Upload 12 files
Browse files- config.json +100 -0
- generation_config.json +7 -0
- model.safetensors +3 -0
- rng_state.pth +3 -0
- scaler.pt +3 -0
- scheduler.pt +3 -0
- special_tokens_map.json +37 -0
- tokenizer.json +0 -0
- tokenizer_config.json +65 -0
- trainer_state.json +2554 -0
- training_args.bin +3 -0
- vocab.txt +0 -0
config.json
ADDED
|
@@ -0,0 +1,100 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"EncoderDecoderModel"
|
| 4 |
+
],
|
| 5 |
+
"decoder": {
|
| 6 |
+
"_name_or_path": "skt/kogpt2-base-v2",
|
| 7 |
+
"_num_labels": 1,
|
| 8 |
+
"activation_function": "gelu_new",
|
| 9 |
+
"add_cross_attention": true,
|
| 10 |
+
"architectures": [
|
| 11 |
+
"GPT2LMHeadModel"
|
| 12 |
+
],
|
| 13 |
+
"attn_pdrop": 0.1,
|
| 14 |
+
"author": "Heewon Jeon(madjakarta@gmail.com)",
|
| 15 |
+
"bos_token_id": 0,
|
| 16 |
+
"created_date": "2021-04-28",
|
| 17 |
+
"dtype": "float32",
|
| 18 |
+
"embd_pdrop": 0.1,
|
| 19 |
+
"eos_token_id": 1,
|
| 20 |
+
"gradient_checkpointing": false,
|
| 21 |
+
"id2label": {
|
| 22 |
+
"0": "LABEL_0"
|
| 23 |
+
},
|
| 24 |
+
"initializer_range": 0.02,
|
| 25 |
+
"is_decoder": true,
|
| 26 |
+
"label2id": {
|
| 27 |
+
"LABEL_0": 0
|
| 28 |
+
},
|
| 29 |
+
"layer_norm_epsilon": 1e-05,
|
| 30 |
+
"license": "CC-BY-NC-SA 4.0",
|
| 31 |
+
"model_type": "gpt2",
|
| 32 |
+
"n_ctx": 1024,
|
| 33 |
+
"n_embd": 768,
|
| 34 |
+
"n_head": 12,
|
| 35 |
+
"n_inner": null,
|
| 36 |
+
"n_layer": 12,
|
| 37 |
+
"n_positions": 1024,
|
| 38 |
+
"pad_token_id": 3,
|
| 39 |
+
"reorder_and_upcast_attn": false,
|
| 40 |
+
"resid_pdrop": 0.1,
|
| 41 |
+
"scale_attn_by_inverse_layer_idx": false,
|
| 42 |
+
"scale_attn_weights": true,
|
| 43 |
+
"summary_activation": null,
|
| 44 |
+
"summary_first_dropout": 0.1,
|
| 45 |
+
"summary_proj_to_labels": true,
|
| 46 |
+
"summary_type": "cls_index",
|
| 47 |
+
"summary_use_proj": true,
|
| 48 |
+
"task_specific_params": {
|
| 49 |
+
"text-generation": {
|
| 50 |
+
"do_sample": true,
|
| 51 |
+
"max_length": 50
|
| 52 |
+
}
|
| 53 |
+
},
|
| 54 |
+
"use_cache": true,
|
| 55 |
+
"vocab_size": 51201
|
| 56 |
+
},
|
| 57 |
+
"decoder_start_token_id": 51200,
|
| 58 |
+
"dtype": "float32",
|
| 59 |
+
"encoder": {
|
| 60 |
+
"_name_or_path": "snunlp/KR-FinBert-SC",
|
| 61 |
+
"architectures": [
|
| 62 |
+
"BertForSequenceClassification"
|
| 63 |
+
],
|
| 64 |
+
"attention_probs_dropout_prob": 0.1,
|
| 65 |
+
"classifier_dropout": null,
|
| 66 |
+
"dtype": "float32",
|
| 67 |
+
"gradient_checkpointing": false,
|
| 68 |
+
"hidden_act": "gelu",
|
| 69 |
+
"hidden_dropout_prob": 0.1,
|
| 70 |
+
"hidden_size": 768,
|
| 71 |
+
"id2label": {
|
| 72 |
+
"0": "negative",
|
| 73 |
+
"1": "neutral",
|
| 74 |
+
"2": "positive"
|
| 75 |
+
},
|
| 76 |
+
"initializer_range": 0.02,
|
| 77 |
+
"intermediate_size": 3072,
|
| 78 |
+
"label2id": {
|
| 79 |
+
"negative": 0,
|
| 80 |
+
"neutral": 1,
|
| 81 |
+
"positive": 2
|
| 82 |
+
},
|
| 83 |
+
"layer_norm_eps": 1e-12,
|
| 84 |
+
"max_position_embeddings": 512,
|
| 85 |
+
"model_type": "bert",
|
| 86 |
+
"num_attention_heads": 12,
|
| 87 |
+
"num_hidden_layers": 12,
|
| 88 |
+
"position_embedding_type": "absolute",
|
| 89 |
+
"problem_type": "single_label_classification",
|
| 90 |
+
"type_vocab_size": 2,
|
| 91 |
+
"use_cache": true,
|
| 92 |
+
"vocab_size": 20000
|
| 93 |
+
},
|
| 94 |
+
"eos_token_id": 51200,
|
| 95 |
+
"is_encoder_decoder": true,
|
| 96 |
+
"model_type": "encoder-decoder",
|
| 97 |
+
"pad_token_id": 51200,
|
| 98 |
+
"transformers_version": "4.56.1",
|
| 99 |
+
"vocab_size": 51201
|
| 100 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 0,
|
| 4 |
+
"eos_token_id": 1,
|
| 5 |
+
"pad_token_id": 3,
|
| 6 |
+
"transformers_version": "4.56.1"
|
| 7 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e8fe4fb752586ac1f80075fe68d5ba1504fc42d0ea565eb2659e5689059e217c
|
| 3 |
+
size 1019783696
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fc796ead0c9402aaf7286e58f527db81e4585e1d90bec5e0601c027bcf74fe94
|
| 3 |
+
size 14645
|
scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dde170cdb556cfcd15ba973ae8c92b53d83c8471ef183cbf7f9bba7b5a8a5101
|
| 3 |
+
size 1383
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:30620eac7fb7636e90cbb857d06e2a62a640eb9dfc2f4d71b6eaa9e81f93a040
|
| 3 |
+
size 1465
|
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 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,65 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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 |
+
"1": {
|
| 12 |
+
"content": "[UNK]",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "[CLS]",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "[SEP]",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"4": {
|
| 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": false,
|
| 45 |
+
"cls_token": "[CLS]",
|
| 46 |
+
"do_basic_tokenize": true,
|
| 47 |
+
"do_lower_case": false,
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "[MASK]",
|
| 50 |
+
"max_length": 300,
|
| 51 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 52 |
+
"never_split": null,
|
| 53 |
+
"pad_to_multiple_of": null,
|
| 54 |
+
"pad_token": "[PAD]",
|
| 55 |
+
"pad_token_type_id": 0,
|
| 56 |
+
"padding_side": "right",
|
| 57 |
+
"sep_token": "[SEP]",
|
| 58 |
+
"stride": 0,
|
| 59 |
+
"strip_accents": null,
|
| 60 |
+
"tokenize_chinese_chars": true,
|
| 61 |
+
"tokenizer_class": "BertTokenizer",
|
| 62 |
+
"truncation_side": "right",
|
| 63 |
+
"truncation_strategy": "longest_first",
|
| 64 |
+
"unk_token": "[UNK]"
|
| 65 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2554 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": null,
|
| 3 |
+
"best_metric": null,
|
| 4 |
+
"best_model_checkpoint": null,
|
| 5 |
+
"epoch": 10.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 3600,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.027777777777777776,
|
| 14 |
+
"grad_norm": 104.33584594726562,
|
| 15 |
+
"learning_rate": 1.35e-06,
|
| 16 |
+
"loss": 13.5343,
|
| 17 |
+
"step": 10
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.05555555555555555,
|
| 21 |
+
"grad_norm": 57.29974365234375,
|
| 22 |
+
"learning_rate": 2.8500000000000002e-06,
|
| 23 |
+
"loss": 12.1877,
|
| 24 |
+
"step": 20
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.08333333333333333,
|
| 28 |
+
"grad_norm": 35.94429016113281,
|
| 29 |
+
"learning_rate": 4.35e-06,
|
| 30 |
+
"loss": 10.9908,
|
| 31 |
+
"step": 30
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.1111111111111111,
|
| 35 |
+
"grad_norm": 36.21198272705078,
|
| 36 |
+
"learning_rate": 5.850000000000001e-06,
|
| 37 |
+
"loss": 10.8169,
|
| 38 |
+
"step": 40
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.1388888888888889,
|
| 42 |
+
"grad_norm": 35.18545150756836,
|
| 43 |
+
"learning_rate": 7.35e-06,
|
| 44 |
+
"loss": 10.226,
|
| 45 |
+
"step": 50
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.16666666666666666,
|
| 49 |
+
"grad_norm": 30.034957885742188,
|
| 50 |
+
"learning_rate": 8.85e-06,
|
| 51 |
+
"loss": 9.8243,
|
| 52 |
+
"step": 60
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.19444444444444445,
|
| 56 |
+
"grad_norm": 26.107460021972656,
|
| 57 |
+
"learning_rate": 1.035e-05,
|
| 58 |
+
"loss": 9.7202,
|
| 59 |
+
"step": 70
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.2222222222222222,
|
| 63 |
+
"grad_norm": 28.405378341674805,
|
| 64 |
+
"learning_rate": 1.185e-05,
|
| 65 |
+
"loss": 9.655,
|
| 66 |
+
"step": 80
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 0.25,
|
| 70 |
+
"grad_norm": 31.871849060058594,
|
| 71 |
+
"learning_rate": 1.3350000000000001e-05,
|
| 72 |
+
"loss": 9.1121,
|
| 73 |
+
"step": 90
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.2777777777777778,
|
| 77 |
+
"grad_norm": 31.50947380065918,
|
| 78 |
+
"learning_rate": 1.485e-05,
|
| 79 |
+
"loss": 9.1896,
|
| 80 |
+
"step": 100
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.3055555555555556,
|
| 84 |
+
"grad_norm": 28.950199127197266,
|
| 85 |
+
"learning_rate": 1.635e-05,
|
| 86 |
+
"loss": 9.2017,
|
| 87 |
+
"step": 110
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.3333333333333333,
|
| 91 |
+
"grad_norm": 25.37229347229004,
|
| 92 |
+
"learning_rate": 1.785e-05,
|
| 93 |
+
"loss": 9.2022,
|
| 94 |
+
"step": 120
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.3611111111111111,
|
| 98 |
+
"grad_norm": 24.216697692871094,
|
| 99 |
+
"learning_rate": 1.935e-05,
|
| 100 |
+
"loss": 9.1344,
|
| 101 |
+
"step": 130
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 0.3888888888888889,
|
| 105 |
+
"grad_norm": 24.010272979736328,
|
| 106 |
+
"learning_rate": 2.085e-05,
|
| 107 |
+
"loss": 8.5873,
|
| 108 |
