": 0.03594537451863289,
|
| 1513 |
+
"learning_rate": 2.143540669856459e-06,
|
| 1514 |
+
"loss": 0.0266,
|
| 1515 |
+
"step": 5375
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 18.94736842105263,
|
| 1519 |
+
"grad_norm": 0.07382319122552872,
|
| 1520 |
+
"learning_rate": 2.0837320574162678e-06,
|
| 1521 |
+
"loss": 0.0954,
|
| 1522 |
+
"step": 5400
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 19.035087719298247,
|
| 1526 |
+
"grad_norm": 0.0792032852768898,
|
| 1527 |
+
"learning_rate": 2.0239234449760764e-06,
|
| 1528 |
+
"loss": 0.0871,
|
| 1529 |
+
"step": 5425
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 19.12280701754386,
|
| 1533 |
+
"grad_norm": 0.06994396448135376,
|
| 1534 |
+
"learning_rate": 1.9641148325358855e-06,
|
| 1535 |
+
"loss": 0.0092,
|
| 1536 |
+
"step": 5450
|
| 1537 |
+
},
|
| 1538 |
+
{
|
| 1539 |
+
"epoch": 19.210526315789473,
|
| 1540 |
+
"grad_norm": 36.98837661743164,
|
| 1541 |
+
"learning_rate": 1.904306220095694e-06,
|
| 1542 |
+
"loss": 0.1442,
|
| 1543 |
+
"step": 5475
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"epoch": 19.29824561403509,
|
| 1547 |
+
"grad_norm": 0.11950813978910446,
|
| 1548 |
+
"learning_rate": 1.8444976076555023e-06,
|
| 1549 |
+
"loss": 0.0344,
|
| 1550 |
+
"step": 5500
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 19.385964912280702,
|
| 1554 |
+
"grad_norm": 0.053549814969301224,
|
| 1555 |
+
"learning_rate": 1.7846889952153112e-06,
|
| 1556 |
+
"loss": 0.0609,
|
| 1557 |
+
"step": 5525
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 19.473684210526315,
|
| 1561 |
+
"grad_norm": 0.05915123224258423,
|
| 1562 |
+
"learning_rate": 1.7248803827751196e-06,
|
| 1563 |
+
"loss": 0.0212,
|
| 1564 |
+
"step": 5550
|
| 1565 |
+
},
|
| 1566 |
+
{
|
| 1567 |
+
"epoch": 19.56140350877193,
|
| 1568 |
+
"grad_norm": 0.03509140759706497,
|
| 1569 |
+
"learning_rate": 1.6650717703349283e-06,
|
| 1570 |
+
"loss": 0.0019,
|
| 1571 |
+
"step": 5575
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 19.649122807017545,
|
| 1575 |
+
"grad_norm": 0.04774395003914833,
|
| 1576 |
+
"learning_rate": 1.605263157894737e-06,
|
| 1577 |
+
"loss": 0.032,
|
| 1578 |
+
"step": 5600
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 19.736842105263158,
|
| 1582 |
+
"grad_norm": 0.1909104436635971,
|
| 1583 |
+
"learning_rate": 1.5454545454545454e-06,
|
| 1584 |
+
"loss": 0.0351,
|
| 1585 |
+
"step": 5625
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"epoch": 19.82456140350877,
|
| 1589 |
+
"grad_norm": 0.11199568212032318,
|
| 1590 |
+
"learning_rate": 1.4856459330143541e-06,
|
| 1591 |
+
"loss": 0.0268,
|
| 1592 |
+
"step": 5650
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 19.912280701754387,
|
| 1596 |
+
"grad_norm": 17.852022171020508,
|
| 1597 |
+
"learning_rate": 1.4258373205741628e-06,
|
| 1598 |
+
"loss": 0.0397,
|
| 1599 |
+
"step": 5675
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 20.0,
|
| 1603 |
+
"grad_norm": 0.18822583556175232,
|
| 1604 |
+
"learning_rate": 1.3660287081339715e-06,
|
| 1605 |
+
"loss": 0.0541,
|
| 1606 |
+
"step": 5700
|
| 1607 |
+
},
|
| 1608 |
+
{
|
| 1609 |
+
"epoch": 20.087719298245613,
