Upload folder using huggingface_hub
Browse files- hw1/results_paragraph/checkpoint-4343/config.json +34 -0
- hw1/results_paragraph/checkpoint-4343/model.safetensors +3 -0
- hw1/results_paragraph/checkpoint-4343/optimizer.pt +3 -0
- hw1/results_paragraph/checkpoint-4343/rng_state.pth +3 -0
- hw1/results_paragraph/checkpoint-4343/scheduler.pt +3 -0
- hw1/results_paragraph/checkpoint-4343/special_tokens_map.json +37 -0
- hw1/results_paragraph/checkpoint-4343/tokenizer.json +0 -0
- hw1/results_paragraph/checkpoint-4343/tokenizer_config.json +55 -0
- hw1/results_paragraph/checkpoint-4343/trainer_state.json +3077 -0
- hw1/results_paragraph/checkpoint-4343/training_args.bin +3 -0
- hw1/results_paragraph/checkpoint-4343/vocab.txt +0 -0
- hw1/results_paragraph/runs/Sep26_12-47-40_d88996431f93/events.out.tfevents.1727354862.d88996431f93.1905835.0 +3 -0
- hw1/results_paragraph/runs/Sep26_13-37-30_d88996431f93/events.out.tfevents.1727357852.d88996431f93.2093465.0 +3 -0
- hw1/results_span/checkpoint-10855/config.json +34 -0
- hw1/results_span/checkpoint-10855/model.safetensors +3 -0
- hw1/results_span/checkpoint-10855/optimizer.pt +3 -0
- hw1/results_span/checkpoint-10855/rng_state.pth +3 -0
- hw1/results_span/checkpoint-10855/scheduler.pt +3 -0
- hw1/results_span/checkpoint-10855/special_tokens_map.json +37 -0
- hw1/results_span/checkpoint-10855/tokenizer.json +0 -0
- hw1/results_span/checkpoint-10855/tokenizer_config.json +55 -0
- hw1/results_span/checkpoint-10855/trainer_state.json +0 -0
- hw1/results_span/checkpoint-10855/training_args.bin +3 -0
- hw1/results_span/checkpoint-10855/vocab.txt +0 -0
- hw1/results_span/runs/Sep30_07-11-20_d88996431f93/events.out.tfevents.1727680284.d88996431f93.2093465.1 +3 -0
- hw1/results_span/runs/Sep30_07-49-45_d88996431f93/events.out.tfevents.1727682589.d88996431f93.906292.0 +3 -0
hw1/results_paragraph/checkpoint-4343/config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "hfl/chinese-roberta-wwm-ext",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"BertForMultipleChoice"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"directionality": "bidi",
|
| 10 |
+
"eos_token_id": 2,
|
| 11 |
+
"hidden_act": "gelu",
|
| 12 |
+
"hidden_dropout_prob": 0.1,
|
| 13 |
+
"hidden_size": 768,
|
| 14 |
+
"initializer_range": 0.02,
|
| 15 |
+
"intermediate_size": 3072,
|
| 16 |
+
"layer_norm_eps": 1e-12,
|
| 17 |
+
"max_position_embeddings": 512,
|
| 18 |
+
"model_type": "bert",
|
| 19 |
+
"num_attention_heads": 12,
|
| 20 |
+
"num_hidden_layers": 12,
|
| 21 |
+
"output_past": true,
|
| 22 |
+
"pad_token_id": 0,
|
| 23 |
+
"pooler_fc_size": 768,
|
| 24 |
+
"pooler_num_attention_heads": 12,
|
| 25 |
+
"pooler_num_fc_layers": 3,
|
| 26 |
+
"pooler_size_per_head": 128,
|
| 27 |
+
"pooler_type": "first_token_transform",
|
| 28 |
+
"position_embedding_type": "absolute",
|
| 29 |
+
"torch_dtype": "float32",
|
| 30 |
+
"transformers_version": "4.39.0.dev0",
|
| 31 |
+
"type_vocab_size": 2,
|
| 32 |
+
"use_cache": true,
|
| 33 |
+
"vocab_size": 21128
|
| 34 |
+
}
|
hw1/results_paragraph/checkpoint-4343/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:17c9d6a348a406e6c7ed8c467d644ab96f5aa0dbe218de7356b0b9fc0c668e0e
|
| 3 |
+
size 409097164
|
hw1/results_paragraph/checkpoint-4343/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ddbef9f6c5dcd2daf26b4fbd7dec3ef911673b9dda39a1eda4d132b8ff215fb0
|
| 3 |
+
size 818315514
|
hw1/results_paragraph/checkpoint-4343/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d3b7a4f86beb75a75169f0d81ac9179af80390f4c150961b33a7c76e0e8d0777
|
| 3 |
+
size 14244
|
hw1/results_paragraph/checkpoint-4343/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b792b92266bb03c4c35b59a8c4acdfd232f54f711960fa00f13b6b294a336964
|
| 3 |
+
size 1064
|
hw1/results_paragraph/checkpoint-4343/special_tokens_map.json
ADDED
|
@@ -0,0 +1,37 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cls_token": {
|
| 3 |
+
"content": "[CLS]",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"mask_token": {
|
| 10 |
+
"content": "[MASK]",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"pad_token": {
|
| 17 |
+
"content": "[PAD]",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"sep_token": {
|
| 24 |
+
"content": "[SEP]",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"unk_token": {
|
| 31 |
+
"content": "[UNK]",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
}
|
| 37 |
+
}
|
hw1/results_paragraph/checkpoint-4343/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
hw1/results_paragraph/checkpoint-4343/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "[PAD]",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"100": {
|
| 12 |
+
"content": "[UNK]",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"101": {
|
| 20 |
+
"content": "[CLS]",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"102": {
|
| 28 |
+
"content": "[SEP]",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"103": {
|
| 36 |
+
"content": "[MASK]",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"clean_up_tokenization_spaces": true,
|
| 45 |
+
"cls_token": "[CLS]",
|
| 46 |
+
"do_lower_case": true,
|
| 47 |
+
"mask_token": "[MASK]",
|
| 48 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 49 |
+
"pad_token": "[PAD]",
|
| 50 |
+
"sep_token": "[SEP]",
|
| 51 |
+
"strip_accents": null,
|
| 52 |
+
"tokenize_chinese_chars": true,
|
| 53 |
+
"tokenizer_class": "BertTokenizer",
|
| 54 |
+
"unk_token": "[UNK]"
|
| 55 |
+
}
|
hw1/results_paragraph/checkpoint-4343/trainer_state.json
ADDED
|
@@ -0,0 +1,3077 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.965769358590894,
|
| 3 |
+
"best_model_checkpoint": "./results_paragraph/checkpoint-4343",
|
| 4 |
+
"epoch": 2.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 4343,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0,
|
| 13 |
+
"grad_norm": 5.42363977432251,
|
| 14 |
+
"learning_rate": 1.998157531091663e-05,
|
| 15 |
+
"loss": 1.2693,
|
| 16 |
+
"step": 10
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.01,
|
| 20 |
+
"grad_norm": 5.5741286277771,
|
| 21 |
+
"learning_rate": 1.996315062183326e-05,
|
| 22 |
+
"loss": 1.0585,
|
| 23 |
+
"step": 20
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.01,
|
| 27 |
+
"grad_norm": 8.118847846984863,
|
| 28 |
+
"learning_rate": 1.9944725932749887e-05,
|
| 29 |
+
"loss": 0.9733,
|
| 30 |
+
"step": 30
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.02,
|
| 34 |
+
"grad_norm": 5.396815299987793,
|
| 35 |
+
"learning_rate": 1.9926301243666515e-05,
|
| 36 |
+
"loss": 0.9898,
|
| 37 |
+
"step": 40
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.02,
|
| 41 |
+
"grad_norm": 6.525942325592041,
|
| 42 |
+
"learning_rate": 1.9907876554583144e-05,
|
| 43 |
+
"loss": 0.9492,
|
| 44 |
+
"step": 50
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.03,
|
| 48 |
+
"grad_norm": 6.160113334655762,
|
| 49 |
+
"learning_rate": 1.9889451865499772e-05,
|
| 50 |
+
"loss": 0.9158,
|
| 51 |
+
"step": 60
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.03,
|
| 55 |
+
"grad_norm": 6.827396392822266,
|
| 56 |
+
"learning_rate": 1.98710271764164e-05,
|
| 57 |
+
"loss": 0.9787,
|
| 58 |
+
"step": 70
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.04,
|
| 62 |
+
"grad_norm": 5.99915885925293,
|
| 63 |
+
"learning_rate": 1.985260248733303e-05,
|
| 64 |
+
"loss": 0.8594,
|
| 65 |
+
"step": 80
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.04,
|
| 69 |
+
"grad_norm": 5.186646938323975,
|
| 70 |
+
"learning_rate": 1.9834177798249658e-05,
|
| 71 |
+
"loss": 0.7502,
|
| 72 |
+
"step": 90
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.05,
|
| 76 |
+
"grad_norm": 15.453600883483887,
|
| 77 |
+
"learning_rate": 1.9815753109166286e-05,
|
| 78 |
+
"loss": 0.4096,
|
| 79 |
+
"step": 100
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.05,
|
| 83 |
+
"grad_norm": 5.537822246551514,
|
| 84 |
+
"learning_rate": 1.979732842008291e-05,
|
| 85 |
+
"loss": 0.3128,
|
| 86 |
+
"step": 110
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.06,
|
| 90 |
+
"grad_norm": 9.06698989868164,
|
| 91 |
+
"learning_rate": 1.977890373099954e-05,
|
| 92 |
+
"loss": 0.2671,
|
| 93 |
+
"step": 120
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.06,
|
| 97 |
+
"grad_norm": 12.036720275878906,
|
| 98 |
+
"learning_rate": 1.9760479041916168e-05,
|
| 99 |
+
"loss": 0.2578,
|
| 100 |
+
"step": 130
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.06,
|
| 104 |
+
"grad_norm": 3.8072216510772705,
|
| 105 |
+
"learning_rate": 1.9742054352832797e-05,
|
| 106 |
+
"loss": 0.1423,
|
| 107 |
+
"step": 140
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.07,
|
| 111 |
+
"grad_norm": 2.7156248092651367,
|
| 112 |
+
"learning_rate": 1.9723629663749425e-05,
|
| 113 |
+
"loss": 0.1837,
|
| 114 |
+
"step": 150
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.07,
|
| 118 |
+
"grad_norm": 10.041704177856445,
|
| 119 |
+
"learning_rate": 1.9705204974666054e-05,
|
| 120 |
+
"loss": 0.1242,
|
| 121 |
+
"step": 160
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.08,
|
| 125 |
+
"grad_norm": 6.0569353103637695,
|
| 126 |
+
"learning_rate": 1.9686780285582682e-05,
|
| 127 |
+
"loss": 0.2228,
|
| 128 |
+
"step": 170
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.08,
|
| 132 |
+
"grad_norm": 11.10827922821045,
|
| 133 |
+
"learning_rate": 1.966835559649931e-05,
|
| 134 |
+
"loss": 0.2136,
|
| 135 |
+
"step": 180
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.09,
|
| 139 |
+
"grad_norm": 6.920894145965576,
|
| 140 |
+
"learning_rate": 1.964993090741594e-05,
|
| 141 |
+
"loss": 0.1595,
|
| 142 |
+
"step": 190
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.09,
|
| 146 |
+
"grad_norm": 7.058196544647217,
|
| 147 |
+
"learning_rate": 1.9631506218332567e-05,
|
| 148 |
+
"loss": 0.1644,
|
| 149 |
+
"step": 200
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.1,
|
| 153 |
+
"grad_norm": 0.476233571767807,
|
| 154 |
+
"learning_rate": 1.9613081529249196e-05,
|
| 155 |
+
"loss": 0.0995,
|
| 156 |
+
"step": 210
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.1,
|
| 160 |
+
"grad_norm": 2.5886449813842773,
|
| 161 |
+
"learning_rate": 1.9594656840165824e-05,
|
| 162 |
+
"loss": 0.1714,
|
| 163 |
+
"step": 220
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.11,
|
| 167 |
+
"grad_norm": 3.322171211242676,
|
| 168 |
+
"learning_rate": 1.9576232151082453e-05,
|
| 169 |
+
"loss": 0.0998,
|
| 170 |
+
"step": 230
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.11,
|
| 174 |
+
"grad_norm": 6.558858871459961,
|
| 175 |
+
"learning_rate": 1.955780746199908e-05,
|
| 176 |
+
"loss": 0.1752,
|
| 177 |
+
"step": 240
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.12,
|
| 181 |
+
"grad_norm": 14.43525505065918,
|
| 182 |
+
"learning_rate": 1.953938277291571e-05,
|
| 183 |
+
"loss": 0.1083,
|
| 184 |
+
"step": 250
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.12,
|
| 188 |
+
"grad_norm": 29.391435623168945,
|
| 189 |
+
"learning_rate": 1.9520958083832338e-05,
|
| 190 |
+
"loss": 0.0928,
|
| 191 |
+
"step": 260
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.12,
|
| 195 |
+
"grad_norm": 16.467206954956055,
|
| 196 |
+
"learning_rate": 1.9502533394748967e-05,
|
| 197 |
+
"loss": 0.1184,
|
| 198 |
+
"step": 270
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.13,
|
| 202 |
+
"grad_norm": 0.2154889702796936,
|
| 203 |
+
"learning_rate": 1.9484108705665595e-05,
|
| 204 |
+
"loss": 0.2665,
|
| 205 |
+
"step": 280
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.13,
|
| 209 |
+
"grad_norm": 7.023002624511719,
|
| 210 |
+
"learning_rate": 1.9465684016582224e-05,
|
| 211 |
+
"loss": 0.1473,
|
| 212 |
+
"step": 290
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.14,
|
| 216 |
+
"grad_norm": 2.650477647781372,
|
| 217 |
+
"learning_rate": 1.9447259327498852e-05,
|
| 218 |
+
"loss": 0.1301,
|
| 219 |
+
"step": 300
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.14,
|
| 223 |
+
"grad_norm": 18.17388343811035,
|
| 224 |
+
"learning_rate": 1.942883463841548e-05,
|
| 225 |
+
"loss": 0.1753,
|
| 226 |
+
"step": 310
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.15,
|
| 230 |
+
"grad_norm": 13.063555717468262,
|
| 231 |
+
"learning_rate": 1.941040994933211e-05,
|
| 232 |
+
"loss": 0.0762,
|
| 233 |
+
"step": 320
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.15,
|
| 237 |
+
"grad_norm": 31.46049690246582,
|
| 238 |
+
"learning_rate": 1.9391985260248734e-05,
|
| 239 |
+
"loss": 0.1489,
|
| 240 |
+
"step": 330
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.16,
|
| 244 |
+
"grad_norm": 0.7956262230873108,
|
| 245 |
+
"learning_rate": 1.9373560571165363e-05,
|
| 246 |
+
"loss": 0.1253,
|
| 247 |
+
"step": 340
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.16,
|
| 251 |
+
"grad_norm": 16.001853942871094,
|
| 252 |
+
"learning_rate": 1.935513588208199e-05,
|
| 253 |
+
"loss": 0.1346,
|
| 254 |
+
"step": 350
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.17,
|