+
"step": 140
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 0.4166666666666667,
|
| 112 |
+
"grad_norm": 24.390907287597656,
|
| 113 |
+
"learning_rate": 2.235e-05,
|
| 114 |
+
"loss": 8.7195,
|
| 115 |
+
"step": 150
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.4444444444444444,
|
| 119 |
+
"grad_norm": 21.80168914794922,
|
| 120 |
+
"learning_rate": 2.385e-05,
|
| 121 |
+
"loss": 8.5485,
|
| 122 |
+
"step": 160
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.4722222222222222,
|
| 126 |
+
"grad_norm": 20.435110092163086,
|
| 127 |
+
"learning_rate": 2.535e-05,
|
| 128 |
+
"loss": 8.6359,
|
| 129 |
+
"step": 170
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.5,
|
| 133 |
+
"grad_norm": 24.11395835876465,
|
| 134 |
+
"learning_rate": 2.6850000000000002e-05,
|
| 135 |
+
"loss": 8.4246,
|
| 136 |
+
"step": 180
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 0.5277777777777778,
|
| 140 |
+
"grad_norm": 25.995811462402344,
|
| 141 |
+
"learning_rate": 2.8349999999999998e-05,
|
| 142 |
+
"loss": 8.5049,
|
| 143 |
+
"step": 190
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 0.5555555555555556,
|
| 147 |
+
"grad_norm": 23.826154708862305,
|
| 148 |
+
"learning_rate": 2.985e-05,
|
| 149 |
+
"loss": 8.4623,
|
| 150 |
+
"step": 200
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 0.5833333333333334,
|
| 154 |
+
"grad_norm": 20.094274520874023,
|
| 155 |
+
"learning_rate": 2.9920588235294118e-05,
|
| 156 |
+
"loss": 8.4495,
|
| 157 |
+
"step": 210
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 0.6111111111111112,
|
| 161 |
+
"grad_norm": 22.793804168701172,
|
| 162 |
+
"learning_rate": 2.983235294117647e-05,
|
| 163 |
+
"loss": 8.4131,
|
| 164 |
+
"step": 220
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.6388888888888888,
|
| 168 |
+
"grad_norm": 26.08730125427246,
|
| 169 |
+
"learning_rate": 2.9744117647058824e-05,
|
| 170 |
+
"loss": 8.3116,
|
| 171 |
+
"step": 230
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.6666666666666666,
|
| 175 |
+
"grad_norm": 24.06973648071289,
|
| 176 |
+
"learning_rate": 2.9655882352941176e-05,
|
| 177 |
+
"loss": 8.4003,
|
| 178 |
+
"step": 240
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 0.6944444444444444,
|
| 182 |
+
"grad_norm": 22.93770408630371,
|
| 183 |
+
"learning_rate": 2.956764705882353e-05,
|
| 184 |
+
"loss": 8.2579,
|
| 185 |
+
"step": 250
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.7222222222222222,
|
| 189 |
+
"grad_norm": 20.36606216430664,
|
| 190 |
+
"learning_rate": 2.9479411764705883e-05,
|
| 191 |
+
"loss": 8.2888,
|
| 192 |
+
"step": 260
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 0.75,
|
| 196 |
+
"grad_norm": 21.670818328857422,
|
| 197 |
+
"learning_rate": 2.9391176470588238e-05,
|
| 198 |
+
"loss": 8.1623,
|
| 199 |
+
"step": 270
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 0.7777777777777778,
|
| 203 |
+
"grad_norm": 17.550491333007812,
|
| 204 |
+
"learning_rate": 2.930294117647059e-05,
|
| 205 |
+
"loss": 8.0683,
|
| 206 |
+
"step": 280
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.8055555555555556,
|
| 210 |
+
"grad_norm": 25.335660934448242,
|
| 211 |
+
"learning_rate": 2.9214705882352944e-05,
|
| 212 |
+
"loss": 8.2784,
|
| 213 |
+
"step": 290
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.8333333333333334,
|
| 217 |
+
"grad_norm": 20.76925277709961,
|
| 218 |
+
"learning_rate": 2.9126470588235292e-05,
|
| 219 |
+
"loss": 8.2923,
|
| 220 |
+
"step": 300
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 0.8611111111111112,
|
| 224 |
+
"grad_norm": 23.108144760131836,
|
| 225 |
+
"learning_rate": 2.9038235294117647e-05,
|
| 226 |
+
"loss": 8.1554,
|
| 227 |
+
"step": 310
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 0.8888888888888888,
|
| 231 |
+
"grad_norm": 16.563575744628906,
|
| 232 |
+
"learning_rate": 2.895e-05,
|
| 233 |
+
"loss": 7.8543,
|
| 234 |
+
"step": 320
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.9166666666666666,
|
| 238 |
+
"grad_norm": 19.03083038330078,
|
| 239 |
+
"learning_rate": 2.8861764705882354e-05,
|
| 240 |
+
"loss": 8.0434,
|
| 241 |
+
"step": 330
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 0.9444444444444444,
|
| 245 |
+
"grad_norm": 18.121356964111328,
|
| 246 |
+
"learning_rate": 2.8773529411764706e-05,
|
| 247 |
+
"loss": 7.93,
|
| 248 |
+
"step": 340
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.9722222222222222,
|
| 252 |
+
"grad_norm": 21.65605926513672,
|
| 253 |
+
"learning_rate": 2.868529411764706e-05,
|
| 254 |
+
"loss": 8.0124,
|
| 255 |
+
"step": 350
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 1.0,
|
| 259 |
+
"grad_norm": 24.55741310119629,
|
| 260 |
+
"learning_rate": 2.8597058823529412e-05,
|
| 261 |
+
"loss": 7.547,
|
| 262 |
+
"step": 360
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 1.0277777777777777,
|
| 266 |
+
"grad_norm": 18.358055114746094,
|
| 267 |
+
"learning_rate": 2.8508823529411767e-05,
|
| 268 |
+
"loss": 6.9283,
|
| 269 |
+
"step": 370
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 1.0555555555555556,
|
| 273 |
+
"grad_norm": 18.642698287963867,
|
| 274 |
+
"learning_rate": 2.842058823529412e-05,
|
| 275 |
+
"loss": 7.1341,
|
| 276 |
+
"step": 380
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 1.0833333333333333,
|
| 280 |
+
"grad_norm": 30.198606491088867,
|
| 281 |
+
"learning_rate": 2.8332352941176474e-05,
|
| 282 |
+
"loss": 7.113,
|
| 283 |
+
"step": 390
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 1.1111111111111112,
|
| 287 |
+
"grad_norm": 19.851511001586914,
|
| 288 |
+
"learning_rate": 2.8244117647058825e-05,
|
| 289 |
+
"loss": 7.1696,
|
| 290 |
+
"step": 400
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 1.1388888888888888,
|
| 294 |
+
"grad_norm": 20.48548698425293,
|
| 295 |
+
"learning_rate": 2.8155882352941177e-05,
|
| 296 |
+
"loss": 7.1552,
|
| 297 |
+
"step": 410
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 1.1666666666666667,
|
| 301 |
+
"grad_norm": 18.470821380615234,
|
| 302 |
+
"learning_rate": 2.806764705882353e-05,
|
| 303 |
+
"loss": 6.8519,
|
| 304 |
+
"step": 420
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 1.1944444444444444,
|
| 308 |
+
"grad_norm": 20.821441650390625,
|
| 309 |
+
"learning_rate": 2.7979411764705883e-05,
|
| 310 |
+
"loss": 7.0113,
|
| 311 |
+
"step": 430
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 1.2222222222222223,
|
| 315 |
+
"grad_norm": 19.294477462768555,
|
| 316 |
+
"learning_rate": 2.7891176470588235e-05,
|
| 317 |
+
"loss": 7.3983,
|
| 318 |
+
"step": 440
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 1.25,
|
| 322 |
+
"grad_norm": 19.5776309967041,
|
| 323 |
+
"learning_rate": 2.780294117647059e-05,
|
| 324 |
+
"loss": 7.0649,
|
| 325 |
+
"step": 450
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 1.2777777777777777,
|
| 329 |
+
"grad_norm": 18.286705017089844,
|
| 330 |
+
"learning_rate": 2.771470588235294e-05,
|
| 331 |
+
"loss": 7.1587,
|
| 332 |
+
"step": 460
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 1.3055555555555556,
|
| 336 |
+
"grad_norm": 19.803571701049805,
|
| 337 |
+
"learning_rate": 2.7626470588235297e-05,
|
| 338 |
+
"loss": 6.9439,
|
| 339 |
+
"step": 470
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 1.3333333333333333,
|
| 343 |
+
"grad_norm": 20.182926177978516,
|
| 344 |
+
"learning_rate": 2.7538235294117648e-05,
|
| 345 |
+
"loss": 6.9072,
|
| 346 |
+
"step": 480
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 1.3611111111111112,
|
| 350 |
+
"grad_norm": 21.21506690979004,
|
| 351 |
+
"learning_rate": 2.7450000000000003e-05,
|
| 352 |
+
"loss": 6.7805,
|
| 353 |
+
"step": 490
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 1.3888888888888888,
|
| 357 |
+
"grad_norm": 18.884464263916016,
|
| 358 |
+
"learning_rate": 2.7361764705882355e-05,
|
| 359 |
+
"loss": 6.7809,
|
| 360 |
+
"step": 500
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 1.4166666666666667,
|
| 364 |
+
"grad_norm": 18.7882080078125,
|
| 365 |
+
"learning_rate": 2.7273529411764706e-05,
|
| 366 |
+
"loss": 7.1101,
|
| 367 |
+
"step": 510
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 1.4444444444444444,
|
| 371 |
+
"grad_norm": 17.172212600708008,
|
| 372 |
+
"learning_rate": 2.7185294117647058e-05,
|
| 373 |
+
"loss": 6.8579,
|
| 374 |
+
"step": 520
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 1.4722222222222223,
|
| 378 |
+
"grad_norm": 18.599943161010742,
|
| 379 |
+
"learning_rate": 2.7097058823529413e-05,
|
| 380 |
+
"loss": 6.9192,
|
| 381 |
+
"step": 530
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 1.5,
|
| 385 |
+
"grad_norm": 22.43140411376953,
|