|
| 1610 |
+
"grad_norm": 50.338890075683594,
|
| 1611 |
+
"learning_rate": 1.30622009569378e-06,
|
| 1612 |
+
"loss": 0.016,
|
| 1613 |
+
"step": 5725
|
| 1614 |
+
},
|
| 1615 |
+
{
|
| 1616 |
+
"epoch": 20.17543859649123,
|
| 1617 |
+
"grad_norm": 0.05340347811579704,
|
| 1618 |
+
"learning_rate": 1.2464114832535886e-06,
|
| 1619 |
+
"loss": 0.0084,
|
| 1620 |
+
"step": 5750
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 20.263157894736842,
|
| 1624 |
+
"grad_norm": 0.04629586637020111,
|
| 1625 |
+
"learning_rate": 1.186602870813397e-06,
|
| 1626 |
+
"loss": 0.0356,
|
| 1627 |
+
"step": 5775
|
| 1628 |
+
},
|
| 1629 |
+
{
|
| 1630 |
+
"epoch": 20.350877192982455,
|
| 1631 |
+
"grad_norm": 0.9073031544685364,
|
| 1632 |
+
"learning_rate": 1.1267942583732058e-06,
|
| 1633 |
+
"loss": 0.0786,
|
| 1634 |
+
"step": 5800
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 20.43859649122807,
|
| 1638 |
+
"grad_norm": 10.095131874084473,
|
| 1639 |
+
"learning_rate": 1.0669856459330144e-06,
|
| 1640 |
+
"loss": 0.0419,
|
| 1641 |
+
"step": 5825
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 20.526315789473685,
|
| 1645 |
+
"grad_norm": 0.16011740267276764,
|
| 1646 |
+
"learning_rate": 1.0071770334928231e-06,
|
| 1647 |
+
"loss": 0.0653,
|
| 1648 |
+
"step": 5850
|
| 1649 |
+
},
|
| 1650 |
+
{
|
| 1651 |
+
"epoch": 20.614035087719298,
|
| 1652 |
+
"grad_norm": 0.18243364989757538,
|
| 1653 |
+
"learning_rate": 9.473684210526316e-07,
|
| 1654 |
+
"loss": 0.0015,
|
| 1655 |
+
"step": 5875
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 20.70175438596491,
|
| 1659 |
+
"grad_norm": 0.14108963310718536,
|
| 1660 |
+
"learning_rate": 8.875598086124402e-07,
|
| 1661 |
+
"loss": 0.0817,
|
| 1662 |
+
"step": 5900
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 20.789473684210527,
|
| 1666 |
+
"grad_norm": 0.03629032149910927,
|
| 1667 |
+
"learning_rate": 8.277511961722487e-07,
|
| 1668 |
+
"loss": 0.002,
|
| 1669 |
+
"step": 5925
|
| 1670 |
+
},
|
| 1671 |
+
{
|
| 1672 |
+
"epoch": 20.87719298245614,
|
| 1673 |
+
"grad_norm": 21.195932388305664,
|
| 1674 |
+
"learning_rate": 7.679425837320574e-07,
|
| 1675 |
+
"loss": 0.1265,
|
| 1676 |
+
"step": 5950
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 20.964912280701753,
|
| 1680 |
+
"grad_norm": 0.036704275757074356,
|
| 1681 |
+
"learning_rate": 7.081339712918661e-07,
|
| 1682 |
+
"loss": 0.0369,
|
| 1683 |
+
"step": 5975
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 21.05263157894737,
|
| 1687 |
+
"grad_norm": 327.310791015625,
|
| 1688 |
+
"learning_rate": 6.483253588516747e-07,
|
| 1689 |
+
"loss": 0.0288,
|
| 1690 |
+
"step": 6000
|
| 1691 |
+
},
|
| 1692 |
+
{
|
| 1693 |
+
"epoch": 21.140350877192983,
|
| 1694 |
+
"grad_norm": 0.17735086381435394,
|
| 1695 |
+
"learning_rate": 5.885167464114833e-07,
|
| 1696 |
+
"loss": 0.0805,
|
| 1697 |
+
"step": 6025
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 21.228070175438596,
|
| 1701 |
+
"grad_norm": 1.9311668872833252,
|
| 1702 |
+
"learning_rate": 5.287081339712919e-07,
|
| 1703 |
+
"loss": 0.012,
|
| 1704 |
+
"step": 6050
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 21.31578947368421,