| 258 |
+
"grad_norm": 16.36872673034668,
|
| 259 |
+
"learning_rate": 1.933671119299862e-05,
|
| 260 |
+
"loss": 0.1792,
|
| 261 |
+
"step": 360
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.17,
|
| 265 |
+
"grad_norm": 10.501505851745605,
|
| 266 |
+
"learning_rate": 1.9318286503915248e-05,
|
| 267 |
+
"loss": 0.1376,
|
| 268 |
+
"step": 370
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.17,
|
| 272 |
+
"grad_norm": 18.060327529907227,
|
| 273 |
+
"learning_rate": 1.9299861814831876e-05,
|
| 274 |
+
"loss": 0.1475,
|
| 275 |
+
"step": 380
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.18,
|
| 279 |
+
"grad_norm": 2.5868239402770996,
|
| 280 |
+
"learning_rate": 1.9281437125748505e-05,
|
| 281 |
+
"loss": 0.1019,
|
| 282 |
+
"step": 390
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.18,
|
| 286 |
+
"grad_norm": 18.798660278320312,
|
| 287 |
+
"learning_rate": 1.9263012436665133e-05,
|
| 288 |
+
"loss": 0.1969,
|
| 289 |
+
"step": 400
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.19,
|
| 293 |
+
"grad_norm": 3.9084548950195312,
|
| 294 |
+
"learning_rate": 1.9244587747581762e-05,
|
| 295 |
+
"loss": 0.1064,
|
| 296 |
+
"step": 410
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.19,
|
| 300 |
+
"grad_norm": 17.513097763061523,
|
| 301 |
+
"learning_rate": 1.922616305849839e-05,
|
| 302 |
+
"loss": 0.0982,
|
| 303 |
+
"step": 420
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.2,
|
| 307 |
+
"grad_norm": 1.9817689657211304,
|
| 308 |
+
"learning_rate": 1.920773836941502e-05,
|
| 309 |
+
"loss": 0.0714,
|
| 310 |
+
"step": 430
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.2,
|
| 314 |
+
"grad_norm": 6.034035682678223,
|
| 315 |
+
"learning_rate": 1.9189313680331644e-05,
|
| 316 |
+
"loss": 0.1101,
|
| 317 |
+
"step": 440
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.21,
|
| 321 |
+
"grad_norm": 7.219274044036865,
|
| 322 |
+
"learning_rate": 1.9170888991248272e-05,
|
| 323 |
+
"loss": 0.1688,
|
| 324 |
+
"step": 450
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.21,
|
| 328 |
+
"grad_norm": 2.1540260314941406,
|
| 329 |
+
"learning_rate": 1.91524643021649e-05,
|
| 330 |
+
"loss": 0.1087,
|
| 331 |
+
"step": 460
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.22,
|
| 335 |
+
"grad_norm": 3.1677350997924805,
|
| 336 |
+
"learning_rate": 1.913403961308153e-05,
|
| 337 |
+
"loss": 0.1229,
|
| 338 |
+
"step": 470
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.22,
|
| 342 |
+
"grad_norm": 4.970607757568359,
|
| 343 |
+
"learning_rate": 1.9115614923998158e-05,
|
| 344 |
+
"loss": 0.0605,
|
| 345 |
+
"step": 480
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.23,
|
| 349 |
+
"grad_norm": 17.171939849853516,
|
| 350 |
+
"learning_rate": 1.9097190234914786e-05,
|
| 351 |
+
"loss": 0.1796,
|
| 352 |
+
"step": 490
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.23,
|
| 356 |
+
"grad_norm": 11.96903133392334,
|
| 357 |
+
"learning_rate": 1.9078765545831415e-05,
|
| 358 |
+
"loss": 0.115,
|
| 359 |
+
"step": 500
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.23,
|
| 363 |
+
"grad_norm": 9.562213897705078,
|
| 364 |
+
"learning_rate": 1.9060340856748043e-05,
|
| 365 |
+
"loss": 0.0944,
|
| 366 |
+
"step": 510
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.24,
|
| 370 |
+
"grad_norm": 5.644919395446777,
|
| 371 |
+
"learning_rate": 1.904191616766467e-05,
|
| 372 |
+
"loss": 0.2544,
|
| 373 |
+
"step": 520
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.24,
|
| 377 |
+
"grad_norm": 13.460328102111816,
|
| 378 |
+
"learning_rate": 1.90234914785813e-05,
|
| 379 |
+
"loss": 0.2247,
|
| 380 |
+
"step": 530
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.25,
|
| 384 |
+
"grad_norm": 11.407585144042969,
|
| 385 |
+
"learning_rate": 1.900506678949793e-05,
|
| 386 |
+
"loss": 0.0726,
|
| 387 |
+
"step": 540
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.25,
|
| 391 |
+
"grad_norm": 21.495920181274414,
|
| 392 |
+
"learning_rate": 1.8986642100414557e-05,
|
| 393 |
+
"loss": 0.1726,
|
| 394 |
+
"step": 550
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.26,
|
| 398 |
+
"grad_norm": 0.34193065762519836,
|
| 399 |
+
"learning_rate": 1.8968217411331185e-05,
|
| 400 |
+
"loss": 0.2303,
|
| 401 |
+
"step": 560
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.26,
|
| 405 |
+
"grad_norm": 11.13711166381836,
|
| 406 |
+
"learning_rate": 1.8949792722247814e-05,
|
| 407 |
+
"loss": 0.1242,
|
| 408 |
+
"step": 570
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.27,
|
| 412 |
+
"grad_norm": 23.82451057434082,
|
| 413 |
+
"learning_rate": 1.8931368033164442e-05,
|
| 414 |
+
"loss": 0.1661,
|
| 415 |
+
"step": 580
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.27,
|
| 419 |
+
"grad_norm": 12.27181339263916,
|
| 420 |
+
"learning_rate": 1.891294334408107e-05,
|
| 421 |
+
"loss": 0.2051,
|
| 422 |
+
"step": 590
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.28,
|
| 426 |
+
"grad_norm": 0.061552494764328,
|
| 427 |
+
"learning_rate": 1.88945186549977e-05,
|
| 428 |
+
"loss": 0.1662,
|
| 429 |
+
"step": 600
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.28,
|
| 433 |
+
"grad_norm": 3.284986972808838,
|
| 434 |
+
"learning_rate": 1.8876093965914328e-05,
|
| 435 |
+
"loss": 0.0499,
|
| 436 |
+
"step": 610
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.29,
|
| 440 |
+
"grad_norm": 9.97547435760498,
|
| 441 |
+
"learning_rate": 1.8857669276830953e-05,
|
| 442 |
+
"loss": 0.1161,
|
| 443 |
+
"step": 620
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.29,
|
| 447 |
+
"grad_norm": 13.979111671447754,
|
| 448 |
+
"learning_rate": 1.883924458774758e-05,
|
| 449 |
+
"loss": 0.2008,
|
| 450 |
+
"step": 630
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.29,
|
| 454 |
+
"grad_norm": 0.7910891771316528,
|
| 455 |
+
"learning_rate": 1.882081989866421e-05,
|
| 456 |
+
"loss": 0.1556,
|
| 457 |
+
"step": 640
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.3,
|
| 461 |
+
"grad_norm": 0.24290454387664795,
|
| 462 |
+
"learning_rate": 1.8802395209580838e-05,
|
| 463 |
+
"loss": 0.0752,
|
| 464 |
+
"step": 650
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.3,
|
| 468 |
+
"grad_norm": 7.324313640594482,
|
| 469 |
+
"learning_rate": 1.8783970520497467e-05,
|
| 470 |
+
"loss": 0.1932,
|
| 471 |
+
"step": 660
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.31,
|
| 475 |
+
"grad_norm": 17.672924041748047,
|
| 476 |
+
"learning_rate": 1.8765545831414095e-05,
|
| 477 |
+
"loss": 0.137,
|
| 478 |
+
"step": 670
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.31,
|
| 482 |
+
"grad_norm": 0.5127893686294556,
|
| 483 |
+
"learning_rate": 1.8747121142330723e-05,
|
| 484 |
+
"loss": 0.0727,
|
| 485 |
+
"step": 680
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.32,
|
| 489 |
+
"grad_norm": 4.361024379730225,
|
| 490 |
+
"learning_rate": 1.8728696453247352e-05,
|
| 491 |
+
"loss": 0.0937,
|
| 492 |
+
"step": 690
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.32,
|
| 496 |
+
"grad_norm": 3.0104458332061768,
|
| 497 |
+
"learning_rate": 1.871027176416398e-05,
|
| 498 |
+
"loss": 0.0724,
|
| 499 |
+
"step": 700
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.33,
|
| 503 |
+
"grad_norm": 12.688844680786133,
|
| 504 |
+
"learning_rate": 1.869184707508061e-05,
|
| 505 |
+
"loss": 0.1766,
|
| 506 |
+
"step": 710
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.33,
|
| 510 |
+
"grad_norm": 14.999696731567383,
|
| 511 |
+
"learning_rate": 1.8673422385997237e-05,
|
| 512 |
+
"loss": 0.1306,
|
| 513 |
+
"step": 720
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.34,
|
| 517 |
+
"grad_norm": 9.746272087097168,
|
| 518 |
+
"learning_rate": 1.8654997696913866e-05,
|
| 519 |
+
"loss": 0.1128,
|
| 520 |
+
"step": 730
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.34,
|
| 524 |
+
"grad_norm": 3.1531741619110107,
|
| 525 |
+
"learning_rate": 1.8636573007830494e-05,
|
| 526 |
+
"loss": 0.0917,
|
| 527 |
+
"step": 740
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.35,
|
| 531 |
+
"grad_norm": 3.493957996368408,
|
| 532 |
+
"learning_rate": 1.8618148318747123e-05,
|
| 533 |
+
"loss": 0.1138,
|
| 534 |
+
"step": 750
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.35,
|
| 538 |
+
"grad_norm": 2.8383517265319824,
|
| 539 |
+
"learning_rate": 1.859972362966375e-05,
|
| 540 |
+
"loss": 0.2223,
|
| 541 |
+
"step": 760
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.35,
|
| 545 |
+
"grad_norm": 16.254199981689453,
|
| 546 |
+
"learning_rate": 1.858129894058038e-05,
|
| 547 |
+
"loss": 0.1461,
|
| 548 |
+
"step": 770
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.36,
|
| 552 |
+
"grad_norm": 0.10699751228094101,
|
| 553 |
+
"learning_rate": 1.8562874251497008e-05,
|
| 554 |
+
"loss": 0.1238,
|
| 555 |
+
"step": 780
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.36,
|
| 559 |
+
"grad_norm": 7.592443466186523,
|
| 560 |
+
"learning_rate": 1.8544449562413637e-05,
|
| 561 |
+
"loss": 0.1773,
|
| 562 |
+
"step": 790
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.37,
|
| 566 |
+
"grad_norm": 10.313605308532715,
|
| 567 |
+
"learning_rate": 1.8526024873330265e-05,
|
| 568 |
+
"loss": 0.1663,
|
| 569 |
+
"step": 800
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.37,
|
| 573 |
+
"grad_norm": 0.4551958739757538,
|
| 574 |
+
"learning_rate": 1.8507600184246893e-05,
|
| 575 |
+
"loss": 0.0778,
|
| 576 |
+
"step": 810
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.38,
|
| 580 |
+
"grad_norm": 6.450429439544678,
|
| 581 |
+
"learning_rate": 1.8489175495163522e-05,
|
| 582 |
+
"loss": 0.1039,
|
| 583 |
+
"step": 820
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.38,
|
| 587 |
+
"grad_norm": 1.3565691709518433,
|
| 588 |
+
"learning_rate": 1.847075080608015e-05,
|
| 589 |
+
"loss": 0.0291,
|
| 590 |
+
"step": 830
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.39,
|
| 594 |
+
"grad_norm": 1.8334587812423706,
|
| 595 |
+
"learning_rate": 1.8452326116996775e-05,
|
| 596 |
+
"loss": 0.0881,
|
| 597 |
+
"step": 840
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.39,
|
| 601 |
+
"grad_norm": 15.387334823608398,
|
| 602 |
+
"learning_rate": 1.8433901427913404e-05,
|
| 603 |
+
"loss": 0.1311,
|
| 604 |
+
"step": 850
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.4,
|
| 608 |
+
"grad_norm": 23.866994857788086,
|
| 609 |
+
"learning_rate": 1.8415476738830032e-05,
|
| 610 |
+
"loss": 0.1434,
|
| 611 |
+
"step": 860
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.4,
|
| 615 |
+
"grad_norm": 4.579155445098877,
|
| 616 |
+
"learning_rate": 1.839705204974666e-05,
|
| 617 |
+
"loss": 0.1681,
|
| 618 |
+
"step": 870
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.41,
|
| 622 |
+
"grad_norm": 17.968294143676758,
|
| 623 |
+
"learning_rate": 1.837862736066329e-05,
|
| 624 |
+
"loss": 0.1115,
|
| 625 |
+
"step": 880
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.41,
|
| 629 |
+
"grad_norm": 1.752333164215088,
|
| 630 |
+
"learning_rate": 1.8360202671579918e-05,
|
| 631 |
+
"loss": 0.0856,
|
| 632 |
+
"step": 890
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.41,
|
| 636 |
+
"grad_norm": 6.359458923339844,
|
| 637 |
+
"learning_rate": 1.8341777982496546e-05,
|
| 638 |
+
"loss": 0.076,
|
| 639 |
+
"step": 900
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.42,
|
| 643 |
+
"grad_norm": 0.0380895622074604,
|
| 644 |
+
"learning_rate": 1.8323353293413175e-05,
|
| 645 |
+
"loss": 0.1085,
|
| 646 |
+
"step": 910
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.42,
|
| 650 |
+
"grad_norm": 1.1666935682296753,
|
| 651 |
+
"learning_rate": 1.8304928604329803e-05,
|
| 652 |
+
"loss": 0.2057,
|
| 653 |
+
"step": 920
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.43,
|
| 657 |
+
"grad_norm": 5.157413959503174,
|
| 658 |
+
"learning_rate": 1.828650391524643e-05,
|
| 659 |
+
"loss": 0.0712,
|
| 660 |
+
"step": 930
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.43,
|
| 664 |
+
"grad_norm": 0.13210444152355194,
|
| 665 |
+
"learning_rate": 1.826807922616306e-05,
|
| 666 |
+
"loss": 0.0844,
|
| 667 |
+
"step": 940
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.44,
|
| 671 |
+
"grad_norm": 20.039241790771484,
|
| 672 |
+
"learning_rate": 1.824965453707969e-05,
|
| 673 |
+
"loss": 0.1452,
|
| 674 |
+
"step": 950
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.44,
|
| 678 |
+
"grad_norm": 0.7453742623329163,
|
| 679 |
+
"learning_rate": 1.8231229847996317e-05,
|
| 680 |
+
"loss": 0.0401,
|
| 681 |
+
"step": 960
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.45,
|
| 685 |
+
"grad_norm": 11.507768630981445,
|
| 686 |
+
"learning_rate": 1.8212805158912945e-05,
|
| 687 |
+
"loss": 0.0655,
|
| 688 |
+
"step": 970
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.45,
|
| 692 |
+
"grad_norm": 0.23590461909770966,
|
| 693 |
+
"learning_rate": 1.8194380469829574e-05,
|
| 694 |
+
"loss": 0.2963,
|
| 695 |
+
"step": 980