| 386 |
+
"learning_rate": 2.7008823529411765e-05,
|
| 387 |
+
"loss": 6.9068,
|
| 388 |
+
"step": 540
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 1.5277777777777777,
|
| 392 |
+
"grad_norm": 16.94782257080078,
|
| 393 |
+
"learning_rate": 2.692058823529412e-05,
|
| 394 |
+
"loss": 7.0826,
|
| 395 |
+
"step": 550
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 1.5555555555555556,
|
| 399 |
+
"grad_norm": 18.50872802734375,
|
| 400 |
+
"learning_rate": 2.683235294117647e-05,
|
| 401 |
+
"loss": 6.8281,
|
| 402 |
+
"step": 560
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 1.5833333333333335,
|
| 406 |
+
"grad_norm": 22.08888053894043,
|
| 407 |
+
"learning_rate": 2.6744117647058826e-05,
|
| 408 |
+
"loss": 6.9378,
|
| 409 |
+
"step": 570
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 1.6111111111111112,
|
| 413 |
+
"grad_norm": 17.366363525390625,
|
| 414 |
+
"learning_rate": 2.6655882352941178e-05,
|
| 415 |
+
"loss": 6.6598,
|
| 416 |
+
"step": 580
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 1.6388888888888888,
|
| 420 |
+
"grad_norm": 17.86391830444336,
|
| 421 |
+
"learning_rate": 2.6567647058823533e-05,
|
| 422 |
+
"loss": 6.8419,
|
| 423 |
+
"step": 590
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 1.6666666666666665,
|
| 427 |
+
"grad_norm": 16.568017959594727,
|
| 428 |
+
"learning_rate": 2.6479411764705884e-05,
|
| 429 |
+
"loss": 6.9137,
|
| 430 |
+
"step": 600
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 1.6944444444444444,
|
| 434 |
+
"grad_norm": 21.780588150024414,
|
| 435 |
+
"learning_rate": 2.6391176470588236e-05,
|
| 436 |
+
"loss": 7.0719,
|
| 437 |
+
"step": 610
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 1.7222222222222223,
|
| 441 |
+
"grad_norm": 20.24903678894043,
|
| 442 |
+
"learning_rate": 2.6302941176470587e-05,
|
| 443 |
+
"loss": 6.8612,
|
| 444 |
+
"step": 620
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 1.75,
|
| 448 |
+
"grad_norm": 23.022014617919922,
|
| 449 |
+
"learning_rate": 2.621470588235294e-05,
|
| 450 |
+
"loss": 6.815,
|
| 451 |
+
"step": 630
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 1.7777777777777777,
|
| 455 |
+
"grad_norm": 19.738136291503906,
|
| 456 |
+
"learning_rate": 2.6126470588235294e-05,
|
| 457 |
+
"loss": 7.0848,
|
| 458 |
+
"step": 640
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 1.8055555555555556,
|
| 462 |
+
"grad_norm": 19.904956817626953,
|
| 463 |
+
"learning_rate": 2.6038235294117646e-05,
|
| 464 |
+
"loss": 7.1127,
|
| 465 |
+
"step": 650
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 1.8333333333333335,
|
| 469 |
+
"grad_norm": 20.379528045654297,
|
| 470 |
+
"learning_rate": 2.595e-05,
|
| 471 |
+
"loss": 7.1542,
|
| 472 |
+
"step": 660
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 1.8611111111111112,
|
| 476 |
+
"grad_norm": 17.21288299560547,
|
| 477 |
+
"learning_rate": 2.5861764705882352e-05,
|
| 478 |
+
"loss": 6.4236,
|
| 479 |
+
"step": 670
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 1.8888888888888888,
|
| 483 |
+
"grad_norm": 19.684776306152344,
|
| 484 |
+
"learning_rate": 2.5773529411764707e-05,
|
| 485 |
+
"loss": 6.97,
|
| 486 |
+
"step": 680
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 1.9166666666666665,
|
| 490 |
+
"grad_norm": 22.62903594970703,
|
| 491 |
+
"learning_rate": 2.568529411764706e-05,
|
| 492 |
+
"loss": 6.7521,
|
| 493 |
+
"step": 690
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 1.9444444444444444,
|
| 497 |
+
"grad_norm": 18.688766479492188,
|
| 498 |
+
"learning_rate": 2.5597058823529414e-05,
|
| 499 |
+
"loss": 6.7485,
|
| 500 |
+
"step": 700
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 1.9722222222222223,
|
| 504 |
+
"grad_norm": 19.558578491210938,
|
| 505 |
+
"learning_rate": 2.5508823529411765e-05,
|
| 506 |
+
"loss": 6.6582,
|
| 507 |
+
"step": 710
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 2.0,
|
| 511 |
+
"grad_norm": 20.0054874420166,
|
| 512 |
+
"learning_rate": 2.542058823529412e-05,
|
| 513 |
+
"loss": 6.4649,
|
| 514 |
+
"step": 720
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 2.0277777777777777,
|
| 518 |
+
"grad_norm": 17.48706817626953,
|
| 519 |
+
"learning_rate": 2.533235294117647e-05,
|
| 520 |
+
"loss": 6.0975,
|
| 521 |
+
"step": 730
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 2.0555555555555554,
|
| 525 |
+
"grad_norm": 21.419391632080078,
|
| 526 |
+
"learning_rate": 2.5244117647058823e-05,
|
| 527 |
+
"loss": 5.8201,
|
| 528 |
+
"step": 740
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 2.0833333333333335,
|
| 532 |
+
"grad_norm": 19.62574005126953,
|
| 533 |
+
"learning_rate": 2.5155882352941175e-05,
|
| 534 |
+
"loss": 5.792,
|
| 535 |
+
"step": 750
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 2.111111111111111,
|
| 539 |
+
"grad_norm": 24.182546615600586,
|
| 540 |
+
"learning_rate": 2.506764705882353e-05,
|
| 541 |
+
"loss": 6.002,
|
| 542 |
+
"step": 760
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 2.138888888888889,
|
| 546 |
+
"grad_norm": 17.403255462646484,
|
| 547 |
+
"learning_rate": 2.497941176470588e-05,
|
| 548 |
+
"loss": 5.8743,
|
| 549 |
+
"step": 770
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 2.1666666666666665,
|
| 553 |
+
"grad_norm": 19.421232223510742,
|
| 554 |
+
"learning_rate": 2.4891176470588237e-05,
|
| 555 |
+
"loss": 6.1047,
|
| 556 |
+
"step": 780
|
| 557 |
+
},
|
| 558 |
+
{
|
| 559 |
+
"epoch": 2.1944444444444446,
|
| 560 |
+
"grad_norm": 22.9908504486084,
|
| 561 |
+
"learning_rate": 2.4802941176470588e-05,
|
| 562 |
+
"loss": 5.9951,
|
| 563 |
+
"step": 790
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 2.2222222222222223,
|
| 567 |
+
"grad_norm": 18.04509925842285,
|
| 568 |
+
"learning_rate": 2.4714705882352943e-05,
|
| 569 |
+
"loss": 6.1774,
|
| 570 |
+
"step": 800
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 2.25,
|
| 574 |
+
"grad_norm": 16.365985870361328,
|
| 575 |
+
"learning_rate": 2.4626470588235295e-05,
|
| 576 |
+
"loss": 5.9879,
|
| 577 |
+
"step": 810
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 2.2777777777777777,
|
| 581 |
+
"grad_norm": 19.309423446655273,
|
| 582 |
+
"learning_rate": 2.453823529411765e-05,
|
| 583 |
+
"loss": 5.9292,
|
| 584 |
+
"step": 820
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 2.3055555555555554,
|
| 588 |
+
"grad_norm": 24.4521541595459,
|
| 589 |
+
"learning_rate": 2.4449999999999998e-05,
|
| 590 |
+
"loss": 6.0532,
|
| 591 |
+
"step": 830
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 2.3333333333333335,
|
| 595 |
+
"grad_norm": 22.1500244140625,
|
| 596 |
+
"learning_rate": 2.4361764705882353e-05,
|
| 597 |
+
"loss": 5.7139,
|
| 598 |
+
"step": 840
|
| 599 |
+
},
|
| 600 |
+
{
|
| 601 |
+
"epoch": 2.361111111111111,
|
| 602 |
+
"grad_norm": 18.496788024902344,
|
| 603 |
+
"learning_rate": 2.4273529411764705e-05,
|
| 604 |
+
"loss": 5.9176,
|
| 605 |
+
"step": 850
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 2.388888888888889,
|
| 609 |
+
"grad_norm": 19.089303970336914,
|
| 610 |
+
"learning_rate": 2.418529411764706e-05,
|
| 611 |
+
"loss": 6.0821,
|
| 612 |
+
"step": 860
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 2.4166666666666665,
|
| 616 |
+
"grad_norm": 18.071590423583984,
|
| 617 |
+
"learning_rate": 2.409705882352941e-05,
|
| 618 |
+
"loss": 5.9829,
|
| 619 |
+
"step": 870
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"epoch": 2.4444444444444446,
|
| 623 |
+
"grad_norm": 19.58377456665039,
|
| 624 |
+
"learning_rate": 2.4008823529411766e-05,
|
| 625 |
+
"loss": 5.894,
|
| 626 |
+
"step": 880
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 2.4722222222222223,
|
| 630 |
+
"grad_norm": 20.550006866455078,
|
| 631 |
+
"learning_rate": 2.3920588235294118e-05,
|
| 632 |
+
"loss": 6.0293,
|
| 633 |
+
"step": 890
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 2.5,
|
| 637 |
+
"grad_norm": 22.616382598876953,
|
| 638 |
+
"learning_rate": 2.3832352941176473e-05,
|
| 639 |
+
"loss": 6.037,
|
| 640 |
+
"step": 900
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 2.5277777777777777,
|
| 644 |
+
"grad_norm": 22.372953414916992,
|
| 645 |
+
"learning_rate": 2.3744117647058824e-05,
|
| 646 |
+
"loss": 5.9353,
|
| 647 |
+
"step": 910
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 2.5555555555555554,
|
| 651 |
+
"grad_norm": 18.03106117248535,
|
| 652 |
+
"learning_rate": 2.365588235294118e-05,
|
| 653 |
+
"loss": 5.8287,
|
| 654 |
+
"step": 920
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 2.5833333333333335,
|
| 658 |
+
"grad_norm": 16.170204162597656,
|
| 659 |
+
"learning_rate": 2.356764705882353e-05,
|
| 660 |
+