|
| 1708 |
+
"grad_norm": 0.029030485078692436,
|
| 1709 |
+
"learning_rate": 4.6889952153110044e-07,
|
| 1710 |
+
"loss": 0.0026,
|
| 1711 |
+
"step": 6075
|
| 1712 |
+
},
|
| 1713 |
+
{
|
| 1714 |
+
"epoch": 21.403508771929825,
|
| 1715 |
+
"grad_norm": 0.030113637447357178,
|
| 1716 |
+
"learning_rate": 4.0909090909090906e-07,
|
| 1717 |
+
"loss": 0.0032,
|
| 1718 |
+
"step": 6100
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 21.49122807017544,
|
| 1722 |
+
"grad_norm": 0.07308195531368256,
|
| 1723 |
+
"learning_rate": 3.492822966507177e-07,
|
| 1724 |
+
"loss": 0.0496,
|
| 1725 |
+
"step": 6125
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 21.57894736842105,
|
| 1729 |
+
"grad_norm": 0.25214284658432007,
|
| 1730 |
+
"learning_rate": 2.894736842105263e-07,
|
| 1731 |
+
"loss": 0.0023,
|
| 1732 |
+
"step": 6150
|
| 1733 |
+
},
|
| 1734 |
+
{
|
| 1735 |
+
"epoch": 21.666666666666668,
|
| 1736 |
+
"grad_norm": 1.1200495958328247,
|
| 1737 |
+
"learning_rate": 2.2966507177033494e-07,
|
| 1738 |
+
"loss": 0.0198,
|
| 1739 |
+
"step": 6175
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 21.75438596491228,
|
| 1743 |
+
"grad_norm": 0.11758319288492203,
|
| 1744 |
+
"learning_rate": 1.6985645933014354e-07,
|
| 1745 |
+
"loss": 0.0227,
|
| 1746 |
+
"step": 6200
|
| 1747 |
+
},
|
| 1748 |
+
{
|
| 1749 |
+
"epoch": 21.842105263157894,
|
| 1750 |
+
"grad_norm": 0.04569339007139206,
|
| 1751 |
+
"learning_rate": 1.1004784688995216e-07,
|
| 1752 |
+
"loss": 0.0989,
|
| 1753 |
+
"step": 6225
|
| 1754 |
+
},
|
| 1755 |
+
{
|
| 1756 |
+
"epoch": 21.92982456140351,
|
| 1757 |
+
"grad_norm": 0.1228533387184143,
|
| 1758 |
+
"learning_rate": 5.023923444976077e-08,
|
| 1759 |
+
"loss": 0.1324,
|
| 1760 |
+
"step": 6250
|
| 1761 |
+
}
|
| 1762 |
+
],
|
| 1763 |
+
"logging_steps": 25,
|
| 1764 |
+
"max_steps": 6270,
|
| 1765 |
+
"num_input_tokens_seen": 0,
|
| 1766 |
+
"num_train_epochs": 22,
|
| 1767 |
+
"save_steps": 500,
|
| 1768 |
+
"stateful_callbacks": {
|
| 1769 |
+
"TrainerControl": {
|
| 1770 |
+
"args": {
|
| 1771 |
+
"should_epoch_stop": false,
|
| 1772 |
+
"should_evaluate": false,
|
| 1773 |
+
"should_log": false,
|
| 1774 |
+
"should_save": true,
|
| 1775 |
+
"should_training_stop": true
|
| 1776 |
+
},
|
| 1777 |
+
"attributes": {}
|
| 1778 |
+
}
|
| 1779 |
+
},
|
| 1780 |
+
"total_flos": 1.690990233877171e+18,
|
| 1781 |
+
"train_batch_size": 4,
|
| 1782 |
+
"trial_name": null,
|
| 1783 |
+
"trial_params": null
|
| 1784 |
+
}
|
checkpoint-6270/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:21cda56f26874205e02c84553dd1454d4b3eefd72224b421d8d696952554b651
|
| 3 |
+
size 5841
|
config.json
ADDED
|
@@ -0,0 +1,139 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"activation_dropout": 0.0,
|
| 3 |
+
"adapter_attn_dim": null,
|
| 4 |
+
"adapter_kernel_size": 3,
|
| 5 |
+
"adapter_stride": 2,
|
| 6 |
+
"add_adapter": false,
|
| 7 |
+
"apply_spec_augment": true,
|
| 8 |
+
"architectures": [
|
| 9 |
+
"Wav2Vec2ForSequenceClassification"
|
| 10 |
+
],
|
| 11 |
+
"attention_dropout": 0.1,
|
| 12 |
+
"bos_token_id": 1,
|
| 13 |
+
"classifier_proj_size": 256,
|
| 14 |
+
"codevector_dim": 256,