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.46,
|
| 699 |
+
"grad_norm": 8.790035247802734,
|
| 700 |
+
"learning_rate": 1.8175955780746202e-05,
|
| 701 |
+
"loss": 0.0552,
|
| 702 |
+
"step": 990
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.46,
|
| 706 |
+
"grad_norm": 0.280657023191452,
|
| 707 |
+
"learning_rate": 1.815753109166283e-05,
|
| 708 |
+
"loss": 0.0751,
|
| 709 |
+
"step": 1000
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.47,
|
| 713 |
+
"grad_norm": 0.0973009243607521,
|
| 714 |
+
"learning_rate": 1.813910640257946e-05,
|
| 715 |
+
"loss": 0.1306,
|
| 716 |
+
"step": 1010
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.47,
|
| 720 |
+
"grad_norm": 2.953173875808716,
|
| 721 |
+
"learning_rate": 1.8120681713496088e-05,
|
| 722 |
+
"loss": 0.1342,
|
| 723 |
+
"step": 1020
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.47,
|
| 727 |
+
"grad_norm": 2.1666452884674072,
|
| 728 |
+
"learning_rate": 1.8102257024412716e-05,
|
| 729 |
+
"loss": 0.1745,
|
| 730 |
+
"step": 1030
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.48,
|
| 734 |
+
"grad_norm": 11.335593223571777,
|
| 735 |
+
"learning_rate": 1.8083832335329345e-05,
|
| 736 |
+
"loss": 0.2019,
|
| 737 |
+
"step": 1040
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.48,
|
| 741 |
+
"grad_norm": 17.71516990661621,
|
| 742 |
+
"learning_rate": 1.8065407646245973e-05,
|
| 743 |
+
"loss": 0.122,
|
| 744 |
+
"step": 1050
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.49,
|
| 748 |
+
"grad_norm": 19.908559799194336,
|
| 749 |
+
"learning_rate": 1.8046982957162598e-05,
|
| 750 |
+
"loss": 0.2149,
|
| 751 |
+
"step": 1060
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.49,
|
| 755 |
+
"grad_norm": 26.181245803833008,
|
| 756 |
+
"learning_rate": 1.8028558268079227e-05,
|
| 757 |
+
"loss": 0.1469,
|
| 758 |
+
"step": 1070
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.5,
|
| 762 |
+
"grad_norm": 0.14569219946861267,
|
| 763 |
+
"learning_rate": 1.8010133578995855e-05,
|
| 764 |
+
"loss": 0.2072,
|
| 765 |
+
"step": 1080
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.5,
|
| 769 |
+
"grad_norm": 13.031170845031738,
|
| 770 |
+
"learning_rate": 1.7991708889912484e-05,
|
| 771 |
+
"loss": 0.1896,
|
| 772 |
+
"step": 1090
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.51,
|
| 776 |
+
"grad_norm": 4.332188606262207,
|
| 777 |
+
"learning_rate": 1.7973284200829112e-05,
|
| 778 |
+
"loss": 0.055,
|
| 779 |
+
"step": 1100
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.51,
|
| 783 |
+
"grad_norm": 3.402667760848999,
|
| 784 |
+
"learning_rate": 1.795485951174574e-05,
|
| 785 |
+
"loss": 0.2013,
|
| 786 |
+
"step": 1110
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.52,
|
| 790 |
+
"grad_norm": 0.19696275889873505,
|
| 791 |
+
"learning_rate": 1.793643482266237e-05,
|
| 792 |
+
"loss": 0.1374,
|
| 793 |
+
"step": 1120
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.52,
|
| 797 |
+
"grad_norm": 0.8259896039962769,
|
| 798 |
+
"learning_rate": 1.7918010133578997e-05,
|
| 799 |
+
"loss": 0.286,
|
| 800 |
+
"step": 1130
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.52,
|
| 804 |
+
"grad_norm": 3.934999704360962,
|
| 805 |
+
"learning_rate": 1.7899585444495626e-05,
|
| 806 |
+
"loss": 0.1217,
|
| 807 |
+
"step": 1140
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.53,
|
| 811 |
+
"grad_norm": 0.3496149778366089,
|
| 812 |
+
"learning_rate": 1.7881160755412254e-05,
|
| 813 |
+
"loss": 0.0286,
|
| 814 |
+
"step": 1150
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.53,
|
| 818 |
+
"grad_norm": 8.93337345123291,
|
| 819 |
+
"learning_rate": 1.7862736066328883e-05,
|
| 820 |
+
"loss": 0.1058,
|
| 821 |
+
"step": 1160
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.54,
|
| 825 |
+
"grad_norm": 2.22180438041687,
|
| 826 |
+
"learning_rate": 1.784431137724551e-05,
|
| 827 |
+
"loss": 0.2495,
|
| 828 |
+
"step": 1170
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.54,
|
| 832 |
+
"grad_norm": 7.9512858390808105,
|
| 833 |
+
"learning_rate": 1.782588668816214e-05,
|
| 834 |
+
"loss": 0.0632,
|
| 835 |
+
"step": 1180
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.55,
|
| 839 |
+
"grad_norm": 2.3138413429260254,
|
| 840 |
+
"learning_rate": 1.7807461999078768e-05,
|
| 841 |
+
"loss": 0.0533,
|
| 842 |
+
"step": 1190
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.55,
|
| 846 |
+
"grad_norm": 5.726136684417725,
|
| 847 |
+
"learning_rate": 1.7789037309995397e-05,
|
| 848 |
+
"loss": 0.0359,
|
| 849 |
+
"step": 1200
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.56,
|
| 853 |
+
"grad_norm": 5.151144981384277,
|
| 854 |
+
"learning_rate": 1.7770612620912025e-05,
|
| 855 |
+
"loss": 0.069,
|
| 856 |
+
"step": 1210
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.56,
|
| 860 |
+
"grad_norm": 4.789042949676514,
|
| 861 |
+
"learning_rate": 1.7752187931828654e-05,
|
| 862 |
+
"loss": 0.1835,
|
| 863 |
+
"step": 1220
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.57,
|
| 867 |
+
"grad_norm": 13.844017028808594,
|
| 868 |
+
"learning_rate": 1.7733763242745282e-05,
|
| 869 |
+
"loss": 0.0905,
|
| 870 |
+
"step": 1230
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.57,
|
| 874 |
+
"grad_norm": 8.862701416015625,
|
| 875 |
+
"learning_rate": 1.771533855366191e-05,
|
| 876 |
+
"loss": 0.209,
|
| 877 |
+
"step": 1240
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.58,
|
| 881 |
+
"grad_norm": 1.3909494876861572,
|
| 882 |
+
"learning_rate": 1.769691386457854e-05,
|
| 883 |
+
"loss": 0.0344,
|
| 884 |
+
"step": 1250
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.58,
|
| 888 |
+
"grad_norm": 0.3043901026248932,
|
| 889 |
+
"learning_rate": 1.7678489175495167e-05,
|
| 890 |
+
"loss": 0.0513,
|
| 891 |
+
"step": 1260
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.58,
|
| 895 |
+
"grad_norm": 3.006042718887329,
|
| 896 |
+
"learning_rate": 1.7660064486411796e-05,
|
| 897 |
+
"loss": 0.0757,
|
| 898 |
+
"step": 1270
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.59,
|
| 902 |
+
"grad_norm": 0.05583062395453453,
|
| 903 |
+
"learning_rate": 1.764163979732842e-05,
|
| 904 |
+
"loss": 0.0508,
|
| 905 |
+
"step": 1280
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.59,
|
| 909 |
+
"grad_norm": 14.672821044921875,
|
| 910 |
+
"learning_rate": 1.762321510824505e-05,
|
| 911 |
+
"loss": 0.2056,
|
| 912 |
+
"step": 1290
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.6,
|
| 916 |
+
"grad_norm": 0.50792396068573,
|
| 917 |
+
"learning_rate": 1.7604790419161678e-05,
|
| 918 |
+
"loss": 0.0526,
|
| 919 |
+
"step": 1300
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.6,
|
| 923 |
+
"grad_norm": 3.128744602203369,
|
| 924 |
+
"learning_rate": 1.7586365730078306e-05,
|
| 925 |
+
"loss": 0.1255,
|
| 926 |
+
"step": 1310
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.61,
|
| 930 |
+
"grad_norm": 5.814742088317871,
|
| 931 |
+
"learning_rate": 1.7567941040994935e-05,
|
| 932 |
+
"loss": 0.0877,
|
| 933 |
+
"step": 1320
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.61,
|
| 937 |
+
"grad_norm": 24.0638484954834,
|
| 938 |
+
"learning_rate": 1.7549516351911563e-05,
|
| 939 |
+
"loss": 0.0784,
|
| 940 |
+
"step": 1330
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.62,
|
| 944 |
+
"grad_norm": 2.3511364459991455,
|
| 945 |
+
"learning_rate": 1.7531091662828192e-05,
|
| 946 |
+
"loss": 0.0829,
|
| 947 |
+
"step": 1340
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.62,
|
| 951 |
+
"grad_norm": 19.565967559814453,
|
| 952 |
+
"learning_rate": 1.7512666973744817e-05,
|
| 953 |
+
"loss": 0.2538,
|
| 954 |
+
"step": 1350
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.63,
|
| 958 |
+
"grad_norm": 21.208799362182617,
|
| 959 |
+
"learning_rate": 1.7494242284661445e-05,
|
| 960 |
+
"loss": 0.0497,
|
| 961 |
+
"step": 1360
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.63,
|
| 965 |
+
"grad_norm": 0.3589443266391754,
|
| 966 |
+
"learning_rate": 1.7475817595578074e-05,
|
| 967 |
+
"loss": 0.0155,
|
| 968 |
+
"step": 1370
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.64,
|
| 972 |
+
"grad_norm": 4.569169521331787,
|
| 973 |
+
"learning_rate": 1.7457392906494702e-05,
|
| 974 |
+
"loss": 0.0638,
|
| 975 |
+
"step": 1380
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.64,
|
| 979 |
+
"grad_norm": 3.0978171825408936,
|
| 980 |
+
"learning_rate": 1.743896821741133e-05,
|
| 981 |
+
"loss": 0.0294,
|
| 982 |
+
"step": 1390
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.64,
|
| 986 |
+
"grad_norm": 23.960538864135742,
|
| 987 |
+
"learning_rate": 1.742054352832796e-05,
|
| 988 |
+
"loss": 0.1144,
|
| 989 |
+
"step": 1400
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.65,
|
| 993 |
+
"grad_norm": 31.21448516845703,
|
| 994 |
+
"learning_rate": 1.7402118839244588e-05,
|
| 995 |
+
"loss": 0.2082,
|
| 996 |
+
"step": 1410
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.65,
|
| 1000 |
+
"grad_norm": 2.9483284950256348,
|
| 1001 |
+
"learning_rate": 1.7383694150161216e-05,
|
| 1002 |
+
"loss": 0.1238,
|
| 1003 |
+
"step": 1420
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.66,
|
| 1007 |
+
"grad_norm": 7.06640625,
|
| 1008 |
+
"learning_rate": 1.7365269461077845e-05,
|
| 1009 |
+
"loss": 0.0257,
|
| 1010 |
+
"step": 1430
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.66,
|
| 1014 |
+
"grad_norm": 17.621641159057617,
|
| 1015 |
+
"learning_rate": 1.7346844771994473e-05,
|
| 1016 |
+
"loss": 0.0769,
|
| 1017 |
+
"step": 1440
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.67,
|
| 1021 |
+
"grad_norm": 0.5330172181129456,
|
| 1022 |
+
"learning_rate": 1.73284200829111e-05,
|
| 1023 |
+
"loss": 0.1089,
|
| 1024 |
+
"step": 1450
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.67,
|
| 1028 |
+
"grad_norm": 3.053800106048584,
|
| 1029 |
+
"learning_rate": 1.730999539382773e-05,
|
| 1030 |
+
"loss": 0.0507,
|
| 1031 |
+
"step": 1460
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.68,
|
| 1035 |
+
"grad_norm": 0.17157615721225739,
|
| 1036 |
+
"learning_rate": 1.729157070474436e-05,
|
| 1037 |
+
"loss": 0.124,
|
| 1038 |
+
"step": 1470
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.68,
|
| 1042 |
+
"grad_norm": 0.049004122614860535,
|
| 1043 |
+
"learning_rate": 1.7273146015660987e-05,
|
| 1044 |
+
"loss": 0.1686,
|
| 1045 |
+
"step": 1480
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.69,
|
| 1049 |
+
"grad_norm": 0.28901880979537964,
|
| 1050 |
+
"learning_rate": 1.7254721326577615e-05,
|
| 1051 |
+
"loss": 0.096,
|
| 1052 |
+
"step": 1490
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.69,
|
| 1056 |
+
"grad_norm": 0.08810051530599594,
|
| 1057 |
+
"learning_rate": 1.7236296637494244e-05,
|
| 1058 |
+
"loss": 0.0706,
|
| 1059 |
+
"step": 1500
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.7,
|
| 1063 |
+
"grad_norm": 0.0002217306464444846,
|
| 1064 |
+
"learning_rate": 1.7217871948410872e-05,
|
| 1065 |
+
"loss": 0.0558,
|
| 1066 |
+
"step": 1510
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.7,
|
| 1070 |
+
"grad_norm": 0.16710884869098663,
|
| 1071 |
+
"learning_rate": 1.71994472593275e-05,
|
| 1072 |
+
"loss": 0.2065,
|
| 1073 |
+
"step": 1520
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.7,
|
| 1077 |
+
"grad_norm": 0.09904110431671143,
|
| 1078 |
+
"learning_rate": 1.718102257024413e-05,
|
| 1079 |
+
"loss": 0.02,
|
| 1080 |
+
"step": 1530
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 0.71,
|
| 1084 |
+
"grad_norm": 0.0009630492422729731,
|
| 1085 |
+
"learning_rate": 1.7162597881160758e-05,
|
| 1086 |
+
"loss": 0.0612,
|
| 1087 |
+
"step": 1540
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.71,
|
| 1091 |
+
"grad_norm": 7.022042274475098,
|
| 1092 |
+
"learning_rate": 1.7144173192077386e-05,
|
| 1093 |
+
"loss": 0.0381,
|
| 1094 |
+
"step": 1550
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.72,
|
| 1098 |
+
"grad_norm": 0.5450890064239502,
|
| 1099 |
+
"learning_rate": 1.7125748502994015e-05,
|
| 1100 |
+
"loss": 0.0512,
|
| 1101 |
+
"step": 1560
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.72,
|
| 1105 |
+
"grad_norm": 2.2123756408691406,
|
| 1106 |
+
"learning_rate": 1.710732381391064e-05,
|
| 1107 |
+
"loss": 0.041,
|
| 1108 |
+
"step": 1570
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 0.73,
|
| 1112 |
+
"grad_norm": 3.81233286857605,
|
| 1113 |
+
"learning_rate": 1.7088899124827268e-05,
|
| 1114 |
+
"loss": 0.1295,
|
| 1115 |
+
"step": 1580
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.73,
|
| 1119 |
+
"grad_norm": 14.528470039367676,
|
| 1120 |
+
"learning_rate": 1.7070474435743897e-05,
|
| 1121 |
+
"loss": 0.1534,
|
| 1122 |
+
"step": 1590
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 0.74,
|
| 1126 |
+
"grad_norm": 13.945511817932129,
|
| 1127 |
+
"learning_rate": 1.7052049746660525e-05,
|
| 1128 |
+
"loss": 0.1612,
|
| 1129 |
+
"step": 1600