"loss": 5.7742,
|
| 661 |
+
"step": 930
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 2.611111111111111,
|
| 665 |
+
"grad_norm": 17.15852165222168,
|
| 666 |
+
"learning_rate": 2.3479411764705882e-05,
|
| 667 |
+
"loss": 6.0491,
|
| 668 |
+
"step": 940
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 2.638888888888889,
|
| 672 |
+
"grad_norm": 20.76968002319336,
|
| 673 |
+
"learning_rate": 2.3391176470588234e-05,
|
| 674 |
+
"loss": 5.9122,
|
| 675 |
+
"step": 950
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 2.6666666666666665,
|
| 679 |
+
"grad_norm": 21.134782791137695,
|
| 680 |
+
"learning_rate": 2.330294117647059e-05,
|
| 681 |
+
"loss": 5.8734,
|
| 682 |
+
"step": 960
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 2.6944444444444446,
|
| 686 |
+
"grad_norm": 18.72968864440918,
|
| 687 |
+
"learning_rate": 2.321470588235294e-05,
|
| 688 |
+
"loss": 5.9093,
|
| 689 |
+
"step": 970
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 2.7222222222222223,
|
| 693 |
+
"grad_norm": 23.75607681274414,
|
| 694 |
+
"learning_rate": 2.3126470588235296e-05,
|
| 695 |
+
"loss": 5.9551,
|
| 696 |
+
"step": 980
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 2.75,
|
| 700 |
+
"grad_norm": 19.960378646850586,
|
| 701 |
+
"learning_rate": 2.3038235294117647e-05,
|
| 702 |
+
"loss": 6.0045,
|
| 703 |
+
"step": 990
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 2.7777777777777777,
|
| 707 |
+
"grad_norm": 16.736736297607422,
|
| 708 |
+
"learning_rate": 2.2950000000000002e-05,
|
| 709 |
+
"loss": 5.9759,
|
| 710 |
+
"step": 1000
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 2.8055555555555554,
|
| 714 |
+
"grad_norm": 22.561372756958008,
|
| 715 |
+
"learning_rate": 2.2861764705882354e-05,
|
| 716 |
+
"loss": 5.9796,
|
| 717 |
+
"step": 1010
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 2.8333333333333335,
|
| 721 |
+
"grad_norm": 16.96518898010254,
|
| 722 |
+
"learning_rate": 2.277352941176471e-05,
|
| 723 |
+
"loss": 6.0321,
|
| 724 |
+
"step": 1020
|
| 725 |
+
},
|
| 726 |
+
{
|
| 727 |
+
"epoch": 2.861111111111111,
|
| 728 |
+
"grad_norm": 20.031225204467773,
|
| 729 |
+
"learning_rate": 2.268529411764706e-05,
|
| 730 |
+
"loss": 5.9221,
|
| 731 |
+
"step": 1030
|
| 732 |
+
},
|
| 733 |
+
{
|
| 734 |
+
"epoch": 2.888888888888889,
|
| 735 |
+
"grad_norm": 16.219301223754883,
|
| 736 |
+
"learning_rate": 2.2597058823529415e-05,
|
| 737 |
+
"loss": 6.0382,
|
| 738 |
+
"step": 1040
|
| 739 |
+
},
|
| 740 |
+
{
|
| 741 |
+
"epoch": 2.9166666666666665,
|
| 742 |
+
"grad_norm": 19.537729263305664,
|
| 743 |
+
"learning_rate": 2.2508823529411764e-05,
|
| 744 |
+
"loss": 5.8556,
|
| 745 |
+
"step": 1050
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 2.9444444444444446,
|
| 749 |
+
"grad_norm": 20.713224411010742,
|
| 750 |
+
"learning_rate": 2.242058823529412e-05,
|
| 751 |
+
"loss": 5.9832,
|
| 752 |
+
"step": 1060
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 2.9722222222222223,
|
| 756 |
+
"grad_norm": 18.94963836669922,
|
| 757 |
+
"learning_rate": 2.233235294117647e-05,
|
| 758 |
+
"loss": 6.0023,
|
| 759 |
+
"step": 1070
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 3.0,
|
| 763 |
+
"grad_norm": 20.285676956176758,
|
| 764 |
+
"learning_rate": 2.2244117647058825e-05,
|
| 765 |
+
"loss": 5.8155,
|
| 766 |
+
"step": 1080
|
| 767 |
+
},
|
| 768 |
+
{
|
| 769 |
+
"epoch": 3.0277777777777777,
|
| 770 |
+
"grad_norm": 17.27924346923828,
|
| 771 |
+
"learning_rate": 2.2155882352941177e-05,
|
| 772 |
+
"loss": 5.2819,
|
| 773 |
+
"step": 1090
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"epoch": 3.0555555555555554,
|
| 777 |
+
"grad_norm": 21.477184295654297,
|
| 778 |
+
"learning_rate": 2.206764705882353e-05,
|
| 779 |
+
"loss": 5.1979,
|
| 780 |
+
"step": 1100
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"epoch": 3.0833333333333335,
|
| 784 |
+
"grad_norm": 15.144262313842773,
|
| 785 |
+
"learning_rate": 2.1979411764705883e-05,
|
| 786 |
+
"loss": 5.1172,
|
| 787 |
+
"step": 1110
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"epoch": 3.111111111111111,
|
| 791 |
+
"grad_norm": 22.37119483947754,
|
| 792 |
+
"learning_rate": 2.1891176470588238e-05,
|
| 793 |
+
"loss": 5.2262,
|
| 794 |
+
"step": 1120
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 3.138888888888889,
|
| 798 |
+
"grad_norm": 20.329545974731445,
|
| 799 |
+
"learning_rate": 2.180294117647059e-05,
|
| 800 |
+
"loss": 5.2281,
|
| 801 |
+
"step": 1130
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 3.1666666666666665,
|
| 805 |
+
"grad_norm": 17.549114227294922,
|
| 806 |
+
"learning_rate": 2.1714705882352945e-05,
|
| 807 |
+
"loss": 5.1716,
|
| 808 |
+
"step": 1140
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"epoch": 3.1944444444444446,
|
| 812 |
+
"grad_norm": 16.489368438720703,
|
| 813 |
+
"learning_rate": 2.1626470588235293e-05,
|
| 814 |
+
"loss": 5.3562,
|
| 815 |
+
"step": 1150
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"epoch": 3.2222222222222223,
|
| 819 |
+
"grad_norm": 18.964805603027344,
|
| 820 |
+
"learning_rate": 2.1538235294117648e-05,
|
| 821 |
+
"loss": 5.5189,
|
| 822 |
+
"step": 1160
|
| 823 |
+
},
|
| 824 |
+
{
|
| 825 |
+
"epoch": 3.25,
|
| 826 |
+
"grad_norm": 18.049579620361328,
|
| 827 |
+
"learning_rate": 2.145e-05,
|
| 828 |
+
"loss": 5.1967,
|
| 829 |
+
"step": 1170
|
| 830 |
+
},
|
| 831 |
+
{
|
| 832 |
+
"epoch": 3.2777777777777777,
|
| 833 |
+
"grad_norm": 18.76259422302246,
|
| 834 |
+
"learning_rate": 2.136176470588235e-05,
|
| 835 |
+
"loss": 5.2686,
|
| 836 |
+
"step": 1180
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 3.3055555555555554,
|
| 840 |
+
"grad_norm": 20.977890014648438,
|
| 841 |
+
"learning_rate": 2.1273529411764706e-05,
|
| 842 |
+
"loss": 5.1385,
|
| 843 |
+
"step": 1190
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 3.3333333333333335,
|
| 847 |
+
"grad_norm": 19.02361488342285,
|
| 848 |
+
"learning_rate": 2.1185294117647058e-05,
|
| 849 |
+
"loss": 5.3758,
|
| 850 |
+
"step": 1200
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"epoch": 3.361111111111111,
|
| 854 |
+
"grad_norm": 17.559446334838867,
|
| 855 |
+
"learning_rate": 2.1097058823529413e-05,
|
| 856 |
+
"loss": 5.3187,
|
| 857 |
+
"step": 1210
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 3.388888888888889,
|
| 861 |
+
"grad_norm": 21.838626861572266,
|
| 862 |
+
"learning_rate": 2.1008823529411764e-05,
|
| 863 |
+
"loss": 5.3414,
|
| 864 |
+
"step": 1220
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 3.4166666666666665,
|
| 868 |
+
"grad_norm": 19.832502365112305,
|
| 869 |
+
"learning_rate": 2.092058823529412e-05,
|
| 870 |
+
"loss": 5.3392,
|
| 871 |
+
"step": 1230
|
| 872 |
+
},
|
| 873 |
+
{
|
| 874 |
+
"epoch": 3.4444444444444446,
|
| 875 |
+
"grad_norm": 17.20013427734375,
|
| 876 |
+
"learning_rate": 2.083235294117647e-05,
|
| 877 |
+
"loss": 5.2786,
|
| 878 |
+
"step": 1240
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 3.4722222222222223,
|
| 882 |
+
"grad_norm": 19.66887855529785,
|
| 883 |
+
"learning_rate": 2.0744117647058826e-05,
|
| 884 |
+
"loss": 5.3352,
|
| 885 |
+
"step": 1250
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 3.5,
|
| 889 |
+
"grad_norm": 19.046735763549805,
|
| 890 |
+
"learning_rate": 2.0655882352941174e-05,
|
| 891 |
+
"loss": 5.2161,
|
| 892 |
+
"step": 1260
|
| 893 |
+
},
|
| 894 |
+
{
|
| 895 |
+
"epoch": 3.5277777777777777,
|
| 896 |
+
"grad_norm": 17.690959930419922,
|
| 897 |
+
"learning_rate": 2.056764705882353e-05,
|
| 898 |
+
"loss": 5.3777,
|
| 899 |
+
"step": 1270
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 3.5555555555555554,
|
| 903 |
+
"grad_norm": 19.2017822265625,
|
| 904 |
+
"learning_rate": 2.047941176470588e-05,
|
| 905 |
+
"loss": 5.3262,
|
| 906 |
+
"step": 1280
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"epoch": 3.5833333333333335,
|
| 910 |
+
"grad_norm": 22.48141098022461,
|
| 911 |
+
"learning_rate": 2.0391176470588236e-05,
|
| 912 |
+
"loss": 5.2689,
|
| 913 |
+
"step": 1290
|
| 914 |
+
},
|
| 915 |
+
{
|
| 916 |
+
"epoch": 3.611111111111111,
|
| 917 |
+
"grad_norm": 21.33321189880371,
|
| 918 |
+
"learning_rate": 2.0302941176470587e-05,
|
| 919 |
+
"loss": 5.1867,
|
| 920 |
+
"step": 1300
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 3.638888888888889,
|
| 924 |
+
"grad_norm": 20.216909408569336,
|
| 925 |
+
"learning_rate": 2.0214705882352942e-05,
|
| 926 |
+
"loss": 5.3314,
|
| 927 |
+
"step": 1310
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 3.6666666666666665,
|
| 931 |
+
"grad_norm": 19.15237808227539,
|
| 932 |
+
"learning_rate": 2.0126470588235294e-05,
|
| 933 |
+
"loss": 5.3206,
|