|
| 15 |
+
"contrastive_logits_temperature": 0.1,
|
| 16 |
+
"conv_bias": false,
|
| 17 |
+
"conv_dim": [
|
| 18 |
+
512,
|
| 19 |
+
512,
|
| 20 |
+
512,
|
| 21 |
+
512,
|
| 22 |
+
512,
|
| 23 |
+
512,
|
| 24 |
+
512
|
| 25 |
+
],
|
| 26 |
+
"conv_kernel": [
|
| 27 |
+
10,
|
| 28 |
+
3,
|
| 29 |
+
3,
|
| 30 |
+
3,
|
| 31 |
+
3,
|
| 32 |
+
2,
|
| 33 |
+
2
|
| 34 |
+
],
|
| 35 |
+
"conv_stride": [
|
| 36 |
+
5,
|
| 37 |
+
2,
|
| 38 |
+
2,
|
| 39 |
+
2,
|
| 40 |
+
2,
|
| 41 |
+
2,
|
| 42 |
+
2
|
| 43 |
+
],
|
| 44 |
+
"ctc_loss_reduction": "sum",
|
| 45 |
+
"ctc_zero_infinity": false,
|
| 46 |
+
"diversity_loss_weight": 0.1,
|
| 47 |
+
"do_stable_layer_norm": false,
|
| 48 |
+
"dtype": "float32",
|
| 49 |
+
"eos_token_id": 2,
|
| 50 |
+
"feat_extract_activation": "gelu",
|
| 51 |
+
"feat_extract_norm": "group",
|
| 52 |
+
"feat_proj_dropout": 0.1,
|
| 53 |
+
"feat_quantizer_dropout": 0.0,
|
| 54 |
+
"final_dropout": 0.0,
|
| 55 |
+
"finetuning_task": "wav2vec2_emotion",
|
| 56 |
+
"freeze_feat_extract_train": true,
|
| 57 |
+
"hidden_act": "gelu",
|
| 58 |
+
"hidden_dropout": 0.1,
|
| 59 |
+
"hidden_size": 768,
|
| 60 |
+
"id2label": {
|
| 61 |
+
"0": "neutral",
|
| 62 |
+
"1": "calm",
|
| 63 |
+
"2": "happy",
|
| 64 |
+
"3": "sad",
|
| 65 |
+
"4": "angry",
|
| 66 |
+
"5": "fearful",
|
| 67 |
+
"6": "disgust",
|
| 68 |
+
"7": "surprised"
|
| 69 |
+
},
|
| 70 |
+
"initializer_range": 0.02,
|
| 71 |
+
"intermediate_size": 3072,
|
| 72 |
+
"label2id": {
|
| 73 |
+
"angry": 4,
|
| 74 |
+
"calm": 1,
|
| 75 |
+
"disgust": 6,
|
| 76 |
+
"fearful": 5,
|
| 77 |
+
"happy": 2,
|
| 78 |
+
"neutral": 0,
|
| 79 |
+
"sad": 3,
|
| 80 |
+
"surprised": 7
|
| 81 |
+
},
|
| 82 |
+
"layer_norm_eps": 1e-05,
|
| 83 |
+
"layerdrop": 0.05,
|
| 84 |
+
"mask_channel_length": 10,
|
| 85 |
+
"mask_channel_min_space": 1,
|
| 86 |
+
"mask_channel_other": 0.0,
|
| 87 |
+
"mask_channel_prob": 0.0,
|
| 88 |
+
"mask_channel_selection": "static",
|
| 89 |
+
"mask_feature_length": 10,
|
| 90 |
+
"mask_feature_min_masks": 0,
|
| 91 |
+
"mask_feature_prob": 0.0,
|
| 92 |
+
"mask_time_length": 10,
|
| 93 |
+
"mask_time_min_masks": 2,
|
| 94 |
+
"mask_time_min_space": 1,
|
| 95 |
+
"mask_time_other": 0.0,
|
| 96 |
+
"mask_time_prob": 0.05,
|
| 97 |
+
"mask_time_selection": "static",
|
| 98 |
+
"model_type": "wav2vec2",
|
| 99 |
+
"no_mask_channel_overlap": false,
|
| 100 |
+
"no_mask_time_overlap": false,
|
| 101 |
+
"num_adapter_layers": 3,
|
| 102 |
+
"num_attention_heads": 12,
|
| 103 |
+
"num_codevector_groups": 2,
|
| 104 |
+
"num_codevectors_per_group": 320,
|
| 105 |
+
"num_conv_pos_embedding_groups": 16,
|
| 106 |
+
"num_conv_pos_embeddings": 128,
|
| 107 |
+
"num_feat_extract_layers": 7,
|
| 108 |
+
"num_hidden_layers": 12,
|
| 109 |
+
"num_negatives": 100,
|
| 110 |
+
"output_hidden_size": 768,
|
| 111 |
+
"pad_token_id": 0,
|
| 112 |
+
"proj_codevector_dim": 256,
|
| 113 |
+
"tdnn_dilation": [
|
| 114 |
+
1,
|
| 115 |
+
2,
|
| 116 |
+
3,
|
| 117 |
+
1,
|
| 118 |
+
1
|
| 119 |
+
],
|
| 120 |
+
"tdnn_dim": [
|
| 121 |
+
512,
|
| 122 |
+
512,
|
| 123 |
+
512,
|
| 124 |
+
512,
|
| 125 |
+
1500
|
| 126 |
+
],
|
| 127 |
+
"tdnn_kernel": [
|
| 128 |