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.74,
|
| 1133 |
+
"grad_norm": 4.4138898849487305,
|
| 1134 |
+
"learning_rate": 1.7033625057577153e-05,
|
| 1135 |
+
"loss": 0.1254,
|
| 1136 |
+
"step": 1610
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.75,
|
| 1140 |
+
"grad_norm": 9.094703674316406,
|
| 1141 |
+
"learning_rate": 1.7015200368493782e-05,
|
| 1142 |
+
"loss": 0.1708,
|
| 1143 |
+
"step": 1620
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.75,
|
| 1147 |
+
"grad_norm": 5.324583053588867,
|
| 1148 |
+
"learning_rate": 1.699677567941041e-05,
|
| 1149 |
+
"loss": 0.1757,
|
| 1150 |
+
"step": 1630
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 0.76,
|
| 1154 |
+
"grad_norm": 16.52385902404785,
|
| 1155 |
+
"learning_rate": 1.697835099032704e-05,
|
| 1156 |
+
"loss": 0.0837,
|
| 1157 |
+
"step": 1640
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.76,
|
| 1161 |
+
"grad_norm": 4.362677097320557,
|
| 1162 |
+
"learning_rate": 1.6959926301243667e-05,
|
| 1163 |
+
"loss": 0.1586,
|
| 1164 |
+
"step": 1650
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 0.76,
|
| 1168 |
+
"grad_norm": 8.995072364807129,
|
| 1169 |
+
"learning_rate": 1.6941501612160296e-05,
|
| 1170 |
+
"loss": 0.2564,
|
| 1171 |
+
"step": 1660
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.77,
|
| 1175 |
+
"grad_norm": 13.925511360168457,
|
| 1176 |
+
"learning_rate": 1.6923076923076924e-05,
|
| 1177 |
+
"loss": 0.1162,
|
| 1178 |
+
"step": 1670
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.77,
|
| 1182 |
+
"grad_norm": 17.29884910583496,
|
| 1183 |
+
"learning_rate": 1.6904652233993553e-05,
|
| 1184 |
+
"loss": 0.1417,
|
| 1185 |
+
"step": 1680
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.78,
|
| 1189 |
+
"grad_norm": 10.468055725097656,
|
| 1190 |
+
"learning_rate": 1.688622754491018e-05,
|
| 1191 |
+
"loss": 0.1014,
|
| 1192 |
+
"step": 1690
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 0.78,
|
| 1196 |
+
"grad_norm": 13.775412559509277,
|
| 1197 |
+
"learning_rate": 1.686780285582681e-05,
|
| 1198 |
+
"loss": 0.2368,
|
| 1199 |
+
"step": 1700
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.79,
|
| 1203 |
+
"grad_norm": 6.128839492797852,
|
| 1204 |
+
"learning_rate": 1.6849378166743438e-05,
|
| 1205 |
+
"loss": 0.0507,
|
| 1206 |
+
"step": 1710
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 0.79,
|
| 1210 |
+
"grad_norm": 10.073379516601562,
|
| 1211 |
+
"learning_rate": 1.6830953477660067e-05,
|
| 1212 |
+
"loss": 0.0755,
|
| 1213 |
+
"step": 1720
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.8,
|
| 1217 |
+
"grad_norm": 0.9331756234169006,
|
| 1218 |
+
"learning_rate": 1.6812528788576695e-05,
|
| 1219 |
+
"loss": 0.0316,
|
| 1220 |
+
"step": 1730
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 0.8,
|
| 1224 |
+
"grad_norm": 0.06095833703875542,
|
| 1225 |
+
"learning_rate": 1.6794104099493323e-05,
|
| 1226 |
+
"loss": 0.1482,
|
| 1227 |
+
"step": 1740
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.81,
|
| 1231 |
+
"grad_norm": 4.112434387207031,
|
| 1232 |
+
"learning_rate": 1.6775679410409952e-05,
|
| 1233 |
+
"loss": 0.1104,
|
| 1234 |
+
"step": 1750
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 0.81,
|
| 1238 |
+
"grad_norm": 0.18547752499580383,
|
| 1239 |
+
"learning_rate": 1.675725472132658e-05,
|
| 1240 |
+
"loss": 0.124,
|
| 1241 |
+
"step": 1760
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.82,
|
| 1245 |
+
"grad_norm": 9.885139465332031,
|
| 1246 |
+
"learning_rate": 1.673883003224321e-05,
|
| 1247 |
+
"loss": 0.1263,
|
| 1248 |
+
"step": 1770
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 0.82,
|
| 1252 |
+
"grad_norm": 1.6185027360916138,
|
| 1253 |
+
"learning_rate": 1.6720405343159837e-05,
|
| 1254 |
+
"loss": 0.194,
|
| 1255 |
+
"step": 1780
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.82,
|
| 1259 |
+
"grad_norm": 5.13451623916626,
|
| 1260 |
+
"learning_rate": 1.6701980654076462e-05,
|
| 1261 |
+
"loss": 0.0769,
|
| 1262 |
+
"step": 1790
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 0.83,
|
| 1266 |
+
"grad_norm": 0.39274197816848755,
|
| 1267 |
+
"learning_rate": 1.668355596499309e-05,
|
| 1268 |
+
"loss": 0.0532,
|
| 1269 |
+
"step": 1800
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.83,
|
| 1273 |
+
"grad_norm": 36.26974105834961,
|
| 1274 |
+
"learning_rate": 1.666513127590972e-05,
|
| 1275 |
+
"loss": 0.0916,
|
| 1276 |
+
"step": 1810
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 0.84,
|
| 1280 |
+
"grad_norm": 8.078559875488281,
|
| 1281 |
+
"learning_rate": 1.6646706586826348e-05,
|
| 1282 |
+
"loss": 0.0945,
|
| 1283 |
+
"step": 1820
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.84,
|
| 1287 |
+
"grad_norm": 0.015051756985485554,
|
| 1288 |
+
"learning_rate": 1.6628281897742976e-05,
|
| 1289 |
+
"loss": 0.068,
|
| 1290 |
+
"step": 1830
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 0.85,
|
| 1294 |
+
"grad_norm": 24.1297550201416,
|
| 1295 |
+
"learning_rate": 1.6609857208659605e-05,
|
| 1296 |
+
"loss": 0.183,
|
| 1297 |
+
"step": 1840
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 0.85,
|
| 1301 |
+
"grad_norm": 0.2485981285572052,
|
| 1302 |
+
"learning_rate": 1.6591432519576233e-05,
|
| 1303 |
+
"loss": 0.2339,
|
| 1304 |
+
"step": 1850
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 0.86,
|
| 1308 |
+
"grad_norm": 24.146484375,
|
| 1309 |
+
"learning_rate": 1.657300783049286e-05,
|
| 1310 |
+
"loss": 0.1163,
|
| 1311 |
+
"step": 1860
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 0.86,
|
| 1315 |
+
"grad_norm": 0.0001443505461793393,
|
| 1316 |
+
"learning_rate": 1.655458314140949e-05,
|
| 1317 |
+
"loss": 0.0871,
|
| 1318 |
+
"step": 1870
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 0.87,
|
| 1322 |
+
"grad_norm": 0.0009001191356219351,
|
| 1323 |
+
"learning_rate": 1.653615845232612e-05,
|
| 1324 |
+
"loss": 0.0791,
|
| 1325 |
+
"step": 1880
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 0.87,
|
| 1329 |
+
"grad_norm": 1.2001979351043701,
|
| 1330 |
+
"learning_rate": 1.6517733763242747e-05,
|
| 1331 |
+
"loss": 0.1238,
|
| 1332 |
+
"step": 1890
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 0.87,
|
| 1336 |
+
"grad_norm": 16.16439437866211,
|
| 1337 |
+
"learning_rate": 1.6499309074159375e-05,
|
| 1338 |
+
"loss": 0.0517,
|
| 1339 |
+
"step": 1900
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 0.88,
|
| 1343 |
+
"grad_norm": 0.0023416222538799047,
|
| 1344 |
+
"learning_rate": 1.6480884385076004e-05,
|
| 1345 |
+
"loss": 0.1066,
|
| 1346 |
+
"step": 1910
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 0.88,
|
| 1350 |
+
"grad_norm": 1.0242302417755127,
|
| 1351 |
+
"learning_rate": 1.6462459695992632e-05,
|
| 1352 |
+
"loss": 0.0659,
|
| 1353 |
+
"step": 1920
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 0.89,
|
| 1357 |
+
"grad_norm": 9.080098152160645,
|
| 1358 |
+
"learning_rate": 1.644403500690926e-05,
|
| 1359 |
+
"loss": 0.0516,
|
| 1360 |
+
"step": 1930
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 0.89,
|
| 1364 |
+
"grad_norm": 14.89382266998291,
|
| 1365 |
+
"learning_rate": 1.642561031782589e-05,
|
| 1366 |
+
"loss": 0.0409,
|
| 1367 |
+
"step": 1940
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 0.9,
|
| 1371 |
+
"grad_norm": 3.238348960876465,
|
| 1372 |
+
"learning_rate": 1.6407185628742518e-05,
|
| 1373 |
+
"loss": 0.0342,
|
| 1374 |
+
"step": 1950
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 0.9,
|
| 1378 |
+
"grad_norm": 11.910837173461914,
|
| 1379 |
+
"learning_rate": 1.6388760939659146e-05,
|
| 1380 |
+
"loss": 0.116,
|
| 1381 |
+
"step": 1960
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 0.91,
|
| 1385 |
+
"grad_norm": 0.4929376542568207,
|
| 1386 |
+
"learning_rate": 1.6370336250575775e-05,
|
| 1387 |
+
"loss": 0.0443,
|
| 1388 |
+
"step": 1970
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 0.91,
|
| 1392 |
+
"grad_norm": 0.0020700031891465187,
|
| 1393 |
+
"learning_rate": 1.6351911561492403e-05,
|
| 1394 |
+
"loss": 0.0326,
|
| 1395 |
+
"step": 1980
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 0.92,
|
| 1399 |
+
"grad_norm": 9.666726112365723,
|
| 1400 |
+
"learning_rate": 1.633348687240903e-05,
|
| 1401 |
+
"loss": 0.1958,
|
| 1402 |
+
"step": 1990
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 0.92,
|
| 1406 |
+
"grad_norm": 4.457545280456543,
|
| 1407 |
+
"learning_rate": 1.631506218332566e-05,
|
| 1408 |
+
"loss": 0.1035,
|
| 1409 |
+
"step": 2000
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 0.93,
|
| 1413 |
+
"grad_norm": 4.343936920166016,
|
| 1414 |
+
"learning_rate": 1.6296637494242285e-05,
|
| 1415 |
+
"loss": 0.1508,
|
| 1416 |
+
"step": 2010
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 0.93,
|
| 1420 |
+
"grad_norm": 20.539628982543945,
|
| 1421 |
+
"learning_rate": 1.6278212805158914e-05,
|
| 1422 |
+
"loss": 0.1334,
|
| 1423 |
+
"step": 2020
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 0.93,
|
| 1427 |
+
"grad_norm": 10.818260192871094,
|
| 1428 |
+
"learning_rate": 1.6259788116075542e-05,
|
| 1429 |
+
"loss": 0.0406,
|
| 1430 |
+
"step": 2030
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 0.94,
|
| 1434 |
+
"grad_norm": 0.017373841255903244,
|
| 1435 |
+
"learning_rate": 1.624136342699217e-05,
|
| 1436 |
+
"loss": 0.0855,
|
| 1437 |
+
"step": 2040
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 0.94,
|
| 1441 |
+
"grad_norm": 0.1721264272928238,
|
| 1442 |
+
"learning_rate": 1.62229387379088e-05,
|
| 1443 |
+
"loss": 0.2521,
|
| 1444 |
+
"step": 2050
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 0.95,
|
| 1448 |
+
"grad_norm": 0.02232915349304676,
|
| 1449 |
+
"learning_rate": 1.6204514048825427e-05,
|
| 1450 |
+
"loss": 0.0925,
|
| 1451 |
+
"step": 2060
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 0.95,
|
| 1455 |
+
"grad_norm": 0.2657681703567505,
|
| 1456 |
+
"learning_rate": 1.6186089359742056e-05,
|
| 1457 |
+
"loss": 0.2016,
|
| 1458 |
+
"step": 2070
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 0.96,
|
| 1462 |
+
"grad_norm": 17.512866973876953,
|
| 1463 |
+
"learning_rate": 1.6167664670658684e-05,
|
| 1464 |
+
"loss": 0.0556,
|
| 1465 |
+
"step": 2080
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 0.96,
|
| 1469 |
+
"grad_norm": 0.29758042097091675,
|
| 1470 |
+
"learning_rate": 1.6149239981575313e-05,
|
| 1471 |
+
"loss": 0.0612,
|
| 1472 |
+
"step": 2090
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 0.97,
|
| 1476 |
+
"grad_norm": 11.794560432434082,
|
| 1477 |
+
"learning_rate": 1.613081529249194e-05,
|
| 1478 |
+
"loss": 0.231,
|
| 1479 |
+
"step": 2100
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 0.97,
|
| 1483 |
+
"grad_norm": 2.616342067718506,
|
| 1484 |
+
"learning_rate": 1.611239060340857e-05,
|
| 1485 |
+
"loss": 0.1411,
|
| 1486 |
+
"step": 2110
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 0.98,
|
| 1490 |
+
"grad_norm": 1.2001636028289795,
|
| 1491 |
+
"learning_rate": 1.6093965914325198e-05,
|
| 1492 |
+
"loss": 0.1238,
|
| 1493 |
+
"step": 2120
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 0.98,
|
| 1497 |
+
"grad_norm": 27.738758087158203,
|
| 1498 |
+
"learning_rate": 1.6075541225241827e-05,
|
| 1499 |
+
"loss": 0.042,
|
| 1500 |
+
"step": 2130
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 0.99,
|
| 1504 |
+
"grad_norm": 2.5175182819366455,
|
| 1505 |
+
"learning_rate": 1.6057116536158455e-05,
|
| 1506 |
+
"loss": 0.0338,
|
| 1507 |
+
"step": 2140
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 0.99,
|
| 1511 |
+
"grad_norm": 0.003502259962260723,
|
| 1512 |
+
"learning_rate": 1.6038691847075084e-05,
|
| 1513 |
+
"loss": 0.1227,
|
| 1514 |
+
"step": 2150
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 0.99,
|
| 1518 |
+
"grad_norm": 0.3118779957294464,
|
| 1519 |
+
"learning_rate": 1.6020267157991712e-05,
|
| 1520 |
+
"loss": 0.2732,
|
| 1521 |
+
"step": 2160
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 1.0,
|
| 1525 |
+
"grad_norm": 1.3505421876907349,
|
| 1526 |
+
"learning_rate": 1.600184246890834e-05,
|
| 1527 |
+
"loss": 0.1363,
|
| 1528 |
+
"step": 2170
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"epoch": 1.0,
|
| 1532 |
+
"eval_accuracy": 0.9607843137254902,
|
| 1533 |
+
"eval_loss": 0.11265985667705536,
|
| 1534 |
+
"eval_runtime": 88.8351,
|
| 1535 |
+
"eval_samples_per_second": 33.872,
|
| 1536 |
+
"eval_steps_per_second": 6.777,
|
| 1537 |
+
"step": 2171
|
| 1538 |
+
},
|
| 1539 |
+
{
|
| 1540 |
+
"epoch": 1.0,
|
| 1541 |
+
"grad_norm": 0.11328113824129105,
|
| 1542 |
+
"learning_rate": 1.5983417779824966e-05,
|
| 1543 |
+
"loss": 0.066,
|
| 1544 |
+
"step": 2180
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 1.01,
|
| 1548 |
+
"grad_norm": 0.006320483051240444,
|
| 1549 |
+
"learning_rate": 1.5964993090741594e-05,
|
| 1550 |
+
"loss": 0.0227,
|
| 1551 |
+
"step": 2190
|
| 1552 |
+
},
|
| 1553 |
+
{
|
| 1554 |
+
"epoch": 1.01,
|
| 1555 |