| 934 |
+
"step": 1320
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 3.6944444444444446,
|
| 938 |
+
"grad_norm": 17.66857147216797,
|
| 939 |
+
"learning_rate": 2.003823529411765e-05,
|
| 940 |
+
"loss": 5.4498,
|
| 941 |
+
"step": 1330
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 3.7222222222222223,
|
| 945 |
+
"grad_norm": 20.253646850585938,
|
| 946 |
+
"learning_rate": 1.995e-05,
|
| 947 |
+
"loss": 5.4104,
|
| 948 |
+
"step": 1340
|
| 949 |
+
},
|
| 950 |
+
{
|
| 951 |
+
"epoch": 3.75,
|
| 952 |
+
"grad_norm": 19.723072052001953,
|
| 953 |
+
"learning_rate": 1.9861764705882355e-05,
|
| 954 |
+
"loss": 5.214,
|
| 955 |
+
"step": 1350
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 3.7777777777777777,
|
| 959 |
+
"grad_norm": 17.601451873779297,
|
| 960 |
+
"learning_rate": 1.9773529411764704e-05,
|
| 961 |
+
"loss": 5.356,
|
| 962 |
+
"step": 1360
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 3.8055555555555554,
|
| 966 |
+
"grad_norm": 19.1397705078125,
|
| 967 |
+
"learning_rate": 1.968529411764706e-05,
|
| 968 |
+
"loss": 5.4965,
|
| 969 |
+
"step": 1370
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 3.8333333333333335,
|
| 973 |
+
"grad_norm": 18.656843185424805,
|
| 974 |
+
"learning_rate": 1.959705882352941e-05,
|
| 975 |
+
"loss": 5.0704,
|
| 976 |
+
"step": 1380
|
| 977 |
+
},
|
| 978 |
+
{
|
| 979 |
+
"epoch": 3.861111111111111,
|
| 980 |
+
"grad_norm": 16.182132720947266,
|
| 981 |
+
"learning_rate": 1.9508823529411765e-05,
|
| 982 |
+
"loss": 5.1655,
|
| 983 |
+
"step": 1390
|
| 984 |
+
},
|
| 985 |
+
{
|
| 986 |
+
"epoch": 3.888888888888889,
|
| 987 |
+
"grad_norm": 18.828876495361328,
|
| 988 |
+
"learning_rate": 1.9420588235294117e-05,
|
| 989 |
+
"loss": 5.0816,
|
| 990 |
+
"step": 1400
|
| 991 |
+
},
|
| 992 |
+
{
|
| 993 |
+
"epoch": 3.9166666666666665,
|
| 994 |
+
"grad_norm": 18.60540199279785,
|
| 995 |
+
"learning_rate": 1.9332352941176472e-05,
|
| 996 |
+
"loss": 5.2968,
|
| 997 |
+
"step": 1410
|
| 998 |
+
},
|
| 999 |
+
{
|
| 1000 |
+
"epoch": 3.9444444444444446,
|
| 1001 |
+
"grad_norm": 17.49435806274414,
|
| 1002 |
+
"learning_rate": 1.9244117647058823e-05,
|
| 1003 |
+
"loss": 5.3383,
|
| 1004 |
+
"step": 1420
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 3.9722222222222223,
|
| 1008 |
+
"grad_norm": 19.32575798034668,
|
| 1009 |
+
"learning_rate": 1.915588235294118e-05,
|
| 1010 |
+
"loss": 5.1811,
|
| 1011 |
+
"step": 1430
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 4.0,
|
| 1015 |
+
"grad_norm": 23.117151260375977,
|
| 1016 |
+
"learning_rate": 1.906764705882353e-05,
|
| 1017 |
+
"loss": 5.3958,
|
| 1018 |
+
"step": 1440
|
| 1019 |
+
},
|
| 1020 |
+
{
|
| 1021 |
+
"epoch": 4.027777777777778,
|
| 1022 |
+
"grad_norm": 14.918752670288086,
|
| 1023 |
+
"learning_rate": 1.8979411764705885e-05,
|
| 1024 |
+
"loss": 4.7873,
|
| 1025 |
+
"step": 1450
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 4.055555555555555,
|
| 1029 |
+
"grad_norm": 16.663442611694336,
|
| 1030 |
+
"learning_rate": 1.8891176470588236e-05,
|
| 1031 |
+
"loss": 4.812,
|
| 1032 |
+
"step": 1460
|
| 1033 |
+
},
|
| 1034 |
+
{
|
| 1035 |
+
"epoch": 4.083333333333333,
|
| 1036 |
+
"grad_norm": 17.157150268554688,
|
| 1037 |
+
"learning_rate": 1.8802941176470588e-05,
|
| 1038 |
+
"loss": 4.7987,
|
| 1039 |
+
"step": 1470
|
| 1040 |
+
},
|
| 1041 |
+
{
|
| 1042 |
+
"epoch": 4.111111111111111,
|
| 1043 |
+
"grad_norm": 18.143966674804688,
|
| 1044 |
+
"learning_rate": 1.871470588235294e-05,
|
| 1045 |
+
"loss": 4.6237,
|
| 1046 |
+
"step": 1480
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 4.138888888888889,
|
| 1050 |
+
"grad_norm": 16.317506790161133,
|
| 1051 |
+
"learning_rate": 1.8626470588235295e-05,
|
| 1052 |
+
"loss": 4.6253,
|
| 1053 |
+
"step": 1490
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 4.166666666666667,
|
| 1057 |
+
"grad_norm": 23.109792709350586,
|
| 1058 |
+
"learning_rate": 1.8538235294117646e-05,
|
| 1059 |
+
"loss": 4.7575,
|
| 1060 |
+
"step": 1500
|
| 1061 |
+
},
|
| 1062 |
+
{
|
| 1063 |
+
"epoch": 4.194444444444445,
|
| 1064 |
+
"grad_norm": 16.878374099731445,
|
| 1065 |
+
"learning_rate": 1.845e-05,
|
| 1066 |
+
"loss": 4.6636,
|
| 1067 |
+
"step": 1510
|
| 1068 |
+
},
|
| 1069 |
+
{
|
| 1070 |
+
"epoch": 4.222222222222222,
|
| 1071 |
+
"grad_norm": 18.059541702270508,
|
| 1072 |
+
"learning_rate": 1.8361764705882353e-05,
|
| 1073 |
+
"loss": 4.7979,
|
| 1074 |
+
"step": 1520
|
| 1075 |
+
},
|
| 1076 |
+
{
|
| 1077 |
+
"epoch": 4.25,
|
| 1078 |
+
"grad_norm": 16.914222717285156,
|
| 1079 |
+
"learning_rate": 1.8273529411764708e-05,
|
| 1080 |
+
"loss": 4.7786,
|
| 1081 |
+
"step": 1530
|
| 1082 |
+
},
|
| 1083 |
+
{
|
| 1084 |
+
"epoch": 4.277777777777778,
|
| 1085 |
+
"grad_norm": 18.730314254760742,
|
| 1086 |
+
"learning_rate": 1.818529411764706e-05,
|
| 1087 |
+
"loss": 4.7931,
|
| 1088 |
+
"step": 1540
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 4.305555555555555,
|
| 1092 |
+
"grad_norm": 17.19397735595703,
|
| 1093 |
+
"learning_rate": 1.8097058823529414e-05,
|
| 1094 |
+
"loss": 4.6813,
|
| 1095 |
+
"step": 1550
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 4.333333333333333,
|
| 1099 |
+
"grad_norm": 17.824308395385742,
|
| 1100 |
+
"learning_rate": 1.8008823529411766e-05,
|
| 1101 |
+
"loss": 4.8799,
|
| 1102 |
+
"step": 1560
|
| 1103 |
+
},
|
| 1104 |
+
{
|
| 1105 |
+
"epoch": 4.361111111111111,
|
| 1106 |
+
"grad_norm": 17.733112335205078,
|
| 1107 |
+
"learning_rate": 1.792058823529412e-05,
|
| 1108 |
+
"loss": 4.7475,
|
| 1109 |
+
"step": 1570
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 4.388888888888889,
|
| 1113 |
+
"grad_norm": 17.163602828979492,
|
| 1114 |
+
"learning_rate": 1.783235294117647e-05,
|
| 1115 |
+
"loss": 4.8217,
|
| 1116 |
+
"step": 1580
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 4.416666666666667,
|
| 1120 |
+
"grad_norm": 19.913904190063477,
|
| 1121 |
+
"learning_rate": 1.7744117647058824e-05,
|
| 1122 |
+
"loss": 4.8949,
|
| 1123 |
+
"step": 1590
|
| 1124 |
+
},
|
| 1125 |
+
{
|
| 1126 |
+
"epoch": 4.444444444444445,
|
| 1127 |
+
"grad_norm": 18.19684600830078,
|
| 1128 |
+
"learning_rate": 1.7655882352941176e-05,
|
| 1129 |
+
"loss": 4.9527,
|
| 1130 |
+
"step": 1600
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 4.472222222222222,
|
| 1134 |
+
"grad_norm": 19.120418548583984,
|
| 1135 |
+
"learning_rate": 1.756764705882353e-05,
|
| 1136 |
+
"loss": 4.8928,
|
| 1137 |
+
"step": 1610
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 4.5,
|
| 1141 |
+
"grad_norm": 18.0157470703125,
|
| 1142 |
+
"learning_rate": 1.7479411764705882e-05,
|
| 1143 |
+
"loss": 4.8177,
|
| 1144 |
+
"step": 1620
|
| 1145 |
+
},
|
| 1146 |
+
{
|
| 1147 |
+
"epoch": 4.527777777777778,
|
| 1148 |
+
"grad_norm": 20.340044021606445,
|
| 1149 |
+
"learning_rate": 1.7391176470588237e-05,
|
| 1150 |
+
"loss": 4.6013,
|
| 1151 |
+
"step": 1630
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 4.555555555555555,
|
| 1155 |
+
"grad_norm": 16.55755043029785,
|
| 1156 |
+
"learning_rate": 1.730294117647059e-05,
|
| 1157 |
+
"loss": 4.8209,
|
| 1158 |
+
"step": 1640
|
| 1159 |
+
},
|
| 1160 |
+
{
|
| 1161 |
+
"epoch": 4.583333333333333,
|
| 1162 |
+
"grad_norm": 17.506793975830078,
|
| 1163 |
+
"learning_rate": 1.7214705882352944e-05,
|
| 1164 |
+
"loss": 5.0675,
|
| 1165 |
+
"step": 1650
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 4.611111111111111,
|
| 1169 |
+
"grad_norm": 22.037546157836914,
|
| 1170 |
+
"learning_rate": 1.7126470588235295e-05,
|
| 1171 |
+
"loss": 4.8392,
|
| 1172 |
+
"step": 1660
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 4.638888888888889,
|
| 1176 |
+
"grad_norm": 17.09819221496582,
|
| 1177 |
+
"learning_rate": 1.703823529411765e-05,
|
| 1178 |
+
"loss": 4.7789,
|
| 1179 |
+
"step": 1670
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 4.666666666666667,
|
| 1183 |
+
"grad_norm": 23.13201904296875,
|
| 1184 |
+
"learning_rate": 1.695e-05,
|
| 1185 |
+
"loss": 4.8017,
|
| 1186 |
+
"step": 1680
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 4.694444444444445,
|
| 1190 |
+
"grad_norm": 19.751766204833984,
|
| 1191 |
+
"learning_rate": 1.6861764705882354e-05,
|
| 1192 |
+
"loss": 4.7095,
|
| 1193 |
+
"step": 1690
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 4.722222222222222,
|
| 1197 |
+
"grad_norm": 17.58235740661621,
|
| 1198 |
+
"learning_rate": 1.6773529411764705e-05,
|
| 1199 |
+
"loss": 4.7251,
|
| 1200 |
+
"step": 1700
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 4.75,
|
| 1204 |
+
"grad_norm": 17.987777709960938,
|
| 1205 |
+
"learning_rate": 1.668529411764706e-05,
|
| 1206 |
+
"loss": 4.8577,