+
5,
|
| 129 |
+
3,
|
| 130 |
+
3,
|
| 131 |
+
1,
|
| 132 |
+
1
|
| 133 |
+
],
|
| 134 |
+
"transformers_version": "4.57.1",
|
| 135 |
+
"use_cache": false,
|
| 136 |
+
"use_weighted_layer_sum": true,
|
| 137 |
+
"vocab_size": 32,
|
| 138 |
+
"xvector_output_dim": 512
|
| 139 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d01726ed5d89e30cdf84b5dbff1273c611514bc8962ba0dfa1601980938c63a6
|
| 3 |
+
size 378308660
|
preprocessor_config.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_normalize": false,
|
| 3 |
+
"feature_extractor_type": "Wav2Vec2FeatureExtractor",
|
| 4 |
+
"feature_size": 1,
|
| 5 |
+
"padding_side": "right",
|
| 6 |
+
"padding_value": 0.0,
|
| 7 |
+
"return_attention_mask": true,
|
| 8 |
+
"sampling_rate": 16000
|
| 9 |
+
}
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"eos_token": "</s>",
|
| 4 |
+
"pad_token": "<pad>",
|
| 5 |
+
"unk_token": "<unk>"
|
| 6 |
+
}
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<pad>",
|
| 5 |
+
"lstrip": true,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": true,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": false
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<s>",
|
| 13 |
+
"lstrip": true,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": true,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": false
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": true,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": true,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": false
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": true,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": true,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": false
|
| 34 |
+
}
|
| 35 |
+
},
|
| 36 |
+
"bos_token": "<s>",
|
| 37 |
+
"clean_up_tokenization_spaces": false,
|
| 38 |
+
"do_lower_case": false,
|
| 39 |
+
"do_normalize": true,
|
| 40 |
+
"eos_token": "</s>",
|
| 41 |
+
"extra_special_tokens": {},
|
| 42 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 43 |
+
"pad_token": "<pad>",
|
| 44 |
+
"processor_class": "Wav2Vec2Processor",
|
| 45 |
+
"replace_word_delimiter_char": " ",
|
| 46 |
+
"return_attention_mask": false,
|
| 47 |
+
"target_lang": null,
|
| 48 |
+
"tokenizer_class": "Wav2Vec2CTCTokenizer",
|
| 49 |
+
"unk_token": "<unk>",
|
| 50 |
+
"word_delimiter_token": "|"
|
| 51 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:21cda56f26874205e02c84553dd1454d4b3eefd72224b421d8d696952554b651
|
| 3 |
+
size 5841
|
vocab.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"'": 27,
|
| 3 |
+
"</s>": 2,
|
| 4 |
+
"<pad>": 0,
|
| 5 |
+
"<s>": 1,
|
| 6 |
+
"<unk>": 3,
|
| 7 |
+
"A": 7,
|
| 8 |
+
"B": 24,
|
| 9 |
+
"C": 19,
|
| 10 |
+
"D": 14,
|
| 11 |
+
"E": 5,
|
| 12 |
+
"F": 20,
|
| 13 |
+
"G": 21,
|
| 14 |
+
"H": 11,
|
| 15 |
+
"I": 10,
|
| 16 |
+
"J": 29,
|
| 17 |
+
"K": 26,
|
| 18 |
+
"L": 15,
|
| 19 |
+
"M": 17,
|
| 20 |
+
"N": 9,
|
| 21 |
+
"O": 8,
|
| 22 |
+
"P": 23,
|
| 23 |
+
"Q": 30,
|
| 24 |
+
"R": 13,
|
| 25 |
+
"S": 12,
|
| 26 |
+
"T": 6,
|
| 27 |
+
"U": 16,
|
| 28 |
+
"V": 25,
|
| 29 |
+
"W": 18,
|
| 30 |
+
"X": 28,
|
| 31 |
+
"Y": 22,
|
| 32 |
+
"Z": 31,
|
| 33 |
+
"|": 4
|
| 34 |
+
}
|