+
"grad_norm": 1.3843839168548584,
|
| 1556 |
+
"learning_rate": 1.5946568401658222e-05,
|
| 1557 |
+
"loss": 0.0141,
|
| 1558 |
+
"step": 2200
|
| 1559 |
+
},
|
| 1560 |
+
{
|
| 1561 |
+
"epoch": 1.02,
|
| 1562 |
+
"grad_norm": 22.27046775817871,
|
| 1563 |
+
"learning_rate": 1.592814371257485e-05,
|
| 1564 |
+
"loss": 0.0995,
|
| 1565 |
+
"step": 2210
|
| 1566 |
+
},
|
| 1567 |
+
{
|
| 1568 |
+
"epoch": 1.02,
|
| 1569 |
+
"grad_norm": 6.569302082061768,
|
| 1570 |
+
"learning_rate": 1.590971902349148e-05,
|
| 1571 |
+
"loss": 0.0697,
|
| 1572 |
+
"step": 2220
|
| 1573 |
+
},
|
| 1574 |
+
{
|
| 1575 |
+
"epoch": 1.03,
|
| 1576 |
+
"grad_norm": 0.008038852363824844,
|
| 1577 |
+
"learning_rate": 1.5891294334408108e-05,
|
| 1578 |
+
"loss": 0.0404,
|
| 1579 |
+
"step": 2230
|
| 1580 |
+
},
|
| 1581 |
+
{
|
| 1582 |
+
"epoch": 1.03,
|
| 1583 |
+
"grad_norm": 0.06073816865682602,
|
| 1584 |
+
"learning_rate": 1.5872869645324736e-05,
|
| 1585 |
+
"loss": 0.009,
|
| 1586 |
+
"step": 2240
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 1.04,
|
| 1590 |
+
"grad_norm": 0.07989092916250229,
|
| 1591 |
+
"learning_rate": 1.5854444956241365e-05,
|
| 1592 |
+
"loss": 0.1219,
|
| 1593 |
+
"step": 2250
|
| 1594 |
+
},
|
| 1595 |
+
{
|
| 1596 |
+
"epoch": 1.04,
|
| 1597 |
+
"grad_norm": 5.781001091003418,
|
| 1598 |
+
"learning_rate": 1.5836020267157993e-05,
|
| 1599 |
+
"loss": 0.013,
|
| 1600 |
+
"step": 2260
|
| 1601 |
+
},
|
| 1602 |
+
{
|
| 1603 |
+
"epoch": 1.05,
|
| 1604 |
+
"grad_norm": 0.14038562774658203,
|
| 1605 |
+
"learning_rate": 1.5817595578074622e-05,
|
| 1606 |
+
"loss": 0.0087,
|
| 1607 |
+
"step": 2270
|
| 1608 |
+
},
|
| 1609 |
+
{
|
| 1610 |
+
"epoch": 1.05,
|
| 1611 |
+
"grad_norm": 0.0014603978488594294,
|
| 1612 |
+
"learning_rate": 1.579917088899125e-05,
|
| 1613 |
+
"loss": 0.0977,
|
| 1614 |
+
"step": 2280
|
| 1615 |
+
},
|
| 1616 |
+
{
|
| 1617 |
+
"epoch": 1.05,
|
| 1618 |
+
"grad_norm": 7.043032169342041,
|
| 1619 |
+
"learning_rate": 1.578074619990788e-05,
|
| 1620 |
+
"loss": 0.0471,
|
| 1621 |
+
"step": 2290
|
| 1622 |
+
},
|
| 1623 |
+
{
|
| 1624 |
+
"epoch": 1.06,
|
| 1625 |
+
"grad_norm": 0.019025901332497597,
|
| 1626 |
+
"learning_rate": 1.5762321510824504e-05,
|
| 1627 |
+
"loss": 0.0124,
|
| 1628 |
+
"step": 2300
|
| 1629 |
+
},
|
| 1630 |
+
{
|
| 1631 |
+
"epoch": 1.06,
|
| 1632 |
+
"grad_norm": 8.663049697875977,
|
| 1633 |
+
"learning_rate": 1.5743896821741132e-05,
|
| 1634 |
+
"loss": 0.0095,
|
| 1635 |
+
"step": 2310
|
| 1636 |
+
},
|
| 1637 |
+
{
|
| 1638 |
+
"epoch": 1.07,
|
| 1639 |
+
"grad_norm": 32.70573806762695,
|
| 1640 |
+
"learning_rate": 1.572547213265776e-05,
|
| 1641 |
+
"loss": 0.0372,
|
| 1642 |
+
"step": 2320
|
| 1643 |
+
},
|
| 1644 |
+
{
|
| 1645 |
+
"epoch": 1.07,
|
| 1646 |
+
"grad_norm": 2.611642837524414,
|
| 1647 |
+
"learning_rate": 1.570704744357439e-05,
|
| 1648 |
+
"loss": 0.0261,
|
| 1649 |
+
"step": 2330
|
| 1650 |
+
},
|
| 1651 |
+
{
|
| 1652 |
+
"epoch": 1.08,
|
| 1653 |
+
"grad_norm": 0.029758945107460022,
|
| 1654 |
+
"learning_rate": 1.5688622754491018e-05,
|
| 1655 |
+
"loss": 0.0575,
|
| 1656 |
+
"step": 2340
|
| 1657 |
+
},
|
| 1658 |
+
{
|
| 1659 |
+
"epoch": 1.08,
|
| 1660 |
+
"grad_norm": 1.4188811779022217,
|
| 1661 |
+
"learning_rate": 1.5670198065407646e-05,
|
| 1662 |
+
"loss": 0.0139,
|
| 1663 |
+
"step": 2350
|
| 1664 |
+
},
|
| 1665 |
+
{
|
| 1666 |
+
"epoch": 1.09,
|
| 1667 |
+
"grad_norm": 7.976904392242432,
|
| 1668 |
+
"learning_rate": 1.5651773376324274e-05,
|
| 1669 |
+
"loss": 0.0378,
|
| 1670 |
+
"step": 2360
|
| 1671 |
+
},
|
| 1672 |
+
{
|
| 1673 |
+
"epoch": 1.09,
|
| 1674 |
+
"grad_norm": 35.63239669799805,
|
| 1675 |
+
"learning_rate": 1.5633348687240903e-05,
|
| 1676 |
+
"loss": 0.0737,
|
| 1677 |
+
"step": 2370
|
| 1678 |
+
},
|
| 1679 |
+
{
|
| 1680 |
+
"epoch": 1.1,
|
| 1681 |
+
"grad_norm": 0.29223495721817017,
|
| 1682 |
+
"learning_rate": 1.561492399815753e-05,
|
| 1683 |
+
"loss": 0.0851,
|
| 1684 |
+
"step": 2380
|
| 1685 |
+
},
|
| 1686 |
+
{
|
| 1687 |
+
"epoch": 1.1,
|
| 1688 |
+
"grad_norm": 0.021279122680425644,
|
| 1689 |
+
"learning_rate": 1.559649930907416e-05,
|
| 1690 |
+
"loss": 0.0401,
|
| 1691 |
+
"step": 2390
|
| 1692 |
+
},
|
| 1693 |
+
{
|
| 1694 |
+
"epoch": 1.11,
|
| 1695 |
+
"grad_norm": 1.3651925325393677,
|
| 1696 |
+
"learning_rate": 1.557807461999079e-05,
|
| 1697 |
+
"loss": 0.0747,
|
| 1698 |
+
"step": 2400
|
| 1699 |
+
},
|
| 1700 |
+
{
|
| 1701 |
+
"epoch": 1.11,
|
| 1702 |
+
"grad_norm": 0.08097986876964569,
|
| 1703 |
+
"learning_rate": 1.5559649930907417e-05,
|
| 1704 |
+
"loss": 0.0531,
|
| 1705 |
+
"step": 2410
|
| 1706 |
+
},
|
| 1707 |
+
{
|
| 1708 |
+
"epoch": 1.11,
|
| 1709 |
+
"grad_norm": 0.15957044064998627,
|
| 1710 |
+
"learning_rate": 1.5541225241824045e-05,
|
| 1711 |
+
"loss": 0.0832,
|
| 1712 |
+
"step": 2420
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 1.12,
|
| 1716 |
+
"grad_norm": 0.009705482050776482,
|
| 1717 |
+
"learning_rate": 1.5522800552740674e-05,
|
| 1718 |
+
"loss": 0.1374,
|
| 1719 |
+
"step": 2430
|
| 1720 |
+
},
|
| 1721 |
+
{
|
| 1722 |
+
"epoch": 1.12,
|
| 1723 |
+
"grad_norm": 1.3060195446014404,
|
| 1724 |
+
"learning_rate": 1.5504375863657302e-05,
|
| 1725 |
+
"loss": 0.0457,
|
| 1726 |
+
"step": 2440
|
| 1727 |
+
},
|
| 1728 |
+
{
|
| 1729 |
+
"epoch": 1.13,
|
| 1730 |
+
"grad_norm": 0.6186008453369141,
|
| 1731 |
+
"learning_rate": 1.548595117457393e-05,
|
| 1732 |
+
"loss": 0.0962,
|
| 1733 |
+
"step": 2450
|
| 1734 |
+
},
|
| 1735 |
+
{
|
| 1736 |
+
"epoch": 1.13,
|
| 1737 |
+
"grad_norm": 33.86445617675781,
|
| 1738 |
+
"learning_rate": 1.546752648549056e-05,
|
| 1739 |
+
"loss": 0.0885,
|
| 1740 |
+
"step": 2460
|
| 1741 |
+
},
|
| 1742 |
+
{
|
| 1743 |
+
"epoch": 1.14,
|
| 1744 |
+
"grad_norm": 13.27135944366455,
|
| 1745 |
+
"learning_rate": 1.5449101796407188e-05,
|
| 1746 |
+
"loss": 0.0662,
|
| 1747 |
+
"step": 2470
|
| 1748 |
+
},
|
| 1749 |
+
{
|
| 1750 |
+
"epoch": 1.14,
|
| 1751 |
+
"grad_norm": 0.023092009127140045,
|
| 1752 |
+
"learning_rate": 1.5430677107323816e-05,
|
| 1753 |
+
"loss": 0.0345,
|
| 1754 |
+
"step": 2480
|
| 1755 |
+
},
|
| 1756 |
+
{
|
| 1757 |
+
"epoch": 1.15,
|
| 1758 |
+
"grad_norm": 0.08246669173240662,
|
| 1759 |
+
"learning_rate": 1.5412252418240444e-05,
|
| 1760 |
+
"loss": 0.0438,
|
| 1761 |
+
"step": 2490
|
| 1762 |
+
},
|
| 1763 |
+
{
|
| 1764 |
+
"epoch": 1.15,
|
| 1765 |
+
"grad_norm": 3.7952194213867188,
|
| 1766 |
+
"learning_rate": 1.5393827729157073e-05,
|
| 1767 |
+
"loss": 0.1456,
|
| 1768 |
+
"step": 2500
|
| 1769 |
+
},
|
| 1770 |
+
{
|
| 1771 |
+
"epoch": 1.16,
|
| 1772 |
+
"grad_norm": 0.0032674316316843033,
|
| 1773 |
+
"learning_rate": 1.53754030400737e-05,
|
| 1774 |
+
"loss": 0.0223,
|
| 1775 |
+
"step": 2510
|
| 1776 |
+
},
|
| 1777 |
+
{
|
| 1778 |
+
"epoch": 1.16,
|
| 1779 |
+
"grad_norm": 4.642292499542236,
|
| 1780 |
+
"learning_rate": 1.5356978350990326e-05,
|
| 1781 |
+
"loss": 0.043,
|
| 1782 |
+
"step": 2520
|
| 1783 |
+
},
|
| 1784 |
+
{
|
| 1785 |
+
"epoch": 1.17,
|
| 1786 |
+
"grad_norm": 0.02049732580780983,
|
| 1787 |
+
"learning_rate": 1.5338553661906955e-05,
|
| 1788 |
+
"loss": 0.0726,
|
| 1789 |
+
"step": 2530
|
| 1790 |
+
},
|
| 1791 |
+
{
|
| 1792 |
+
"epoch": 1.17,
|
| 1793 |
+
"grad_norm": 2.948366641998291,
|
| 1794 |
+
"learning_rate": 1.5320128972823583e-05,
|
| 1795 |
+
"loss": 0.1252,
|
| 1796 |
+
"step": 2540
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"epoch": 1.17,
|
| 1800 |
+
"grad_norm": 19.290773391723633,
|
| 1801 |
+
"learning_rate": 1.5301704283740212e-05,
|
| 1802 |
+
"loss": 0.0639,
|
| 1803 |
+
"step": 2550
|
| 1804 |
+
},
|
| 1805 |
+
{
|
| 1806 |
+
"epoch": 1.18,
|
| 1807 |
+
"grad_norm": 0.0013690482592210174,
|
| 1808 |
+
"learning_rate": 1.528327959465684e-05,
|
| 1809 |
+
"loss": 0.0147,
|
| 1810 |
+
"step": 2560
|
| 1811 |
+
},
|
| 1812 |
+
{
|
| 1813 |
+
"epoch": 1.18,
|
| 1814 |
+
"grad_norm": 5.165115833282471,
|
| 1815 |
+
"learning_rate": 1.526485490557347e-05,
|
| 1816 |
+
"loss": 0.0354,
|
| 1817 |
+
"step": 2570
|
| 1818 |
+
},
|
| 1819 |
+
{
|
| 1820 |
+
"epoch": 1.19,
|
| 1821 |
+
"grad_norm": 1.9532759189605713,
|
| 1822 |
+
"learning_rate": 1.5246430216490097e-05,
|
| 1823 |
+
"loss": 0.0395,
|
| 1824 |
+
"step": 2580
|
| 1825 |
+
},
|
| 1826 |
+
{
|
| 1827 |
+
"epoch": 1.19,
|
| 1828 |
+
"grad_norm": 0.008152857422828674,
|
| 1829 |
+
"learning_rate": 1.5228005527406726e-05,
|
| 1830 |
+
"loss": 0.0347,
|
| 1831 |
+
"step": 2590
|
| 1832 |
+
},
|
| 1833 |
+
{
|
| 1834 |
+
"epoch": 1.2,
|
| 1835 |
+
"grad_norm": 0.00506425928324461,
|
| 1836 |
+
"learning_rate": 1.5209580838323354e-05,
|
| 1837 |
+
"loss": 0.0894,
|
| 1838 |
+
"step": 2600
|
| 1839 |
+
},
|
| 1840 |
+
{
|
| 1841 |
+
"epoch": 1.2,
|
| 1842 |
+
"grad_norm": 2.3241424560546875,
|
| 1843 |
+
"learning_rate": 1.5191156149239983e-05,
|
| 1844 |
+
"loss": 0.0734,
|
| 1845 |
+
"step": 2610
|
| 1846 |
+
},
|
| 1847 |
+
{
|
| 1848 |
+
"epoch": 1.21,
|
| 1849 |
+
"grad_norm": 0.004548648837953806,
|
| 1850 |
+
"learning_rate": 1.5172731460156611e-05,
|
| 1851 |
+
"loss": 0.0877,
|
| 1852 |
+
"step": 2620
|
| 1853 |
+
},
|
| 1854 |
+
{
|
| 1855 |
+
"epoch": 1.21,
|
| 1856 |
+
"grad_norm": 0.3995460271835327,
|
| 1857 |
+
"learning_rate": 1.515430677107324e-05,
|
| 1858 |
+
"loss": 0.0235,
|
| 1859 |
+
"step": 2630
|
| 1860 |
+
},
|
| 1861 |
+
{
|
| 1862 |
+
"epoch": 1.22,
|
| 1863 |
+
"grad_norm": 0.08874616026878357,
|
| 1864 |
+
"learning_rate": 1.5135882081989868e-05,
|
| 1865 |
+
"loss": 0.0237,
|
| 1866 |
+
"step": 2640
|
| 1867 |
+
},
|
| 1868 |
+
{
|
| 1869 |
+
"epoch": 1.22,
|
| 1870 |
+
"grad_norm": 9.560503959655762,
|
| 1871 |
+
"learning_rate": 1.5117457392906496e-05,
|
| 1872 |
+
"loss": 0.0215,
|
| 1873 |
+
"step": 2650
|
| 1874 |
+
},
|
| 1875 |
+
{
|
| 1876 |
+
"epoch": 1.22,
|
| 1877 |
+
"grad_norm": 2.1006712913513184,
|
| 1878 |
+
"learning_rate": 1.5099032703823125e-05,
|
| 1879 |
+
"loss": 0.1345,
|
| 1880 |
+
"step": 2660
|
| 1881 |
+
},
|
| 1882 |
+
{
|
| 1883 |
+
"epoch": 1.23,
|
| 1884 |
+
"grad_norm": 0.0892830416560173,
|
| 1885 |
+
"learning_rate": 1.5080608014739753e-05,
|
| 1886 |
+
"loss": 0.0803,
|
| 1887 |
+
"step": 2670
|
| 1888 |
+
},
|
| 1889 |
+
{
|
| 1890 |
+
"epoch": 1.23,
|
| 1891 |
+
"grad_norm": 1.4158849716186523,
|
| 1892 |
+
"learning_rate": 1.506218332565638e-05,
|
| 1893 |
+
"loss": 0.0036,
|
| 1894 |
+
"step": 2680
|
| 1895 |
+
},
|
| 1896 |
+
{
|
| 1897 |
+
"epoch": 1.24,
|
| 1898 |
+
"grad_norm": 0.02232968993484974,
|
| 1899 |
+
"learning_rate": 1.5043758636573009e-05,
|
| 1900 |
+
"loss": 0.0463,
|
| 1901 |
+
"step": 2690
|
| 1902 |
+
},
|
| 1903 |
+
{
|
| 1904 |
+
"epoch": 1.24,
|
| 1905 |
+
"grad_norm": 24.494020462036133,
|
| 1906 |
+
"learning_rate": 1.5025333947489637e-05,
|
| 1907 |
+
"loss": 0.0955,
|
| 1908 |
+
"step": 2700
|
| 1909 |
+
},
|
| 1910 |
+
{
|
| 1911 |
+
"epoch": 1.25,
|
| 1912 |
+
"grad_norm": 3.047916889190674,
|
| 1913 |
+
"learning_rate": 1.5006909258406266e-05,
|
| 1914 |
+
"loss": 0.0047,
|
| 1915 |
+
"step": 2710
|
| 1916 |
+
},
|
| 1917 |
+
{
|
| 1918 |
+
"epoch": 1.25,
|
| 1919 |
+
"grad_norm": 0.0400189571082592,
|
| 1920 |
+
"learning_rate": 1.4988484569322894e-05,
|
| 1921 |
+
"loss": 0.0165,
|
| 1922 |
+
"step": 2720
|
| 1923 |
+
},
|
| 1924 |
+
{
|
| 1925 |
+
"epoch": 1.26,
|
| 1926 |
+
"grad_norm": 0.13102641701698303,
|
| 1927 |
+
"learning_rate": 1.4970059880239522e-05,
|
| 1928 |
+
"loss": 0.0074,
|
| 1929 |
+
"step": 2730
|
| 1930 |
+
},
|
| 1931 |
+
{
|
| 1932 |
+
"epoch": 1.26,
|
| 1933 |
+
"grad_norm": 2.2329273223876953,
|
| 1934 |
+
"learning_rate": 1.4951635191156151e-05,
|
| 1935 |
+
"loss": 0.0413,
|
| 1936 |
+
"step": 2740
|
| 1937 |
+
},
|
| 1938 |
+
{
|
| 1939 |
+
"epoch": 1.27,
|
| 1940 |
+
"grad_norm": 0.0005666582728736103,
|
| 1941 |
+
"learning_rate": 1.493321050207278e-05,
|
| 1942 |
+
"loss": 0.0428,
|
| 1943 |
+
"step": 2750
|
| 1944 |
+
},
|
| 1945 |
+
{
|
| 1946 |
+
"epoch": 1.27,
|
| 1947 |
+
"grad_norm": 0.00015929104120004922,
|
| 1948 |
+
"learning_rate": 1.4914785812989408e-05,
|
| 1949 |
+
"loss": 0.0038,
|
| 1950 |
+
"step": 2760
|
| 1951 |
+
},
|
| 1952 |
+
{
|
| 1953 |
+
"epoch": 1.28,
|
| 1954 |
+
"grad_norm": 14.374834060668945,
|
| 1955 |
+
"learning_rate": 1.4896361123906036e-05,
|
| 1956 |
+
"loss": 0.0859,
|
| 1957 |
+
"step": 2770
|
| 1958 |
+
},
|
| 1959 |
+
{
|
| 1960 |
+
"epoch": 1.28,
|
| 1961 |
+
"grad_norm": 0.16000565886497498,
|
| 1962 |
+
"learning_rate": 1.4877936434822665e-05,
|
| 1963 |
+
"loss": 0.0119,
|
| 1964 |
+
"step": 2780
|
| 1965 |
+
},
|
| 1966 |
+
{
|
| 1967 |
+
"epoch": 1.28,
|
| 1968 |
+
"grad_norm": 1.4262840747833252,
|
| 1969 |
+
"learning_rate": 1.4859511745739292e-05,
|
| 1970 |
+
"loss": 0.0953,
|
| 1971 |
+
"step": 2790
|
| 1972 |
+
},
|
| 1973 |
+
{
|
| 1974 |
+
"epoch": 1.29,
|
| 1975 |
+
"grad_norm": 0.778024435043335,
|
| 1976 |
+
"learning_rate": 1.484108705665592e-05,
|
| 1977 |
+
"loss": 0.0802,
|
| 1978 |
+