|
| 1207 |
+
"step": 1710
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 4.777777777777778,
|
| 1211 |
+
"grad_norm": 21.950464248657227,
|
| 1212 |
+
"learning_rate": 1.6597058823529412e-05,
|
| 1213 |
+
"loss": 4.8194,
|
| 1214 |
+
"step": 1720
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 4.805555555555555,
|
| 1218 |
+
"grad_norm": 18.807058334350586,
|
| 1219 |
+
"learning_rate": 1.6508823529411767e-05,
|
| 1220 |
+
"loss": 4.8405,
|
| 1221 |
+
"step": 1730
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 4.833333333333333,
|
| 1225 |
+
"grad_norm": 19.066274642944336,
|
| 1226 |
+
"learning_rate": 1.642058823529412e-05,
|
| 1227 |
+
"loss": 4.8167,
|
| 1228 |
+
"step": 1740
|
| 1229 |
+
},
|
| 1230 |
+
{
|
| 1231 |
+
"epoch": 4.861111111111111,
|
| 1232 |
+
"grad_norm": 16.53117561340332,
|
| 1233 |
+
"learning_rate": 1.633235294117647e-05,
|
| 1234 |
+
"loss": 4.9061,
|
| 1235 |
+
"step": 1750
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 4.888888888888889,
|
| 1239 |
+
"grad_norm": 19.067041397094727,
|
| 1240 |
+
"learning_rate": 1.6244117647058825e-05,
|
| 1241 |
+
"loss": 4.772,
|
| 1242 |
+
"step": 1760
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 4.916666666666667,
|
| 1246 |
+
"grad_norm": 17.190061569213867,
|
| 1247 |
+
"learning_rate": 1.6155882352941177e-05,
|
| 1248 |
+
"loss": 4.8326,
|
| 1249 |
+
"step": 1770
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 4.944444444444445,
|
| 1253 |
+
"grad_norm": 22.074304580688477,
|
| 1254 |
+
"learning_rate": 1.606764705882353e-05,
|
| 1255 |
+
"loss": 5.0498,
|
| 1256 |
+
"step": 1780
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 4.972222222222222,
|
| 1260 |
+
"grad_norm": 17.11249542236328,
|
| 1261 |
+
"learning_rate": 1.597941176470588e-05,
|
| 1262 |
+
"loss": 4.8728,
|
| 1263 |
+
"step": 1790
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 5.0,
|
| 1267 |
+
"grad_norm": 17.489036560058594,
|
| 1268 |
+
"learning_rate": 1.5891176470588235e-05,
|
| 1269 |
+
"loss": 4.7712,
|
| 1270 |
+
"step": 1800
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 5.027777777777778,
|
| 1274 |
+
"grad_norm": 18.673629760742188,
|
| 1275 |
+
"learning_rate": 1.5802941176470586e-05,
|
| 1276 |
+
"loss": 4.318,
|
| 1277 |
+
"step": 1810
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 5.055555555555555,
|
| 1281 |
+
"grad_norm": 18.343915939331055,
|
| 1282 |
+
"learning_rate": 1.571470588235294e-05,
|
| 1283 |
+
"loss": 4.4046,
|
| 1284 |
+
"step": 1820
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 5.083333333333333,
|
| 1288 |
+
"grad_norm": 18.69623565673828,
|
| 1289 |
+
"learning_rate": 1.5626470588235293e-05,
|
| 1290 |
+
"loss": 4.4717,
|
| 1291 |
+
"step": 1830
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 5.111111111111111,
|
| 1295 |
+
"grad_norm": 17.767257690429688,
|
| 1296 |
+
"learning_rate": 1.5538235294117648e-05,
|
| 1297 |
+
"loss": 4.3915,
|
| 1298 |
+
"step": 1840
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 5.138888888888889,
|
| 1302 |
+
"grad_norm": 18.257091522216797,
|
| 1303 |
+
"learning_rate": 1.545e-05,
|
| 1304 |
+
"loss": 4.4195,
|
| 1305 |
+
"step": 1850
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 5.166666666666667,
|
| 1309 |
+
"grad_norm": 18.255918502807617,
|
| 1310 |
+
"learning_rate": 1.5361764705882354e-05,
|
| 1311 |
+
"loss": 4.466,
|
| 1312 |
+
"step": 1860
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"epoch": 5.194444444444445,
|
| 1316 |
+
"grad_norm": 17.16253089904785,
|
| 1317 |
+
"learning_rate": 1.5273529411764706e-05,
|
| 1318 |
+
"loss": 4.4639,
|
| 1319 |
+
"step": 1870
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 5.222222222222222,
|
| 1323 |
+
"grad_norm": 23.036542892456055,
|
| 1324 |
+
"learning_rate": 1.5185294117647061e-05,
|
| 1325 |
+
"loss": 4.4155,
|
| 1326 |
+
"step": 1880
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 5.25,
|
| 1330 |
+
"grad_norm": 16.37624168395996,
|
| 1331 |
+
"learning_rate": 1.5097058823529414e-05,
|
| 1332 |
+
"loss": 4.4856,
|
| 1333 |
+
"step": 1890
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 5.277777777777778,
|
| 1337 |
+
"grad_norm": 15.804655075073242,
|
| 1338 |
+
"learning_rate": 1.5008823529411764e-05,
|
| 1339 |
+
"loss": 4.3916,
|
| 1340 |
+
"step": 1900
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 5.305555555555555,
|
| 1344 |
+
"grad_norm": 18.654705047607422,
|
| 1345 |
+
"learning_rate": 1.492058823529412e-05,
|
| 1346 |
+
"loss": 4.5315,
|
| 1347 |
+
"step": 1910
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 5.333333333333333,
|
| 1351 |
+
"grad_norm": 17.40276336669922,
|
| 1352 |
+
"learning_rate": 1.483235294117647e-05,
|
| 1353 |
+
"loss": 4.4648,
|
| 1354 |
+
"step": 1920
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 5.361111111111111,
|
| 1358 |
+
"grad_norm": 14.804203987121582,
|
| 1359 |
+
"learning_rate": 1.4744117647058824e-05,
|
| 1360 |
+
"loss": 4.5326,
|
| 1361 |
+
"step": 1930
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 5.388888888888889,
|
| 1365 |
+
"grad_norm": 18.393539428710938,
|
| 1366 |
+
"learning_rate": 1.4655882352941177e-05,
|
| 1367 |
+
"loss": 4.3771,
|
| 1368 |
+
"step": 1940
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 5.416666666666667,
|
| 1372 |
+
"grad_norm": 21.74736213684082,
|
| 1373 |
+
"learning_rate": 1.456764705882353e-05,
|
| 1374 |
+
"loss": 4.424,
|
| 1375 |
+
"step": 1950
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 5.444444444444445,
|
| 1379 |
+
"grad_norm": 18.945545196533203,
|
| 1380 |
+
"learning_rate": 1.4479411764705884e-05,
|
| 1381 |
+
"loss": 4.3918,
|
| 1382 |
+
"step": 1960
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 5.472222222222222,
|
| 1386 |
+
"grad_norm": 21.048032760620117,
|
| 1387 |
+
"learning_rate": 1.4391176470588237e-05,
|
| 1388 |
+
"loss": 4.4165,
|
| 1389 |
+
"step": 1970
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 5.5,
|
| 1393 |
+
"grad_norm": 17.606342315673828,
|
| 1394 |
+
"learning_rate": 1.4302941176470589e-05,
|
| 1395 |
+
"loss": 4.4718,
|
| 1396 |
+
"step": 1980
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 5.527777777777778,
|
| 1400 |
+
"grad_norm": 16.075519561767578,
|
| 1401 |
+
"learning_rate": 1.4214705882352942e-05,
|
| 1402 |
+
"loss": 4.3442,
|
| 1403 |
+
"step": 1990
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 5.555555555555555,
|
| 1407 |
+
"grad_norm": 15.067805290222168,
|
| 1408 |
+
"learning_rate": 1.4126470588235295e-05,
|
| 1409 |
+
"loss": 4.4641,
|
| 1410 |
+
"step": 2000
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 5.583333333333333,
|
| 1414 |
+
"grad_norm": 16.70688819885254,
|
| 1415 |
+
"learning_rate": 1.4038235294117649e-05,
|
| 1416 |
+
"loss": 4.4619,
|
| 1417 |
+
"step": 2010
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 5.611111111111111,
|
| 1421 |
+
"grad_norm": 20.45810317993164,
|
| 1422 |
+
"learning_rate": 1.395e-05,
|
| 1423 |
+
"loss": 4.4448,
|
| 1424 |
+
"step": 2020
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 5.638888888888889,
|
| 1428 |
+
"grad_norm": 19.671323776245117,
|
| 1429 |
+
"learning_rate": 1.3861764705882352e-05,
|
| 1430 |
+
"loss": 4.5246,
|
| 1431 |
+
"step": 2030
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 5.666666666666667,
|
| 1435 |
+
"grad_norm": 17.035329818725586,
|
| 1436 |
+
"learning_rate": 1.3773529411764705e-05,
|
| 1437 |
+
"loss": 4.3425,
|
| 1438 |
+
"step": 2040
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 5.694444444444445,
|
| 1442 |
+
"grad_norm": 18.257850646972656,
|
| 1443 |
+
"learning_rate": 1.3685294117647058e-05,
|
| 1444 |
+
"loss": 4.3883,
|
| 1445 |
+
"step": 2050
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 5.722222222222222,
|
| 1449 |
+
"grad_norm": 17.29006004333496,
|
| 1450 |
+
"learning_rate": 1.3597058823529412e-05,
|
| 1451 |
+
"loss": 4.4048,
|
| 1452 |
+
"step": 2060
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"epoch": 5.75,
|
| 1456 |
+
"grad_norm": 16.57717514038086,
|
| 1457 |
+
"learning_rate": 1.3508823529411765e-05,
|
| 1458 |
+
"loss": 4.4456,
|
| 1459 |
+
"step": 2070
|
| 1460 |
+
},
|
| 1461 |
+
{
|
| 1462 |
+
"epoch": 5.777777777777778,
|
| 1463 |
+
"grad_norm": 21.21561622619629,
|
| 1464 |
+
"learning_rate": 1.3420588235294117e-05,
|
| 1465 |
+
"loss": 4.5329,
|
| 1466 |
+
"step": 2080
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 5.805555555555555,
|
| 1470 |
+
"grad_norm": 18.407976150512695,
|
| 1471 |
+
"learning_rate": 1.333235294117647e-05,
|
| 1472 |
+
"loss": 4.6047,
|
| 1473 |
+
"step": 2090
|
| 1474 |
+
},
|
| 1475 |
+
{
|
| 1476 |
+
"epoch": 5.833333333333333,
|
| 1477 |