"step": 2800
|
| 1979 |
+
},
|
| 1980 |
+
{
|
| 1981 |
+
"epoch": 1.29,
|
| 1982 |
+
"grad_norm": 2.633740186691284,
|
| 1983 |
+
"learning_rate": 1.4822662367572548e-05,
|
| 1984 |
+
"loss": 0.0826,
|
| 1985 |
+
"step": 2810
|
| 1986 |
+
},
|
| 1987 |
+
{
|
| 1988 |
+
"epoch": 1.3,
|
| 1989 |
+
"grad_norm": 1.6573383808135986,
|
| 1990 |
+
"learning_rate": 1.4804237678489177e-05,
|
| 1991 |
+
"loss": 0.0048,
|
| 1992 |
+
"step": 2820
|
| 1993 |
+
},
|
| 1994 |
+
{
|
| 1995 |
+
"epoch": 1.3,
|
| 1996 |
+
"grad_norm": 0.1440638154745102,
|
| 1997 |
+
"learning_rate": 1.4785812989405805e-05,
|
| 1998 |
+
"loss": 0.0174,
|
| 1999 |
+
"step": 2830
|
| 2000 |
+
},
|
| 2001 |
+
{
|
| 2002 |
+
"epoch": 1.31,
|
| 2003 |
+
"grad_norm": 2.362022876739502,
|
| 2004 |
+
"learning_rate": 1.4767388300322434e-05,
|
| 2005 |
+
"loss": 0.0166,
|
| 2006 |
+
"step": 2840
|
| 2007 |
+
},
|
| 2008 |
+
{
|
| 2009 |
+
"epoch": 1.31,
|
| 2010 |
+
"grad_norm": 0.39531031250953674,
|
| 2011 |
+
"learning_rate": 1.4748963611239062e-05,
|
| 2012 |
+
"loss": 0.0311,
|
| 2013 |
+
"step": 2850
|
| 2014 |
+
},
|
| 2015 |
+
{
|
| 2016 |
+
"epoch": 1.32,
|
| 2017 |
+
"grad_norm": 4.888799667358398,
|
| 2018 |
+
"learning_rate": 1.473053892215569e-05,
|
| 2019 |
+
"loss": 0.0182,
|
| 2020 |
+
"step": 2860
|
| 2021 |
+
},
|
| 2022 |
+
{
|
| 2023 |
+
"epoch": 1.32,
|
| 2024 |
+
"grad_norm": 0.0002681002370081842,
|
| 2025 |
+
"learning_rate": 1.471211423307232e-05,
|
| 2026 |
+
"loss": 0.0154,
|
| 2027 |
+
"step": 2870
|
| 2028 |
+
},
|
| 2029 |
+
{
|
| 2030 |
+
"epoch": 1.33,
|
| 2031 |
+
"grad_norm": 11.085668563842773,
|
| 2032 |
+
"learning_rate": 1.4693689543988948e-05,
|
| 2033 |
+
"loss": 0.0195,
|
| 2034 |
+
"step": 2880
|
| 2035 |
+
},
|
| 2036 |
+
{
|
| 2037 |
+
"epoch": 1.33,
|
| 2038 |
+
"grad_norm": 0.005156415514647961,
|
| 2039 |
+
"learning_rate": 1.4675264854905576e-05,
|
| 2040 |
+
"loss": 0.0209,
|
| 2041 |
+
"step": 2890
|
| 2042 |
+
},
|
| 2043 |
+
{
|
| 2044 |
+
"epoch": 1.34,
|
| 2045 |
+
"grad_norm": 33.43428421020508,
|
| 2046 |
+
"learning_rate": 1.4656840165822203e-05,
|
| 2047 |
+
"loss": 0.0296,
|
| 2048 |
+
"step": 2900
|
| 2049 |
+
},
|
| 2050 |
+
{
|
| 2051 |
+
"epoch": 1.34,
|
| 2052 |
+
"grad_norm": 0.05032246932387352,
|
| 2053 |
+
"learning_rate": 1.4638415476738831e-05,
|
| 2054 |
+
"loss": 0.0257,
|
| 2055 |
+
"step": 2910
|
| 2056 |
+
},
|
| 2057 |
+
{
|
| 2058 |
+
"epoch": 1.34,
|
| 2059 |
+
"grad_norm": 1.5708962678909302,
|
| 2060 |
+
"learning_rate": 1.461999078765546e-05,
|
| 2061 |
+
"loss": 0.0075,
|
| 2062 |
+
"step": 2920
|
| 2063 |
+
},
|
| 2064 |
+
{
|
| 2065 |
+
"epoch": 1.35,
|
| 2066 |
+
"grad_norm": 0.3271735906600952,
|
| 2067 |
+
"learning_rate": 1.4601566098572088e-05,
|
| 2068 |
+
"loss": 0.0874,
|
| 2069 |
+
"step": 2930
|
| 2070 |
+
},
|
| 2071 |
+
{
|
| 2072 |
+
"epoch": 1.35,
|
| 2073 |
+
"grad_norm": 4.142122745513916,
|
| 2074 |
+
"learning_rate": 1.4583141409488717e-05,
|
| 2075 |
+
"loss": 0.0445,
|
| 2076 |
+
"step": 2940
|
| 2077 |
+
},
|
| 2078 |
+
{
|
| 2079 |
+
"epoch": 1.36,
|
| 2080 |
+
"grad_norm": 0.00015666706894990057,
|
| 2081 |
+
"learning_rate": 1.4564716720405345e-05,
|
| 2082 |
+
"loss": 0.0558,
|
| 2083 |
+
"step": 2950
|
| 2084 |
+
},
|
| 2085 |
+
{
|
| 2086 |
+
"epoch": 1.36,
|
| 2087 |
+
"grad_norm": 2.2890191078186035,
|
| 2088 |
+
"learning_rate": 1.4546292031321974e-05,
|
| 2089 |
+
"loss": 0.0591,
|
| 2090 |
+
"step": 2960
|
| 2091 |
+
},
|
| 2092 |
+
{
|
| 2093 |
+
"epoch": 1.37,
|
| 2094 |
+
"grad_norm": 21.041645050048828,
|
| 2095 |
+
"learning_rate": 1.4527867342238602e-05,
|
| 2096 |
+
"loss": 0.0377,
|
| 2097 |
+
"step": 2970
|
| 2098 |
+
},
|
| 2099 |
+
{
|
| 2100 |
+
"epoch": 1.37,
|
| 2101 |
+
"grad_norm": 0.0033177861478179693,
|
| 2102 |
+
"learning_rate": 1.450944265315523e-05,
|
| 2103 |
+
"loss": 0.0422,
|
| 2104 |
+
"step": 2980
|
| 2105 |
+
},
|
| 2106 |
+
{
|
| 2107 |
+
"epoch": 1.38,
|
| 2108 |
+
"grad_norm": 0.44971880316734314,
|
| 2109 |
+
"learning_rate": 1.4491017964071859e-05,
|
| 2110 |
+
"loss": 0.1966,
|
| 2111 |
+
"step": 2990
|
| 2112 |
+
},
|
| 2113 |
+
{
|
| 2114 |
+
"epoch": 1.38,
|
| 2115 |
+
"grad_norm": 0.04920297861099243,
|
| 2116 |
+
"learning_rate": 1.4472593274988488e-05,
|
| 2117 |
+
"loss": 0.0522,
|
| 2118 |
+
"step": 3000
|
| 2119 |
+
},
|
| 2120 |
+
{
|
| 2121 |
+
"epoch": 1.39,
|
| 2122 |
+
"grad_norm": 19.32337188720703,
|
| 2123 |
+
"learning_rate": 1.4454168585905114e-05,
|
| 2124 |
+
"loss": 0.1246,
|
| 2125 |
+
"step": 3010
|
| 2126 |
+
},
|
| 2127 |
+
{
|
| 2128 |
+
"epoch": 1.39,
|
| 2129 |
+
"grad_norm": 0.5419002771377563,
|
| 2130 |
+
"learning_rate": 1.4435743896821743e-05,
|
| 2131 |
+
"loss": 0.0285,
|
| 2132 |
+
"step": 3020
|
| 2133 |
+
},
|
| 2134 |
+
{
|
| 2135 |
+
"epoch": 1.4,
|
| 2136 |
+
"grad_norm": 2.8589115142822266,
|
| 2137 |
+
"learning_rate": 1.4417319207738371e-05,
|
| 2138 |
+
"loss": 0.0038,
|
| 2139 |
+
"step": 3030
|
| 2140 |
+
},
|
| 2141 |
+
{
|
| 2142 |
+
"epoch": 1.4,
|
| 2143 |
+
"grad_norm": 0.0985371470451355,
|
| 2144 |
+
"learning_rate": 1.4398894518654998e-05,
|
| 2145 |
+
"loss": 0.0533,
|
| 2146 |
+
"step": 3040
|
| 2147 |
+
},
|
| 2148 |
+
{
|
| 2149 |
+
"epoch": 1.4,
|
| 2150 |
+
"grad_norm": 3.628220319747925,
|
| 2151 |
+
"learning_rate": 1.4380469829571626e-05,
|
| 2152 |
+
"loss": 0.1248,
|
| 2153 |
+
"step": 3050
|
| 2154 |
+
},
|
| 2155 |
+
{
|
| 2156 |
+
"epoch": 1.41,
|
| 2157 |
+
"grad_norm": 0.026494158431887627,
|
| 2158 |
+
"learning_rate": 1.4362045140488255e-05,
|
| 2159 |
+
"loss": 0.0122,
|
| 2160 |
+
"step": 3060
|
| 2161 |
+
},
|
| 2162 |
+
{
|
| 2163 |
+
"epoch": 1.41,
|
| 2164 |
+
"grad_norm": 0.029013484716415405,
|
| 2165 |
+
"learning_rate": 1.4343620451404883e-05,
|
| 2166 |
+
"loss": 0.0884,
|
| 2167 |
+
"step": 3070
|
| 2168 |
+
},
|
| 2169 |
+
{
|
| 2170 |
+
"epoch": 1.42,
|
| 2171 |
+
"grad_norm": 0.5892958641052246,
|
| 2172 |
+
"learning_rate": 1.432519576232151e-05,
|
| 2173 |
+
"loss": 0.0921,
|
| 2174 |
+
"step": 3080
|
| 2175 |
+
},
|
| 2176 |
+
{
|
| 2177 |
+
"epoch": 1.42,
|
| 2178 |
+
"grad_norm": 28.236888885498047,
|
| 2179 |
+
"learning_rate": 1.4306771073238139e-05,
|
| 2180 |
+
"loss": 0.0536,
|
| 2181 |
+
"step": 3090
|
| 2182 |
+
},
|
| 2183 |
+
{
|
| 2184 |
+
"epoch": 1.43,
|
| 2185 |
+
"grad_norm": 8.166137558873743e-05,
|
| 2186 |
+
"learning_rate": 1.4288346384154767e-05,
|
| 2187 |
+
"loss": 0.0255,
|
| 2188 |
+
"step": 3100
|
| 2189 |
+
},
|
| 2190 |
+
{
|
| 2191 |
+
"epoch": 1.43,
|
| 2192 |
+
"grad_norm": 0.002859271364286542,
|
| 2193 |
+
"learning_rate": 1.4269921695071396e-05,
|
| 2194 |
+
"loss": 0.0212,
|
| 2195 |
+
"step": 3110
|
| 2196 |
+
},
|
| 2197 |
+
{
|
| 2198 |
+
"epoch": 1.44,
|
| 2199 |
+
"grad_norm": 12.254012107849121,
|
| 2200 |
+
"learning_rate": 1.4251497005988024e-05,
|
| 2201 |
+
"loss": 0.0369,
|
| 2202 |
+
"step": 3120
|
| 2203 |
+
},
|
| 2204 |
+
{
|
| 2205 |
+
"epoch": 1.44,
|
| 2206 |
+
"grad_norm": 0.5498998761177063,
|
| 2207 |
+
"learning_rate": 1.4233072316904652e-05,
|
| 2208 |
+
"loss": 0.072,
|
| 2209 |
+
"step": 3130
|
| 2210 |
+
},
|
| 2211 |
+
{
|
| 2212 |
+
"epoch": 1.45,
|
| 2213 |
+
"grad_norm": 0.18189552426338196,
|
| 2214 |
+
"learning_rate": 1.4214647627821281e-05,
|
| 2215 |
+
"loss": 0.0949,
|
| 2216 |
+
"step": 3140
|
| 2217 |
+
},
|
| 2218 |
+
{
|
| 2219 |
+
"epoch": 1.45,
|
| 2220 |
+
"grad_norm": 0.14849336445331573,
|
| 2221 |
+
"learning_rate": 1.419622293873791e-05,
|
| 2222 |
+
"loss": 0.0165,
|
| 2223 |
+
"step": 3150
|
| 2224 |
+
},
|
| 2225 |
+
{
|
| 2226 |
+
"epoch": 1.46,
|
| 2227 |
+
"grad_norm": 0.083045095205307,
|
| 2228 |
+
"learning_rate": 1.4177798249654538e-05,
|
| 2229 |
+
"loss": 0.0144,
|
| 2230 |
+
"step": 3160
|
| 2231 |
+
},
|
| 2232 |
+
{
|
| 2233 |
+
"epoch": 1.46,
|
| 2234 |
+
"grad_norm": 7.960810035001487e-05,
|
| 2235 |
+
"learning_rate": 1.4159373560571166e-05,
|
| 2236 |
+
"loss": 0.0426,
|
| 2237 |
+
"step": 3170
|
| 2238 |
+
},
|
| 2239 |
+
{
|
| 2240 |
+
"epoch": 1.46,
|
| 2241 |
+
"grad_norm": 1.107560634613037,
|
| 2242 |
+
"learning_rate": 1.4140948871487795e-05,
|
| 2243 |
+
"loss": 0.0132,
|
| 2244 |
+
"step": 3180
|
| 2245 |
+
},
|
| 2246 |
+
{
|
| 2247 |
+
"epoch": 1.47,
|
| 2248 |
+
"grad_norm": 2.026191234588623,
|
| 2249 |
+
"learning_rate": 1.4122524182404422e-05,
|
| 2250 |
+
"loss": 0.0026,
|
| 2251 |
+
"step": 3190
|
| 2252 |
+
},
|
| 2253 |
+
{
|
| 2254 |
+
"epoch": 1.47,
|
| 2255 |
+
"grad_norm": 0.006679296027868986,
|
| 2256 |
+
"learning_rate": 1.410409949332105e-05,
|
| 2257 |
+
"loss": 0.0645,
|
| 2258 |
+
"step": 3200
|
| 2259 |
+
},
|
| 2260 |
+
{
|
| 2261 |
+
"epoch": 1.48,
|
| 2262 |
+
"grad_norm": 9.621331214904785,
|
| 2263 |
+
"learning_rate": 1.4085674804237678e-05,
|
| 2264 |
+
"loss": 0.0326,
|
| 2265 |
+
"step": 3210
|
| 2266 |
+
},
|
| 2267 |
+
{
|
| 2268 |
+
"epoch": 1.48,
|
| 2269 |
+
"grad_norm": 0.31459417939186096,
|
| 2270 |
+
"learning_rate": 1.4067250115154307e-05,
|
| 2271 |
+
"loss": 0.2026,
|
| 2272 |
+
"step": 3220
|
| 2273 |
+
},
|
| 2274 |
+
{
|
| 2275 |
+
"epoch": 1.49,
|
| 2276 |
+
"grad_norm": 0.00020696005958598107,
|
| 2277 |
+
"learning_rate": 1.4048825426070935e-05,
|
| 2278 |
+
"loss": 0.0391,
|
| 2279 |
+
"step": 3230
|
| 2280 |
+
},
|
| 2281 |
+
{
|
| 2282 |
+
"epoch": 1.49,
|
| 2283 |
+
"grad_norm": 0.056342847645282745,
|
| 2284 |
+
"learning_rate": 1.4030400736987564e-05,
|
| 2285 |
+
"loss": 0.0326,
|
| 2286 |
+
"step": 3240
|
| 2287 |
+
},
|
| 2288 |
+
{
|
| 2289 |
+
"epoch": 1.5,
|
| 2290 |
+
"grad_norm": 19.293376922607422,
|
| 2291 |
+
"learning_rate": 1.4011976047904192e-05,
|
| 2292 |
+
"loss": 0.2364,
|
| 2293 |
+
"step": 3250
|
| 2294 |
+
},
|
| 2295 |
+
{
|
| 2296 |
+
"epoch": 1.5,
|
| 2297 |
+
"grad_norm": 0.3475209176540375,
|
| 2298 |
+
"learning_rate": 1.399355135882082e-05,
|
| 2299 |
+
"loss": 0.0665,
|
| 2300 |
+
"step": 3260
|
| 2301 |
+
},
|
| 2302 |
+
{
|
| 2303 |
+
"epoch": 1.51,
|
| 2304 |
+
"grad_norm": 0.09548962861299515,
|
| 2305 |
+
"learning_rate": 1.397512666973745e-05,
|
| 2306 |
+
"loss": 0.0353,
|
| 2307 |
+
"step": 3270
|
| 2308 |
+
},
|
| 2309 |
+
{
|
| 2310 |
+
"epoch": 1.51,
|
| 2311 |
+
"grad_norm": 1.0094457864761353,
|
| 2312 |
+
"learning_rate": 1.3956701980654078e-05,
|
| 2313 |
+
"loss": 0.0093,
|
| 2314 |
+
"step": 3280
|
| 2315 |
+
},
|
| 2316 |
+
{
|
| 2317 |
+
"epoch": 1.52,
|
| 2318 |
+
"grad_norm": 1.0635191202163696,
|
| 2319 |
+
"learning_rate": 1.3938277291570706e-05,
|
| 2320 |
+
"loss": 0.0182,
|
| 2321 |
+
"step": 3290
|
| 2322 |
+
},
|
| 2323 |
+
{
|
| 2324 |
+
"epoch": 1.52,
|
| 2325 |
+
"grad_norm": 0.19485218822956085,
|
| 2326 |
+
"learning_rate": 1.3919852602487333e-05,
|
| 2327 |
+
"loss": 0.0538,
|
| 2328 |
+
"step": 3300
|
| 2329 |
+
},
|
| 2330 |
+
{
|
| 2331 |
+
"epoch": 1.52,
|
| 2332 |
+
"grad_norm": 0.00029300845926627517,
|
| 2333 |
+
"learning_rate": 1.3901427913403961e-05,
|
| 2334 |
+
"loss": 0.0604,
|
| 2335 |
+
"step": 3310
|
| 2336 |
+
},
|
| 2337 |
+
{
|
| 2338 |
+
"epoch": 1.53,
|
| 2339 |
+
"grad_norm": 0.07005924731492996,
|
| 2340 |
+
"learning_rate": 1.388300322432059e-05,
|
| 2341 |
+
"loss": 0.0293,
|
| 2342 |
+
"step": 3320
|
| 2343 |
+
},
|
| 2344 |
+
{
|
| 2345 |
+
"epoch": 1.53,
|
| 2346 |
+
"grad_norm": 49.332820892333984,
|
| 2347 |
+
"learning_rate": 1.3864578535237218e-05,
|
| 2348 |
+
"loss": 0.1365,
|
| 2349 |
+
"step": 3330
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"epoch": 1.54,
|
| 2353 |
+
"grad_norm": 0.05095929279923439,
|
| 2354 |
+
"learning_rate": 1.3846153846153847e-05,
|
| 2355 |
+
"loss": 0.0251,
|
| 2356 |
+
"step": 3340
|
| 2357 |
+
},
|
| 2358 |
+
{
|
| 2359 |
+
"epoch": 1.54,
|
| 2360 |
+
"grad_norm": 1.2569241523742676,
|
| 2361 |
+
"learning_rate": 1.3827729157070475e-05,
|
| 2362 |
+
"loss": 0.0085,
|
| 2363 |
+
"step": 3350
|
| 2364 |
+
},
|
| 2365 |
+
{
|
| 2366 |
+
"epoch": 1.55,
|
| 2367 |
+
"grad_norm": 0.24515914916992188,
|
| 2368 |
+
"learning_rate": 1.3809304467987104e-05,
|
| 2369 |
+
"loss": 0.015,
|
| 2370 |
+
"step": 3360
|
| 2371 |
+
},
|
| 2372 |
+
{
|
| 2373 |
+
"epoch": 1.55,
|
| 2374 |
+
"grad_norm": 0.6555165648460388,
|
| 2375 |
+
"learning_rate": 1.3790879778903732e-05,
|
| 2376 |
+
"loss": 0.0176,
|
| 2377 |
+
"step": 3370
|
| 2378 |
+
},
|
| 2379 |
+
{
|
| 2380 |
+
"epoch": 1.56,
|
| 2381 |
+
"grad_norm": 0.000826065253932029,
|
| 2382 |
+
"learning_rate": 1.377245508982036e-05,
|
| 2383 |
+
"loss": 0.029,
|
| 2384 |
+
"step": 3380
|
| 2385 |
+
},
|
| 2386 |
+
{
|
| 2387 |
+
"epoch": 1.56,
|
| 2388 |
+
"grad_norm": 0.0031832808163017035,
|
| 2389 |
+
"learning_rate": 1.3754030400736989e-05,
|
| 2390 |
+
"loss": 0.0041,
|
| 2391 |
+
"step": 3390
|
| 2392 |
+
},
|
| 2393 |
+
{
|
| 2394 |
+
"epoch": 1.57,
|
| 2395 |
+
"grad_norm": 5.125019550323486,
|
| 2396 |
+
"learning_rate": 1.3735605711653618e-05,
|
| 2397 |
+
"loss": 0.0324,
|
| 2398 |
+
"step": 3400
|
| 2399 |
+
},
|
| 2400 |
+
{
|
| 2401 |
+
"epoch": 1.57,
|
| 2402 |
+
"grad_norm": 0.3939912021160126,
|
| 2403 |
+
"learning_rate": 1.3717181022570244e-05,
|
| 2404 |
+
"loss": 0.0572,
|
| 2405 |
+
"step": 3410
|
| 2406 |
+
},
|
| 2407 |
+
{
|
| 2408 |
+
"epoch": 1.57,
|
| 2409 |
+
"grad_norm": 0.043806612491607666,
|
| 2410 |
+
"learning_rate": 1.3698756333486873e-05,
|
| 2411 |
+
"loss": 0.0267,
|
| 2412 |
+
"step": 3420
|
| 2413 |
+
},
|
| 2414 |
+
{
|
| 2415 |
+
"epoch": 1.58,
|
| 2416 |
+
"grad_norm": 0.6995906233787537,
|
| 2417 |
+
"learning_rate": 1.3680331644403501e-05,
|
| 2418 |
+
"loss": 0.0042,
|
| 2419 |
+
"step": 3430
|
| 2420 |
+
},
|
| 2421 |
+
{
|
| 2422 |
+
"epoch": 1.58,
|
| 2423 |
+
"grad_norm": 13.469160079956055,
|
| 2424 |
+
"learning_rate": 1.366190695532013e-05,
|
| 2425 |
+
"loss": 0.0547,
|
| 2426 |
+
"step": 3440
|
| 2427 |
+
},
|
| 2428 |
+
{
|
| 2429 |
+
"epoch": 1.59,
|
| 2430 |
+
"grad_norm": 2.43892502784729,
|
| 2431 |
+
"learning_rate": 1.3643482266236758e-05,
|
| 2432 |
+
"loss": 0.0217,
|
| 2433 |
+
"step": 3450
|
| 2434 |
+
},
|
| 2435 |
+
{
|
| 2436 |
+
"epoch": 1.59,
|
| 2437 |
+
"grad_norm": 0.11497724056243896,
|
| 2438 |
+
"learning_rate": 1.3625057577153387e-05,
|
| 2439 |
+
"loss": 0.0732,
|
| 2440 |
+
"step": 3460
|
| 2441 |
+
},
|
| 2442 |
+
{
|
| 2443 |
+
"epoch": 1.6,
|
| 2444 |
+
"grad_norm": 37.20389938354492,
|
| 2445 |
+
"learning_rate": 1.3606632888070015e-05,
|
| 2446 |
+
"loss": 0.0803,
|
| 2447 |
+
"step": 3470
|
| 2448 |
+
},
|
| 2449 |
+
{
|
| 2450 |
+
"epoch": 1.6,
|
| 2451 |
+
"grad_norm": 0.26896199584007263,
|
| 2452 |
+
"learning_rate": 1.3588208198986644e-05,
|
| 2453 |
+
"loss": 0.001,
|
| 2454 |
+
"step": 3480
|
| 2455 |
+
},
|
| 2456 |
+
{
|
| 2457 |
+
"epoch": 1.61,
|
| 2458 |
+
"grad_norm": 1.2094647900084965e-05,
|
| 2459 |
+
"learning_rate": 1.3569783509903272e-05,
|
| 2460 |
+
"loss": 0.0481,
|
| 2461 |
+
"step": 3490
|
| 2462 |
+
},
|
| 2463 |
+
{
|
| 2464 |
+
"epoch": 1.61,
|
| 2465 |
+
"grad_norm": 5.7832192396745086e-05,
|
| 2466 |
+
"learning_rate": 1.35513588208199e-05,
|
| 2467 |
+
"loss": 0.0118,
|
| 2468 |
+
"step": 3500
|
| 2469 |
+
},
|
| 2470 |
+
{
|
| 2471 |
+
"epoch": 1.62,
|
| 2472 |
+
"grad_norm": 1.7425185433239676e-05,
|
| 2473 |
+
"learning_rate": 1.3532934131736529e-05,
|
| 2474 |
+
"loss": 0.0005,
|
| 2475 |
+
"step": 3510
|
| 2476 |
+
},
|
| 2477 |
+
{
|
| 2478 |
+
"epoch": 1.62,
|
| 2479 |
+
"grad_norm": 0.2973109185695648,
|
| 2480 |
+
"learning_rate": 1.3514509442653156e-05,
|
| 2481 |
+
"loss": 0.1256,
|
| 2482 |
+
"step": 3520
|
| 2483 |
+
},
|
| 2484 |
+
{
|
| 2485 |
+
"epoch": 1.63,
|
| 2486 |
+
"grad_norm": 0.008225671947002411,
|
| 2487 |
+
"learning_rate": 1.3496084753569784e-05,
|
| 2488 |
+
"loss": 0.0481,
|
| 2489 |
+
"step": 3530
|
| 2490 |
+
},
|
| 2491 |
+
{
|
| 2492 |
+
"epoch": 1.63,
|
| 2493 |
+
"grad_norm": 12.407868385314941,
|
| 2494 |
+
"learning_rate": 1.3477660064486413e-05,
|
| 2495 |
+
"loss": 0.0758,
|
| 2496 |
+
"step": 3540
|
| 2497 |
+
},
|
| 2498 |
+
{
|
| 2499 |
+
"epoch": 1.63,
|
| 2500 |
+
"grad_norm": 0.29406219720840454,
|
| 2501 |
+
"learning_rate": 1.3459235375403041e-05,
|
| 2502 |
+
"loss": 0.2475,
|
| 2503 |
+
"step": 3550
|
| 2504 |
+
},
|
| 2505 |
+
{
|
| 2506 |
+
"epoch": 1.64,
|
| 2507 |
+
"grad_norm": 0.6741734147071838,
|
| 2508 |
+
"learning_rate": 1.344081068631967e-05,
|
| 2509 |
+
"loss": 0.0266,
|
| 2510 |
+
"step": 3560
|
| 2511 |
+
},
|
| 2512 |
+
{
|
| 2513 |
+
"epoch": 1.64,
|
| 2514 |
+
"grad_norm": 0.002058387268334627,
|
| 2515 |
+
"learning_rate": 1.3422385997236298e-05,
|
| 2516 |
+
"loss": 0.0387,
|
| 2517 |
+
"step": 3570
|
| 2518 |
+
},
|
| 2519 |
+
{
|
| 2520 |
+
"epoch": 1.65,
|
| 2521 |
+
"grad_norm": 1.848854481067974e-05,
|
| 2522 |
+
"learning_rate": 1.3403961308152926e-05,
|
| 2523 |
+
"loss": 0.0057,
|
| 2524 |
+
"step": 3580
|
| 2525 |
+
},
|
| 2526 |
+
{
|
| 2527 |
+
"epoch": 1.65,
|
| 2528 |
+
"grad_norm": 0.13114599883556366,
|
| 2529 |
+
"learning_rate": 1.3385536619069555e-05,
|
| 2530 |
+
"loss": 0.0251,
|
| 2531 |
+
"step": 3590
|
| 2532 |
+
},
|
| 2533 |
+
{
|
| 2534 |
+
"epoch": 1.66,
|
| 2535 |
+
"grad_norm": 3.7949016094207764,
|
| 2536 |
+
"learning_rate": 1.3367111929986183e-05,
|
| 2537 |
+
"loss": 0.0172,
|
| 2538 |
+
"step": 3600
|
| 2539 |
+
},
|
| 2540 |
+
{
|
| 2541 |
+
"epoch": 1.66,
|
| 2542 |
+
"grad_norm": 0.050431422889232635,
|
| 2543 |
+
"learning_rate": 1.3348687240902812e-05,
|
| 2544 |
+
"loss": 0.075,
|
| 2545 |
+
"step": 3610
|
| 2546 |
+
},
|
| 2547 |
+
{
|
| 2548 |
+
"epoch": 1.67,
|
| 2549 |
+
"grad_norm": 6.660561561584473,
|
| 2550 |
+
"learning_rate": 1.333026255181944e-05,
|
| 2551 |
+
"loss": 0.0673,
|
| 2552 |
+
"step": 3620
|
| 2553 |
+
},
|
| 2554 |
+
{
|
| 2555 |
+
"epoch": 1.67,
|
| 2556 |
+
"grad_norm": 3.854702949523926,
|
| 2557 |
+
"learning_rate": 1.3311837862736067e-05,
|
| 2558 |
+
"loss": 0.0402,
|
| 2559 |
+
"step": 3630
|
| 2560 |
+
},
|
| 2561 |
+
{
|
| 2562 |
+
"epoch": 1.68,
|
| 2563 |
+
"grad_norm": 0.07829450815916061,
|
| 2564 |
+
"learning_rate": 1.3293413173652696e-05,
|
| 2565 |
+
"loss": 0.0073,
|
| 2566 |
+
"step": 3640
|
| 2567 |
+
},
|
| 2568 |
+
{
|
| 2569 |
+
"epoch": 1.68,
|
| 2570 |
+
"grad_norm": 19.043968200683594,
|
| 2571 |
+
"learning_rate": 1.3274988484569324e-05,
|
| 2572 |
+
"loss": 0.0787,
|
| 2573 |
+
"step": 3650
|
| 2574 |
+
},
|
| 2575 |
+
{
|
| 2576 |
+
"epoch": 1.69,
|
| 2577 |
+
"grad_norm": 0.08531194925308228,
|
| 2578 |
+
"learning_rate": 1.3256563795485952e-05,
|
| 2579 |
+
"loss": 0.1169,
|
| 2580 |
+
"step": 3660
|
| 2581 |
+
},
|
| 2582 |
+
{
|
| 2583 |
+
"epoch": 1.69,
|
| 2584 |
+
"grad_norm": 0.03738537058234215,
|
| 2585 |
+
"learning_rate": 1.3238139106402581e-05,
|
| 2586 |
+
"loss": 0.0595,
|
| 2587 |
+
"step": 3670
|
| 2588 |
+
},
|
| 2589 |
+
{
|
| 2590 |
+
"epoch": 1.69,
|
| 2591 |
+
"grad_norm": 0.20163363218307495,
|
| 2592 |
+
"learning_rate": 1.321971441731921e-05,
|
| 2593 |
+
"loss": 0.0383,
|
| 2594 |
+
"step": 3680
|
| 2595 |
+
},
|
| 2596 |
+
{
|
| 2597 |
+
"epoch": 1.7,
|
| 2598 |
+
"grad_norm": 0.0277603380382061,
|
| 2599 |
+
"learning_rate": 1.3201289728235838e-05,
|
| 2600 |
+
"loss": 0.0264,
|
| 2601 |
+
"step": 3690
|
| 2602 |
+
},
|
| 2603 |
+
{
|
| 2604 |
+
"epoch": 1.7,
|
| 2605 |
+
"grad_norm": 0.0004097486671525985,
|
| 2606 |
+
"learning_rate": 1.3182865039152466e-05,
|
| 2607 |
+
"loss": 0.1356,
|
| 2608 |
+
"step": 3700
|
| 2609 |
+
},
|
| 2610 |
+
{
|
| 2611 |
+
"epoch": 1.71,
|
| 2612 |
+
"grad_norm": 4.949276924133301,
|
| 2613 |
+
"learning_rate": 1.3164440350069095e-05,
|
| 2614 |
+
"loss": 0.101,
|
| 2615 |
+
"step": 3710
|
| 2616 |
+
},
|
| 2617 |
+
{
|
| 2618 |
+
"epoch": 1.71,
|
| 2619 |
+
"grad_norm": 0.34668001532554626,
|
| 2620 |
+
"learning_rate": 1.3146015660985723e-05,
|
| 2621 |
+
"loss": 0.0631,
|
| 2622 |
+
"step": 3720
|
| 2623 |
+
},
|
| 2624 |
+
{
|
| 2625 |
+
"epoch": 1.72,
|
| 2626 |
+
"grad_norm": 35.14409637451172,
|
| 2627 |
+
"learning_rate": 1.3127590971902352e-05,
|
| 2628 |
+
"loss": 0.0994,
|
| 2629 |
+
"step": 3730
|
| 2630 |
+
},
|
| 2631 |
+
{
|
| 2632 |
+
"epoch": 1.72,
|
| 2633 |
+
"grad_norm": 1.829344391822815,
|
| 2634 |
+
"learning_rate": 1.3109166282818978e-05,
|
| 2635 |
+
"loss": 0.0986,
|
| 2636 |
+
"step": 3740
|
| 2637 |
+
},
|
| 2638 |
+
{
|
| 2639 |
+
"epoch": 1.73,
|
| 2640 |
+
"grad_norm": 0.9819942712783813,
|
| 2641 |
+
"learning_rate": 1.3090741593735607e-05,
|
| 2642 |
+
"loss": 0.0028,
|
| 2643 |
+
"step": 3750
|
| 2644 |
+
},
|
| 2645 |
+
{
|
| 2646 |
+
"epoch": 1.73,
|
| 2647 |
+
"grad_norm": 0.08805956691503525,
|
| 2648 |
+
"learning_rate": 1.3072316904652235e-05,
|
| 2649 |
+
"loss": 0.1736,
|
| 2650 |
+
"step": 3760
|
| 2651 |
+
},
|
| 2652 |
+
{
|
| 2653 |
+
"epoch": 1.74,
|
| 2654 |
+
"grad_norm": 0.7243494987487793,
|
| 2655 |
+
"learning_rate": 1.3053892215568864e-05,
|
| 2656 |
+
"loss": 0.0723,
|
| 2657 |
+
"step": 3770
|
| 2658 |
+
},
|
| 2659 |
+
{
|
| 2660 |
+
"epoch": 1.74,
|
| 2661 |
+
"grad_norm": 0.01116830762475729,
|
| 2662 |
+
"learning_rate": 1.3035467526485492e-05,
|
| 2663 |
+
"loss": 0.0169,
|
| 2664 |
+
"step": 3780
|
| 2665 |
+
},
|
| 2666 |
+
{
|
| 2667 |
+
"epoch": 1.75,
|
| 2668 |
+
"grad_norm": 0.04502970352768898,
|
| 2669 |
+
"learning_rate": 1.301704283740212e-05,
|
| 2670 |
+
"loss": 0.0029,
|
| 2671 |
+
"step": 3790
|
| 2672 |
+
},
|
| 2673 |
+
{
|
| 2674 |
+
"epoch": 1.75,
|
| 2675 |
+
"grad_norm": 0.044204231351614,
|
| 2676 |
+
"learning_rate": 1.299861814831875e-05,
|
| 2677 |
+
"loss": 0.1873,
|
| 2678 |
+
"step": 3800
|
| 2679 |
+
},
|
| 2680 |
+
{
|
| 2681 |
+
"epoch": 1.75,
|
| 2682 |
+
"grad_norm": 0.00019235633953940123,
|
| 2683 |
+
"learning_rate": 1.2980193459235378e-05,
|
| 2684 |
+
"loss": 0.0303,
|
| 2685 |
+
"step": 3810
|
| 2686 |
+
},
|
| 2687 |
+
{
|
| 2688 |
+
"epoch": 1.76,
|
| 2689 |
+
"grad_norm": 0.03495045378804207,
|
| 2690 |
+
"learning_rate": 1.2961768770152006e-05,
|
| 2691 |
+
"loss": 0.1467,
|
| 2692 |
+
"step": 3820
|
| 2693 |
+
},
|
| 2694 |
+
{
|
| 2695 |
+
"epoch": 1.76,
|
| 2696 |
+
"grad_norm": 1.1287940740585327,
|
| 2697 |
+
"learning_rate": 1.2943344081068635e-05,
|
| 2698 |
+
"loss": 0.0127,
|
| 2699 |
+
"step": 3830
|
| 2700 |
+
},
|
| 2701 |
+
{
|
| 2702 |
+
"epoch": 1.77,
|
| 2703 |
+
"grad_norm": 2.7415771484375,
|
| 2704 |
+
"learning_rate": 1.2924919391985263e-05,
|
| 2705 |
+
"loss": 0.0939,
|
| 2706 |
+
"step": 3840
|
| 2707 |
+
},
|
| 2708 |
+
{
|
| 2709 |
+
"epoch": 1.77,
|
| 2710 |
+
"grad_norm": 4.2156548500061035,
|
| 2711 |
+
"learning_rate": 1.290649470290189e-05,
|
| 2712 |
+
"loss": 0.0088,
|
| 2713 |
+
"step": 3850
|
| 2714 |
+
},
|
| 2715 |
+
{
|
| 2716 |
+
"epoch": 1.78,
|
| 2717 |
+
"grad_norm": 3.3675363063812256,
|
| 2718 |
+
"learning_rate": 1.2888070013818518e-05,
|
| 2719 |
+
"loss": 0.1552,
|
| 2720 |
+
"step": 3860
|
| 2721 |
+
},
|
| 2722 |
+
{
|
| 2723 |
+
"epoch": 1.78,
|
| 2724 |
+
"grad_norm": 0.001867489772848785,
|
| 2725 |
+
"learning_rate": 1.2869645324735147e-05,
|
| 2726 |
+
"loss": 0.082,
|
| 2727 |
+
"step": 3870
|
| 2728 |
+
},
|
| 2729 |
+
{
|
| 2730 |
+
"epoch": 1.79,
|
| 2731 |
+
"grad_norm": 3.630979699664749e-05,
|
| 2732 |
+
"learning_rate": 1.2851220635651775e-05,
|
| 2733 |
+
"loss": 0.0293,
|
| 2734 |
+
"step": 3880
|
| 2735 |
+
},
|
| 2736 |
+
{
|
| 2737 |
+
"epoch": 1.79,
|
| 2738 |
+
"grad_norm": 3.107681914116256e-05,
|
| 2739 |
+
"learning_rate": 1.2832795946568404e-05,
|
| 2740 |
+
"loss": 0.131,
|
| 2741 |
+
"step": 3890
|
| 2742 |
+
},
|
| 2743 |
+
{
|
| 2744 |
+
"epoch": 1.8,
|
| 2745 |
+
"grad_norm": 6.011440277099609,
|
| 2746 |
+
"learning_rate": 1.2814371257485032e-05,
|
| 2747 |
+
"loss": 0.0605,
|
| 2748 |
+
"step": 3900
|
| 2749 |
+
},
|
| 2750 |
+
{
|
| 2751 |
+
"epoch": 1.8,
|
| 2752 |
+
"grad_norm": 0.0299521554261446,
|
| 2753 |
+
"learning_rate": 1.2795946568401659e-05,
|
| 2754 |
+
"loss": 0.0286,
|
| 2755 |
+
"step": 3910
|
| 2756 |
+
},
|
| 2757 |
+
{
|
| 2758 |
+
"epoch": 1.81,
|
| 2759 |
+
"grad_norm": 0.13250289857387543,
|
| 2760 |
+
"learning_rate": 1.2777521879318286e-05,
|
| 2761 |
+
"loss": 0.0198,
|
| 2762 |
+
"step": 3920
|
| 2763 |
+
},
|
| 2764 |
+
{
|
| 2765 |
+
"epoch": 1.81,
|
| 2766 |
+
"grad_norm": 0.02828848361968994,
|
| 2767 |
+
"learning_rate": 1.2759097190234914e-05,
|
| 2768 |
+
"loss": 0.0271,
|
| 2769 |
+
"step": 3930
|
| 2770 |
+
},
|
| 2771 |
+
{
|
| 2772 |
+
"epoch": 1.81,
|
| 2773 |
+
"grad_norm": 2.291139841079712,
|
| 2774 |
+
"learning_rate": 1.2740672501151543e-05,
|
| 2775 |
+
"loss": 0.0373,
|
| 2776 |
+
"step": 3940
|
| 2777 |
+
},
|
| 2778 |
+
{
|
| 2779 |
+
"epoch": 1.82,
|
| 2780 |
+
"grad_norm": 0.017028339207172394,
|
| 2781 |
+
"learning_rate": 1.2722247812068171e-05,
|
| 2782 |
+
"loss": 0.0178,
|
| 2783 |
+
"step": 3950
|
| 2784 |
+
},
|
| 2785 |
+
{
|
| 2786 |
+
"epoch": 1.82,
|
| 2787 |
+
"grad_norm": 0.7271348237991333,
|
| 2788 |
+
"learning_rate": 1.27038231229848e-05,
|
| 2789 |
+
"loss": 0.0701,
|
| 2790 |
+
"step": 3960
|
| 2791 |
+
},
|
| 2792 |
+
{
|
| 2793 |
+
"epoch": 1.83,
|
| 2794 |
+
"grad_norm": 0.0013385345228016376,
|
| 2795 |
+
"learning_rate": 1.2685398433901428e-05,
|
| 2796 |
+
"loss": 0.0861,
|
| 2797 |
+
"step": 3970
|
| 2798 |
+
},
|
| 2799 |
+
{
|
| 2800 |
+
"epoch": 1.83,
|
| 2801 |
+
"grad_norm": 14.219289779663086,
|
| 2802 |
+
"learning_rate": 1.2666973744818056e-05,
|
| 2803 |
+
"loss": 0.0253,
|
| 2804 |
+
"step": 3980
|
| 2805 |
+
},
|
| 2806 |
+
{
|
| 2807 |
+
"epoch": 1.84,
|
| 2808 |
+
"grad_norm": 44.00162124633789,
|
| 2809 |
+
"learning_rate": 1.2648549055734685e-05,
|
| 2810 |
+
"loss": 0.0755,
|
| 2811 |
+
"step": 3990
|
| 2812 |
+
},
|
| 2813 |
+
{
|
| 2814 |
+
"epoch": 1.84,
|
| 2815 |
+
"grad_norm": 1.8568840026855469,
|
| 2816 |
+
"learning_rate": 1.2630124366651313e-05,
|
| 2817 |
+
"loss": 0.0386,
|
| 2818 |
+
"step": 4000
|
| 2819 |
+
},
|
| 2820 |
+
{
|
| 2821 |
+
"epoch": 1.85,
|
| 2822 |
+
"grad_norm": 0.011160570196807384,
|
| 2823 |
+
"learning_rate": 1.2611699677567942e-05,
|
| 2824 |
+
"loss": 0.1926,
|
| 2825 |
+
"step": 4010
|
| 2826 |
+
},
|
| 2827 |
+