+
"grad_norm": 15.929880142211914,
|
| 1478 |
+
"learning_rate": 1.3244117647058823e-05,
|
| 1479 |
+
"loss": 4.421,
|
| 1480 |
+
"step": 2100
|
| 1481 |
+
},
|
| 1482 |
+
{
|
| 1483 |
+
"epoch": 5.861111111111111,
|
| 1484 |
+
"grad_norm": 16.92407989501953,
|
| 1485 |
+
"learning_rate": 1.3155882352941176e-05,
|
| 1486 |
+
"loss": 4.4857,
|
| 1487 |
+
"step": 2110
|
| 1488 |
+
},
|
| 1489 |
+
{
|
| 1490 |
+
"epoch": 5.888888888888889,
|
| 1491 |
+
"grad_norm": 17.74669647216797,
|
| 1492 |
+
"learning_rate": 1.306764705882353e-05,
|
| 1493 |
+
"loss": 4.4654,
|
| 1494 |
+
"step": 2120
|
| 1495 |
+
},
|
| 1496 |
+
{
|
| 1497 |
+
"epoch": 5.916666666666667,
|
| 1498 |
+
"grad_norm": 19.744373321533203,
|
| 1499 |
+
"learning_rate": 1.2979411764705881e-05,
|
| 1500 |
+
"loss": 4.3837,
|
| 1501 |
+
"step": 2130
|
| 1502 |
+
},
|
| 1503 |
+
{
|
| 1504 |
+
"epoch": 5.944444444444445,
|
| 1505 |
+
"grad_norm": 15.323151588439941,
|
| 1506 |
+
"learning_rate": 1.2891176470588235e-05,
|
| 1507 |
+
"loss": 4.3457,
|
| 1508 |
+
"step": 2140
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 5.972222222222222,
|
| 1512 |
+
"grad_norm": 18.93714141845703,
|
| 1513 |
+
"learning_rate": 1.2802941176470588e-05,
|
| 1514 |
+
"loss": 4.3657,
|
| 1515 |
+
"step": 2150
|
| 1516 |
+
},
|
| 1517 |
+
{
|
| 1518 |
+
"epoch": 6.0,
|
| 1519 |
+
"grad_norm": 19.87427520751953,
|
| 1520 |
+
"learning_rate": 1.2714705882352941e-05,
|
| 1521 |
+
"loss": 4.4179,
|
| 1522 |
+
"step": 2160
|
| 1523 |
+
},
|
| 1524 |
+
{
|
| 1525 |
+
"epoch": 6.027777777777778,
|
| 1526 |
+
"grad_norm": 15.826274871826172,
|
| 1527 |
+
"learning_rate": 1.2626470588235295e-05,
|
| 1528 |
+
"loss": 4.1858,
|
| 1529 |
+
"step": 2170
|
| 1530 |
+
},
|
| 1531 |
+
{
|
| 1532 |
+
"epoch": 6.055555555555555,
|
| 1533 |
+
"grad_norm": 16.08074378967285,
|
| 1534 |
+
"learning_rate": 1.2538235294117648e-05,
|
| 1535 |
+
"loss": 4.1317,
|
| 1536 |
+
"step": 2180
|
| 1537 |
+
},
|
| 1538 |
+
{
|
| 1539 |
+
"epoch": 6.083333333333333,
|
| 1540 |
+
"grad_norm": 15.976150512695312,
|
| 1541 |
+
"learning_rate": 1.245e-05,
|
| 1542 |
+
"loss": 4.1214,
|
| 1543 |
+
"step": 2190
|
| 1544 |
+
},
|
| 1545 |
+
{
|
| 1546 |
+
"epoch": 6.111111111111111,
|
| 1547 |
+
"grad_norm": 17.065427780151367,
|
| 1548 |
+
"learning_rate": 1.2361764705882353e-05,
|
| 1549 |
+
"loss": 4.0471,
|
| 1550 |
+
"step": 2200
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 6.138888888888889,
|
| 1554 |
+
"grad_norm": 15.556553840637207,
|
| 1555 |
+
"learning_rate": 1.2273529411764706e-05,
|
| 1556 |
+
"loss": 4.0961,
|
| 1557 |
+
"step": 2210
|
| 1558 |
+
},
|
| 1559 |
+
{
|
| 1560 |
+
"epoch": 6.166666666666667,
|
| 1561 |
+
"grad_norm": 16.178955078125,
|
| 1562 |
+
"learning_rate": 1.218529411764706e-05,
|
| 1563 |
+
"loss": 4.1129,
|
| 1564 |
+
"step": 2220
|
| 1565 |
+
},
|
| 1566 |
+
{
|
| 1567 |
+
"epoch": 6.194444444444445,
|
| 1568 |
+
"grad_norm": 17.971904754638672,
|
| 1569 |
+
"learning_rate": 1.2097058823529413e-05,
|
| 1570 |
+
"loss": 4.111,
|
| 1571 |
+
"step": 2230
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 6.222222222222222,
|
| 1575 |
+
"grad_norm": 17.23676109313965,
|
| 1576 |
+
"learning_rate": 1.2008823529411764e-05,
|
| 1577 |
+
"loss": 4.209,
|
| 1578 |
+
"step": 2240
|
| 1579 |
+
},
|
| 1580 |
+
{
|
| 1581 |
+
"epoch": 6.25,
|
| 1582 |
+
"grad_norm": 18.453350067138672,
|
| 1583 |
+
"learning_rate": 1.1920588235294117e-05,
|
| 1584 |
+
"loss": 4.1997,
|
| 1585 |
+
"step": 2250
|
| 1586 |
+
},
|
| 1587 |
+
{
|
| 1588 |
+
"epoch": 6.277777777777778,
|
| 1589 |
+
"grad_norm": 17.884140014648438,
|
| 1590 |
+
"learning_rate": 1.183235294117647e-05,
|
| 1591 |
+
"loss": 4.1343,
|
| 1592 |
+
"step": 2260
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 6.305555555555555,
|
| 1596 |
+
"grad_norm": 18.340280532836914,
|
| 1597 |
+
"learning_rate": 1.1744117647058824e-05,
|
| 1598 |
+
"loss": 4.1262,
|
| 1599 |
+
"step": 2270
|
| 1600 |
+
},
|
| 1601 |
+
{
|
| 1602 |
+
"epoch": 6.333333333333333,
|
| 1603 |
+
"grad_norm": 18.307395935058594,
|
| 1604 |
+
"learning_rate": 1.1655882352941177e-05,
|
| 1605 |
+
"loss": 4.1307,
|
| 1606 |
+
"step": 2280
|
| 1607 |
+
},
|
| 1608 |
+
{
|
| 1609 |
+
"epoch": 6.361111111111111,
|
| 1610 |
+
"grad_norm": 17.725727081298828,
|
| 1611 |
+
"learning_rate": 1.1567647058823529e-05,
|
| 1612 |
+
"loss": 4.0369,
|
| 1613 |
+
"step": 2290
|
| 1614 |
+
},
|
| 1615 |
+
{
|
| 1616 |
+
"epoch": 6.388888888888889,
|
| 1617 |
+
"grad_norm": 14.209752082824707,
|
| 1618 |
+
"learning_rate": 1.1479411764705882e-05,
|
| 1619 |
+
"loss": 4.1071,
|
| 1620 |
+
"step": 2300
|
| 1621 |
+
},
|
| 1622 |
+
{
|
| 1623 |
+
"epoch": 6.416666666666667,
|
| 1624 |
+
"grad_norm": 21.862701416015625,
|
| 1625 |
+
"learning_rate": 1.1391176470588235e-05,
|
| 1626 |
+
"loss": 4.223,
|
| 1627 |
+
"step": 2310
|
| 1628 |
+
},
|
| 1629 |
+
{
|
| 1630 |
+
"epoch": 6.444444444444445,
|
| 1631 |
+
"grad_norm": 13.577404975891113,
|
| 1632 |
+
"learning_rate": 1.1302941176470589e-05,
|
| 1633 |
+
"loss": 4.2441,
|
| 1634 |
+
"step": 2320
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 6.472222222222222,
|
| 1638 |
+
"grad_norm": 16.62024688720703,
|
| 1639 |
+
"learning_rate": 1.1214705882352942e-05,
|
| 1640 |
+
"loss": 4.1793,
|
| 1641 |
+
"step": 2330
|
| 1642 |
+
},
|
| 1643 |
+
{
|
| 1644 |
+
"epoch": 6.5,
|
| 1645 |
+
"grad_norm": 17.887189865112305,
|
| 1646 |
+
"learning_rate": 1.1126470588235295e-05,
|
| 1647 |
+
"loss": 4.1837,
|
| 1648 |
+
"step": 2340
|
| 1649 |
+
},
|
| 1650 |
+
{
|
| 1651 |
+
"epoch": 6.527777777777778,
|
| 1652 |
+
"grad_norm": 16.476367950439453,
|
| 1653 |
+
"learning_rate": 1.1038235294117647e-05,
|
| 1654 |
+
"loss": 4.1526,
|
| 1655 |
+
"step": 2350
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 6.555555555555555,
|
| 1659 |
+
"grad_norm": 17.012876510620117,
|
| 1660 |
+
"learning_rate": 1.095e-05,
|
| 1661 |
+
"loss": 4.1965,
|
| 1662 |
+
"step": 2360
|
| 1663 |
+
},
|
| 1664 |
+
{
|
| 1665 |
+
"epoch": 6.583333333333333,
|
| 1666 |
+
"grad_norm": 17.46245765686035,
|
| 1667 |
+
"learning_rate": 1.0861764705882353e-05,
|
| 1668 |
+
"loss": 4.1394,
|
| 1669 |
+
"step": 2370
|
| 1670 |
+
},
|
| 1671 |
+
{
|
| 1672 |
+
"epoch": 6.611111111111111,
|
| 1673 |
+
"grad_norm": 17.8790340423584,
|
| 1674 |
+
"learning_rate": 1.0773529411764707e-05,
|
| 1675 |
+
"loss": 4.1404,
|
| 1676 |
+
"step": 2380
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 6.638888888888889,
|
| 1680 |
+
"grad_norm": 16.047483444213867,
|
| 1681 |
+
"learning_rate": 1.068529411764706e-05,
|
| 1682 |
+
"loss": 4.1839,
|
| 1683 |
+
"step": 2390
|
| 1684 |
+
},
|
| 1685 |
+
{
|
| 1686 |
+
"epoch": 6.666666666666667,
|
| 1687 |
+
"grad_norm": 15.634733200073242,
|
| 1688 |
+
"learning_rate": 1.0597058823529412e-05,
|
| 1689 |
+
"loss": 4.1608,
|
| 1690 |
+
"step": 2400
|
| 1691 |
+
},
|
| 1692 |
+
{
|
| 1693 |
+
"epoch": 6.694444444444445,
|
| 1694 |
+
"grad_norm": 15.73880386352539,
|
| 1695 |
+
"learning_rate": 1.0508823529411765e-05,
|
| 1696 |
+
"loss": 4.1989,
|
| 1697 |
+
"step": 2410
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 6.722222222222222,
|
| 1701 |
+
"grad_norm": 18.751541137695312,
|
| 1702 |
+
"learning_rate": 1.0420588235294118e-05,
|
| 1703 |
+
"loss": 4.215,
|
| 1704 |
+
"step": 2420
|
| 1705 |
+
},
|
| 1706 |
+
{
|
| 1707 |
+
"epoch": 6.75,
|
| 1708 |
+
"grad_norm": 18.6307430267334,
|
| 1709 |
+
"learning_rate": 1.0332352941176472e-05,
|
| 1710 |
+
"loss": 4.1345,
|
| 1711 |
+
"step": 2430
|
| 1712 |
+
},
|
| 1713 |
+
{
|
| 1714 |
+
"epoch": 6.777777777777778,
|
| 1715 |
+
"grad_norm": 18.48142433166504,
|
| 1716 |
+
"learning_rate": 1.0244117647058825e-05,
|
| 1717 |
+
"loss": 4.2856,
|
| 1718 |
+
"step": 2440
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 6.805555555555555,
|
| 1722 |
+
"grad_norm": 17.484237670898438,
|
| 1723 |
+
"learning_rate": 1.0155882352941176e-05,
|
| 1724 |
+
"loss": 4.223,
|
| 1725 |
+
"step": 2450
|
| 1726 |
+
},
|
| 1727 |
+
{
|
| 1728 |
+
"epoch": 6.833333333333333,
|
| 1729 |
+
"grad_norm": 16.67469024658203,
|
| 1730 |
+
"learning_rate": 1.006764705882353e-05,
|
| 1731 |
+
"loss": 4.1748,
|
| 1732 |
+
"step": 2460
|
| 1733 |
+
},
|
| 1734 |
+
{
|
| 1735 |
+
"epoch": 6.861111111111111,
|
| 1736 |
+
"grad_norm": 18.385995864868164,
|
| 1737 |
+
"learning_rate": 9.979411764705883e-06,
|
| 1738 |
+
"loss": 4.1886,
|
| 1739 |
+
"step": 2470
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 6.888888888888889,
|
| 1743 |
+
"grad_norm": 20.845624923706055,
|
| 1744 |
+
"learning_rate": 9.891176470588236e-06,
|
| 1745 |
+
"loss": 4.1832,
|
| 1746 |
+
"step": 2480
|
| 1747 |
+
},
|