{
|
| 2828 |
+
"epoch": 1.85,
|
| 2829 |
+
"grad_norm": 0.5079384446144104,
|
| 2830 |
+
"learning_rate": 1.259327498848457e-05,
|
| 2831 |
+
"loss": 0.1599,
|
| 2832 |
+
"step": 4020
|
| 2833 |
+
},
|
| 2834 |
+
{
|
| 2835 |
+
"epoch": 1.86,
|
| 2836 |
+
"grad_norm": 0.009441236034035683,
|
| 2837 |
+
"learning_rate": 1.2574850299401197e-05,
|
| 2838 |
+
"loss": 0.0247,
|
| 2839 |
+
"step": 4030
|
| 2840 |
+
},
|
| 2841 |
+
{
|
| 2842 |
+
"epoch": 1.86,
|
| 2843 |
+
"grad_norm": 12.928335189819336,
|
| 2844 |
+
"learning_rate": 1.2556425610317826e-05,
|
| 2845 |
+
"loss": 0.0664,
|
| 2846 |
+
"step": 4040
|
| 2847 |
+
},
|
| 2848 |
+
{
|
| 2849 |
+
"epoch": 1.87,
|
| 2850 |
+
"grad_norm": 0.022305959835648537,
|
| 2851 |
+
"learning_rate": 1.2538000921234454e-05,
|
| 2852 |
+
"loss": 0.0335,
|
| 2853 |
+
"step": 4050
|
| 2854 |
+
},
|
| 2855 |
+
{
|
| 2856 |
+
"epoch": 1.87,
|
| 2857 |
+
"grad_norm": 9.200892448425293,
|
| 2858 |
+
"learning_rate": 1.2519576232151082e-05,
|
| 2859 |
+
"loss": 0.0132,
|
| 2860 |
+
"step": 4060
|
| 2861 |
+
},
|
| 2862 |
+
{
|
| 2863 |
+
"epoch": 1.87,
|
| 2864 |
+
"grad_norm": 6.198980808258057,
|
| 2865 |
+
"learning_rate": 1.2501151543067711e-05,
|
| 2866 |
+
"loss": 0.0186,
|
| 2867 |
+
"step": 4070
|
| 2868 |
+
},
|
| 2869 |
+
{
|
| 2870 |
+
"epoch": 1.88,
|
| 2871 |
+
"grad_norm": 24.16031837463379,
|
| 2872 |
+
"learning_rate": 1.248272685398434e-05,
|
| 2873 |
+
"loss": 0.0467,
|
| 2874 |
+
"step": 4080
|
| 2875 |
+
},
|
| 2876 |
+
{
|
| 2877 |
+
"epoch": 1.88,
|
| 2878 |
+
"grad_norm": 3.352095365524292,
|
| 2879 |
+
"learning_rate": 1.2464302164900968e-05,
|
| 2880 |
+
"loss": 0.0291,
|
| 2881 |
+
"step": 4090
|
| 2882 |
+
},
|
| 2883 |
+
{
|
| 2884 |
+
"epoch": 1.89,
|
| 2885 |
+
"grad_norm": 0.002663030056282878,
|
| 2886 |
+
"learning_rate": 1.2445877475817596e-05,
|
| 2887 |
+
"loss": 0.0195,
|
| 2888 |
+
"step": 4100
|
| 2889 |
+
},
|
| 2890 |
+
{
|
| 2891 |
+
"epoch": 1.89,
|
| 2892 |
+
"grad_norm": 22.6085262298584,
|
| 2893 |
+
"learning_rate": 1.2427452786734225e-05,
|
| 2894 |
+
"loss": 0.1135,
|
| 2895 |
+
"step": 4110
|
| 2896 |
+
},
|
| 2897 |
+
{
|
| 2898 |
+
"epoch": 1.9,
|
| 2899 |
+
"grad_norm": 6.5170392990112305,
|
| 2900 |
+
"learning_rate": 1.2409028097650853e-05,
|
| 2901 |
+
"loss": 0.0842,
|
| 2902 |
+
"step": 4120
|
| 2903 |
+
},
|
| 2904 |
+
{
|
| 2905 |
+
"epoch": 1.9,
|
| 2906 |
+
"grad_norm": 2.683942739167833e-07,
|
| 2907 |
+
"learning_rate": 1.2390603408567482e-05,
|
| 2908 |
+
"loss": 0.0047,
|
| 2909 |
+
"step": 4130
|
| 2910 |
+
},
|
| 2911 |
+
{
|
| 2912 |
+
"epoch": 1.91,
|
| 2913 |
+
"grad_norm": 0.02569938451051712,
|
| 2914 |
+
"learning_rate": 1.2372178719484108e-05,
|
| 2915 |
+
"loss": 0.0447,
|
| 2916 |
+
"step": 4140
|
| 2917 |
+
},
|
| 2918 |
+
{
|
| 2919 |
+
"epoch": 1.91,
|
| 2920 |
+
"grad_norm": 9.391770362854004,
|
| 2921 |
+
"learning_rate": 1.2353754030400737e-05,
|
| 2922 |
+
"loss": 0.0653,
|
| 2923 |
+
"step": 4150
|
| 2924 |
+
},
|
| 2925 |
+
{
|
| 2926 |
+
"epoch": 1.92,
|
| 2927 |
+
"grad_norm": 15.271596908569336,
|
| 2928 |
+
"learning_rate": 1.2335329341317365e-05,
|
| 2929 |
+
"loss": 0.2332,
|
| 2930 |
+
"step": 4160
|
| 2931 |
+
},
|
| 2932 |
+
{
|
| 2933 |
+
"epoch": 1.92,
|
| 2934 |
+
"grad_norm": 0.18177984654903412,
|
| 2935 |
+
"learning_rate": 1.2316904652233994e-05,
|
| 2936 |
+
"loss": 0.0545,
|
| 2937 |
+
"step": 4170
|
| 2938 |
+
},
|
| 2939 |
+
{
|
| 2940 |
+
"epoch": 1.92,
|
| 2941 |
+
"grad_norm": 0.024433467537164688,
|
| 2942 |
+
"learning_rate": 1.2298479963150622e-05,
|
| 2943 |
+
"loss": 0.0902,
|
| 2944 |
+
"step": 4180
|
| 2945 |
+
},
|
| 2946 |
+
{
|
| 2947 |
+
"epoch": 1.93,
|
| 2948 |
+
"grad_norm": 1.9357306957244873,
|
| 2949 |
+
"learning_rate": 1.228005527406725e-05,
|
| 2950 |
+
"loss": 0.0168,
|
| 2951 |
+
"step": 4190
|
| 2952 |
+
},
|
| 2953 |
+
{
|
| 2954 |
+
"epoch": 1.93,
|
| 2955 |
+
"grad_norm": 5.406333923339844,
|
| 2956 |
+
"learning_rate": 1.226163058498388e-05,
|
| 2957 |
+
"loss": 0.0299,
|
| 2958 |
+
"step": 4200
|
| 2959 |
+
},
|
| 2960 |
+
{
|
| 2961 |
+
"epoch": 1.94,
|
| 2962 |
+
"grad_norm": 0.0019183546537533402,
|
| 2963 |
+
"learning_rate": 1.2243205895900508e-05,
|
| 2964 |
+
"loss": 0.0018,
|
| 2965 |
+
"step": 4210
|
| 2966 |
+
},
|
| 2967 |
+
{
|
| 2968 |
+
"epoch": 1.94,
|
| 2969 |
+
"grad_norm": 7.152927398681641,
|
| 2970 |
+
"learning_rate": 1.2224781206817136e-05,
|
| 2971 |
+
"loss": 0.04,
|
| 2972 |
+
"step": 4220
|
| 2973 |
+
},
|
| 2974 |
+
{
|
| 2975 |
+
"epoch": 1.95,
|
| 2976 |
+
"grad_norm": 14.792912483215332,
|
| 2977 |
+
"learning_rate": 1.2206356517733765e-05,
|
| 2978 |
+
"loss": 0.0992,
|
| 2979 |
+
"step": 4230
|
| 2980 |
+
},
|
| 2981 |
+
{
|
| 2982 |
+
"epoch": 1.95,
|
| 2983 |
+
"grad_norm": 3.2190780639648438,
|
| 2984 |
+
"learning_rate": 1.2187931828650393e-05,
|
| 2985 |
+
"loss": 0.0629,
|
| 2986 |
+
"step": 4240
|
| 2987 |
+
},
|
| 2988 |
+
{
|
| 2989 |
+
"epoch": 1.96,
|
| 2990 |
+
"grad_norm": 0.8792375326156616,
|
| 2991 |
+
"learning_rate": 1.216950713956702e-05,
|
| 2992 |
+
"loss": 0.2232,
|
| 2993 |
+
"step": 4250
|
| 2994 |
+
},
|
| 2995 |
+
{
|
| 2996 |
+
"epoch": 1.96,
|
| 2997 |
+
"grad_norm": 16.167522430419922,
|
| 2998 |
+
"learning_rate": 1.2151082450483648e-05,
|
| 2999 |
+
"loss": 0.1202,
|
| 3000 |
+
"step": 4260
|
| 3001 |
+
},
|
| 3002 |
+
{
|
| 3003 |
+
"epoch": 1.97,
|
| 3004 |
+
"grad_norm": 0.4953334927558899,
|
| 3005 |
+
"learning_rate": 1.2132657761400277e-05,
|
| 3006 |
+
"loss": 0.0458,
|
| 3007 |
+
"step": 4270
|
| 3008 |
+
},
|
| 3009 |
+
{
|
| 3010 |
+
"epoch": 1.97,
|
| 3011 |
+
"grad_norm": 1.541945457458496,
|
| 3012 |
+
"learning_rate": 1.2114233072316905e-05,
|
| 3013 |
+
"loss": 0.0788,
|
| 3014 |
+
"step": 4280
|
| 3015 |
+
},
|
| 3016 |
+
{
|
| 3017 |
+
"epoch": 1.98,
|
| 3018 |
+
"grad_norm": 15.827666282653809,
|
| 3019 |
+
"learning_rate": 1.2095808383233534e-05,
|
| 3020 |
+
"loss": 0.0443,
|
| 3021 |
+
"step": 4290
|
| 3022 |
+
},
|
| 3023 |
+
{
|
| 3024 |
+
"epoch": 1.98,
|
| 3025 |
+
"grad_norm": 4.886255264282227,
|
| 3026 |
+
"learning_rate": 1.2077383694150162e-05,
|
| 3027 |
+
"loss": 0.0342,
|
| 3028 |
+
"step": 4300
|
| 3029 |
+
},
|
| 3030 |
+
{
|
| 3031 |
+
"epoch": 1.98,
|
| 3032 |
+
"grad_norm": 0.23300336301326752,
|
| 3033 |
+
"learning_rate": 1.205895900506679e-05,
|
| 3034 |
+
"loss": 0.015,
|
| 3035 |
+
"step": 4310
|
| 3036 |
+
},
|
| 3037 |
+
{
|
| 3038 |
+
"epoch": 1.99,
|
| 3039 |
+
"grad_norm": 0.032339416444301605,
|
| 3040 |
+
"learning_rate": 1.2040534315983419e-05,
|
| 3041 |
+
"loss": 0.0745,
|
| 3042 |
+
"step": 4320
|
| 3043 |
+
},
|
| 3044 |
+
{
|
| 3045 |
+
"epoch": 1.99,
|
| 3046 |
+
"grad_norm": 2.028648853302002,
|
| 3047 |
+
"learning_rate": 1.2022109626900048e-05,
|
| 3048 |
+
"loss": 0.0902,
|
| 3049 |
+
"step": 4330
|
| 3050 |
+
},
|
| 3051 |
+
{
|
| 3052 |
+
"epoch": 2.0,
|
| 3053 |
+
"grad_norm": 0.48279210925102234,
|
| 3054 |
+
"learning_rate": 1.2003684937816676e-05,
|
| 3055 |
+
"loss": 0.0185,
|
| 3056 |
+
"step": 4340
|
| 3057 |
+
},
|
| 3058 |
+
{
|
| 3059 |
+
"epoch": 2.0,
|
| 3060 |
+
"eval_accuracy": 0.965769358590894,
|
| 3061 |
+
"eval_loss": 0.13561101257801056,
|
| 3062 |
+
"eval_runtime": 88.9935,
|
| 3063 |
+
"eval_samples_per_second": 33.811,
|
| 3064 |
+
"eval_steps_per_second": 6.765,
|
| 3065 |
+
"step": 4343
|
| 3066 |
+
}
|
| 3067 |
+
],
|
| 3068 |
+
"logging_steps": 10,
|
| 3069 |
+
"max_steps": 10855,
|
| 3070 |
+
"num_input_tokens_seen": 0,
|
| 3071 |
+
"num_train_epochs": 5,
|
| 3072 |
+
"save_steps": 500,
|
| 3073 |
+
"total_flos": 4.57051372770263e+16,
|
| 3074 |
+
"train_batch_size": 5,
|
| 3075 |
+
"trial_name": null,
|
| 3076 |
+
"trial_params": null
|
| 3077 |
+
}
|
hw1/results_paragraph/checkpoint-4343/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:790ffd4d5b93e8e921d4c443c73e0e94cbd60d7ccce45b20666b7277f4d695ec
|
| 3 |
+
size 4984
|
hw1/results_paragraph/checkpoint-4343/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
hw1/results_paragraph/runs/Sep26_12-47-40_d88996431f93/events.out.tfevents.1727354862.d88996431f93.1905835.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2e974d62874f6e1099e47ac378f932fd4961fd0e96fe5e0e628edadcd34bbc57
|
| 3 |
+
size 4845
|
hw1/results_paragraph/runs/Sep26_13-37-30_d88996431f93/events.out.tfevents.1727357852.d88996431f93.2093465.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:49f57e5a33cdb5934224138264809ad420816ba83a053e66f45910f0eaef0ebf
|
| 3 |
+
size 189591
|
hw1/results_span/checkpoint-10855/config.json
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "hfl/chinese-roberta-wwm-ext",
|
| 3 |
+
"architectures": [
|
| 4 |
+
"BertForQuestionAnswering"
|
| 5 |
+
],
|
| 6 |
+
"attention_probs_dropout_prob": 0.1,
|
| 7 |
+
"bos_token_id": 0,
|
| 8 |
+
"classifier_dropout": null,
|
| 9 |
+
"directionality": "bidi",
|
| 10 |
+
"eos_token_id": 2,
|
| 11 |
+
"hidden_act": "gelu",
|
| 12 |
+
"hidden_dropout_prob": 0.1,
|
| 13 |
+
"hidden_size": 768,
|
| 14 |
+
"initializer_range": 0.02,
|
| 15 |
+
"intermediate_size": 3072,
|
| 16 |
+
"layer_norm_eps": 1e-12,
|
| 17 |
+
"max_position_embeddings": 512,
|
| 18 |
+
"model_type": "bert",
|
| 19 |
+
"num_attention_heads": 12,
|
| 20 |
+
"num_hidden_layers": 12,
|
| 21 |
+
"output_past": true,
|
| 22 |
+
"pad_token_id": 0,
|
| 23 |
+
"pooler_fc_size": 768,
|
| 24 |
+
"pooler_num_attention_heads": 12,
|
| 25 |
+
"pooler_num_fc_layers": 3,
|
| 26 |
+
"pooler_size_per_head": 128,
|
| 27 |
+
"pooler_type": "first_token_transform",
|
| 28 |
+
"position_embedding_type": "absolute",
|
| 29 |
+
"torch_dtype": "float32",
|
| 30 |
+
"transformers_version": "4.39.0.dev0",
|
| 31 |
+
"type_vocab_size": 2,
|
| 32 |
+
"use_cache": true,
|
| 33 |
+
"vocab_size": 21128
|
| 34 |
+
}
|
hw1/results_span/checkpoint-10855/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ed30dc25599070d6e99273766e9e3e214e0f65bc22228d72db6e793123a69e20
|
| 3 |
+
size 406737680
|
hw1/results_span/checkpoint-10855/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:33bbea16a054c26ff8ca50503cd9b6e8a6fb3c14c1e9f9371f31c29280305d83
|
| 3 |
+
size 813595194
|
hw1/results_span/checkpoint-10855/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a95b97d93943377b1a89efc8eac5f5873f333aab57a5e5338ce23cbe0dc86a7c
|
| 3 |
+
size 14244
|
hw1/results_span/checkpoint-10855/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6668b378e95163e7feae2889383adc2e33fe7df813bcb8cb500352e1ad546cc9
|
| 3 |
+
size 1064
|
hw1/results_span/checkpoint-10855/special_tokens_map.json
ADDED
|
@@ -0,0 +1,37 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cls_token": {
|
| 3 |
+
"content": "[CLS]",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"mask_token": {
|
| 10 |
+
"content": "[MASK]",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"pad_token": {
|
| 17 |
+
"content": "[PAD]",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"sep_token": {
|
| 24 |
+
"content": "[SEP]",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"unk_token": {
|
| 31 |
+
"content": "[UNK]",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
}
|
| 37 |
+
}
|
hw1/results_span/checkpoint-10855/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
hw1/results_span/checkpoint-10855/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "[PAD]",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"100": {
|
| 12 |
+
"content": "[UNK]",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"101": {
|
| 20 |
+
"content": "[CLS]",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"102": {
|
| 28 |
+
"content": "[SEP]",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"103": {
|
| 36 |
+
"content": "[MASK]",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"clean_up_tokenization_spaces": true,
|
| 45 |
+
"cls_token": "[CLS]",
|
| 46 |
+
"do_lower_case": true,
|
| 47 |
+
"mask_token": "[MASK]",
|
| 48 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 49 |
+
"pad_token": "[PAD]",
|
| 50 |
+
"sep_token": "[SEP]",
|
| 51 |
+
"strip_accents": null,
|
| 52 |
+
"tokenize_chinese_chars": true,
|
| 53 |
+
"tokenizer_class": "BertTokenizer",
|
| 54 |
+
"unk_token": "[UNK]"
|
| 55 |
+
}
|
hw1/results_span/checkpoint-10855/trainer_state.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
hw1/results_span/checkpoint-10855/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6a6fc70f2067f04bba006b499ca1f7a90f0a85d05f736be8e541e7c7a8d5a9e4
|
| 3 |
+
size 4984
|
hw1/results_span/checkpoint-10855/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
hw1/results_span/runs/Sep30_07-11-20_d88996431f93/events.out.tfevents.1727680284.d88996431f93.2093465.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a9150a4ff3befc64053b061481928f7e50de332c93875900aeee154d9c2a6ada
|
| 3 |
+
size 50834
|
hw1/results_span/runs/Sep30_07-49-45_d88996431f93/events.out.tfevents.1727682589.d88996431f93.906292.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d07e7bebc1f18d71ba2f906529bcc2ebea19c858ecf1aa2e7a55a722e83e49cb
|
| 3 |
+
size 235422
|