| 1748 |
+
{
|
| 1749 |
+
"epoch": 6.916666666666667,
|
| 1750 |
+
"grad_norm": 18.856298446655273,
|
| 1751 |
+
"learning_rate": 9.80294117647059e-06,
|
| 1752 |
+
"loss": 4.2403,
|
| 1753 |
+
"step": 2490
|
| 1754 |
+
},
|
| 1755 |
+
{
|
| 1756 |
+
"epoch": 6.944444444444445,
|
| 1757 |
+
"grad_norm": 16.965234756469727,
|
| 1758 |
+
"learning_rate": 9.714705882352943e-06,
|
| 1759 |
+
"loss": 4.2439,
|
| 1760 |
+
"step": 2500
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 6.972222222222222,
|
| 1764 |
+
"grad_norm": 16.04508399963379,
|
| 1765 |
+
"learning_rate": 9.626470588235294e-06,
|
| 1766 |
+
"loss": 4.1588,
|
| 1767 |
+
"step": 2510
|
| 1768 |
+
},
|
| 1769 |
+
{
|
| 1770 |
+
"epoch": 7.0,
|
| 1771 |
+
"grad_norm": 17.192468643188477,
|
| 1772 |
+
"learning_rate": 9.538235294117648e-06,
|
| 1773 |
+
"loss": 4.2357,
|
| 1774 |
+
"step": 2520
|
| 1775 |
+
},
|
| 1776 |
+
{
|
| 1777 |
+
"epoch": 7.027777777777778,
|
| 1778 |
+
"grad_norm": 14.738517761230469,
|
| 1779 |
+
"learning_rate": 9.450000000000001e-06,
|
| 1780 |
+
"loss": 3.8684,
|
| 1781 |
+
"step": 2530
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 7.055555555555555,
|
| 1785 |
+
"grad_norm": 16.8319034576416,
|
| 1786 |
+
"learning_rate": 9.361764705882354e-06,
|
| 1787 |
+
"loss": 3.8483,
|
| 1788 |
+
"step": 2540
|
| 1789 |
+
},
|
| 1790 |
+
{
|
| 1791 |
+
"epoch": 7.083333333333333,
|
| 1792 |
+
"grad_norm": 14.79024600982666,
|
| 1793 |
+
"learning_rate": 9.273529411764708e-06,
|
| 1794 |
+
"loss": 3.9072,
|
| 1795 |
+
"step": 2550
|
| 1796 |
+
},
|
| 1797 |
+
{
|
| 1798 |
+
"epoch": 7.111111111111111,
|
| 1799 |
+
"grad_norm": 12.333316802978516,
|
| 1800 |
+
"learning_rate": 9.18529411764706e-06,
|
| 1801 |
+
"loss": 3.9609,
|
| 1802 |
+
"step": 2560
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 7.138888888888889,
|
| 1806 |
+
"grad_norm": 14.567424774169922,
|
| 1807 |
+
"learning_rate": 9.097058823529412e-06,
|
| 1808 |
+
"loss": 3.9129,
|
| 1809 |
+
"step": 2570
|
| 1810 |
+
},
|
| 1811 |
+
{
|
| 1812 |
+
"epoch": 7.166666666666667,
|
| 1813 |
+
"grad_norm": 18.647308349609375,
|
| 1814 |
+
"learning_rate": 9.008823529411766e-06,
|
| 1815 |
+
"loss": 3.9278,
|
| 1816 |
+
"step": 2580
|
| 1817 |
+
},
|
| 1818 |
+
{
|
| 1819 |
+
"epoch": 7.194444444444445,
|
| 1820 |
+
"grad_norm": 16.815906524658203,
|
| 1821 |
+
"learning_rate": 8.920588235294117e-06,
|
| 1822 |
+
"loss": 3.9316,
|
| 1823 |
+
"step": 2590
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"epoch": 7.222222222222222,
|
| 1827 |
+
"grad_norm": 19.90522575378418,
|
| 1828 |
+
"learning_rate": 8.83235294117647e-06,
|
| 1829 |
+
"loss": 3.9444,
|
| 1830 |
+
"step": 2600
|
| 1831 |
+
},
|
| 1832 |
+
{
|
| 1833 |
+
"epoch": 7.25,
|
| 1834 |
+
"grad_norm": 15.773906707763672,
|
| 1835 |
+
"learning_rate": 8.744117647058822e-06,
|
| 1836 |
+
"loss": 4.037,
|
| 1837 |
+
"step": 2610
|
| 1838 |
+
},
|
| 1839 |
+
{
|
| 1840 |
+
"epoch": 7.277777777777778,
|
| 1841 |
+
"grad_norm": 14.414016723632812,
|
| 1842 |
+
"learning_rate": 8.655882352941176e-06,
|
| 1843 |
+
"loss": 3.9834,
|
| 1844 |
+
"step": 2620
|
| 1845 |
+
},
|
| 1846 |
+
{
|
| 1847 |
+
"epoch": 7.305555555555555,
|
| 1848 |
+
"grad_norm": 18.313098907470703,
|
| 1849 |
+
"learning_rate": 8.567647058823529e-06,
|
| 1850 |
+
"loss": 4.0229,
|
| 1851 |
+
"step": 2630
|
| 1852 |
+
},
|
| 1853 |
+
{
|
| 1854 |
+
"epoch": 7.333333333333333,
|
| 1855 |
+
"grad_norm": 15.021878242492676,
|
| 1856 |
+
"learning_rate": 8.479411764705882e-06,
|
| 1857 |
+
"loss": 3.9891,
|
| 1858 |
+
"step": 2640
|
| 1859 |
+
},
|
| 1860 |
+
{
|
| 1861 |
+
"epoch": 7.361111111111111,
|
| 1862 |
+
"grad_norm": 16.228864669799805,
|
| 1863 |
+
"learning_rate": 8.391176470588235e-06,
|
| 1864 |
+
"loss": 3.9871,
|
| 1865 |
+
"step": 2650
|
| 1866 |
+
},
|
| 1867 |
+
{
|
| 1868 |
+
"epoch": 7.388888888888889,
|
| 1869 |
+
"grad_norm": 11.939830780029297,
|
| 1870 |
+
"learning_rate": 8.302941176470589e-06,
|
| 1871 |
+
"loss": 3.8708,
|
| 1872 |
+
"step": 2660
|
| 1873 |
+
},
|
| 1874 |
+
{
|
| 1875 |
+
"epoch": 7.416666666666667,
|
| 1876 |
+
"grad_norm": 13.20190143585205,
|
| 1877 |
+
"learning_rate": 8.21470588235294e-06,
|
| 1878 |
+
"loss": 3.8982,
|
| 1879 |
+
"step": 2670
|
| 1880 |
+
},
|
| 1881 |
+
{
|
| 1882 |
+
"epoch": 7.444444444444445,
|
| 1883 |
+
"grad_norm": 15.111628532409668,
|
| 1884 |
+
"learning_rate": 8.126470588235294e-06,
|
| 1885 |
+
"loss": 3.9871,
|
| 1886 |
+
"step": 2680
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 7.472222222222222,
|
| 1890 |
+
"grad_norm": 15.908904075622559,
|
| 1891 |
+
"learning_rate": 8.038235294117647e-06,
|
| 1892 |
+
"loss": 3.976,
|
| 1893 |
+
"step": 2690
|
| 1894 |
+
},
|
| 1895 |
+
{
|
| 1896 |
+
"epoch": 7.5,
|
| 1897 |
+
"grad_norm": 18.098617553710938,
|
| 1898 |
+
"learning_rate": 7.95e-06,
|
| 1899 |
+
"loss": 4.0415,
|
| 1900 |
+
"step": 2700
|
| 1901 |
+
},
|
| 1902 |
+
{
|
| 1903 |
+
"epoch": 7.527777777777778,
|
| 1904 |
+
"grad_norm": 16.921981811523438,
|
| 1905 |
+
"learning_rate": 7.861764705882353e-06,
|
| 1906 |
+
"loss": 4.0462,
|
| 1907 |
+
"step": 2710
|
| 1908 |
+
},
|
| 1909 |
+
{
|
| 1910 |
+
"epoch": 7.555555555555555,
|
| 1911 |
+
"grad_norm": 14.457706451416016,
|
| 1912 |
+
"learning_rate": 7.773529411764705e-06,
|
| 1913 |
+
"loss": 3.9405,
|
| 1914 |
+
"step": 2720
|
| 1915 |
+
},
|
| 1916 |
+
{
|
| 1917 |
+
"epoch": 7.583333333333333,
|
| 1918 |
+
"grad_norm": 17.50982093811035,
|
| 1919 |
+
"learning_rate": 7.685294117647058e-06,
|
| 1920 |
+
"loss": 3.9974,
|
| 1921 |
+
"step": 2730
|
| 1922 |
+
},
|
| 1923 |
+
{
|
| 1924 |
+
"epoch": 7.611111111111111,
|
| 1925 |
+
"grad_norm": 15.285282135009766,
|
| 1926 |
+
"learning_rate": 7.5970588235294116e-06,
|
| 1927 |
+
"loss": 3.9969,
|
| 1928 |
+
"step": 2740
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 7.638888888888889,
|
| 1932 |
+
"grad_norm": 14.417180061340332,
|
| 1933 |
+
"learning_rate": 7.508823529411765e-06,
|
| 1934 |
+
"loss": 3.97,
|
| 1935 |
+
"step": 2750
|
| 1936 |
+
},
|
| 1937 |
+
{
|
| 1938 |
+
"epoch": 7.666666666666667,
|
| 1939 |
+
"grad_norm": 14.781904220581055,
|
| 1940 |
+
"learning_rate": 7.420588235294117e-06,
|
| 1941 |
+
"loss": 3.9797,
|
| 1942 |
+
"step": 2760
|
| 1943 |
+
},
|
| 1944 |
+
{
|
| 1945 |
+
"epoch": 7.694444444444445,
|
| 1946 |
+
"grad_norm": 16.317949295043945,
|
| 1947 |
+
"learning_rate": 7.332352941176471e-06,
|
| 1948 |
+
"loss": 3.9236,
|
| 1949 |
+
"step": 2770
|
| 1950 |
+
},
|
| 1951 |
+
{
|
| 1952 |
+
"epoch": 7.722222222222222,
|
| 1953 |
+
"grad_norm": 16.340526580810547,
|
| 1954 |
+
"learning_rate": 7.244117647058824e-06,
|
| 1955 |
+
"loss": 4.0323,
|
| 1956 |
+
"step": 2780
|
| 1957 |
+
},
|
| 1958 |
+
{
|
| 1959 |
+
"epoch": 7.75,
|
| 1960 |
+
"grad_norm": 13.924703598022461,
|
| 1961 |
+
"learning_rate": 7.155882352941176e-06,
|
| 1962 |
+
"loss": 3.8719,
|
| 1963 |
+
"step": 2790
|
| 1964 |
+
},
|
| 1965 |
+
{
|
| 1966 |
+
"epoch": 7.777777777777778,
|
| 1967 |
+
"grad_norm": 19.019325256347656,
|
| 1968 |
+
"learning_rate": 7.06764705882353e-06,
|
| 1969 |
+
"loss": 3.9529,
|
| 1970 |
+
"step": 2800
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 7.805555555555555,
|
| 1974 |
+
"grad_norm": 17.188499450683594,
|
| 1975 |
+
"learning_rate": 6.979411764705882e-06,
|
| 1976 |
+
"loss": 4.0392,
|
| 1977 |
+
"step": 2810
|
| 1978 |
+
},
|
| 1979 |
+
{
|
| 1980 |
+
"epoch": 7.833333333333333,
|
| 1981 |
+
"grad_norm": 13.893120765686035,
|
| 1982 |
+
"learning_rate": 6.891176470588235e-06,
|
| 1983 |
+
"loss": 3.9727,
|
| 1984 |
+
"step": 2820
|
| 1985 |
+
},
|
| 1986 |
+
{
|
| 1987 |
+
"epoch": 7.861111111111111,
|
| 1988 |
+
"grad_norm": 13.257885932922363,
|
| 1989 |
+
"learning_rate": 6.802941176470589e-06,
|
| 1990 |
+
"loss": 4.0164,
|
| 1991 |
+
"step": 2830
|
| 1992 |
+
},
|
| 1993 |
+
{
|
| 1994 |
+
"epoch": 7.888888888888889,
|
| 1995 |
+
"grad_norm": 15.29055118560791,
|
| 1996 |
+
"learning_rate": 6.714705882352941e-06,
|
| 1997 |
+
"loss": 4.0283,
|
| 1998 |
+
"step": 2840
|
| 1999 |
+
},
|
| 2000 |
+
{
|
| 2001 |
+
"epoch": 7.916666666666667,
|
| 2002 |
+
"grad_norm": 15.208207130432129,
|
| 2003 |
+
"learning_rate": 6.626470588235294e-06,
|
| 2004 |
+
"loss": 3.9515,
|
| 2005 |
+
"step": 2850
|
| 2006 |
+
},
|
| 2007 |
+
{
|
| 2008 |
+
"epoch": 7.944444444444445,
|
| 2009 |
+
"grad_norm": 15.335555076599121,
|
| 2010 |
+
"learning_rate": 6.538235294117647e-06,
|
| 2011 |
+
"loss": 4.0672,
|
| 2012 |
+
"step": 2860
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 7.972222222222222,
|
| 2016 |
+
"grad_norm": 14.554072380065918,
|
| 2017 |
+
"learning_rate": 6.45e-06,
|
| 2018 |
+
"loss": 4.0757,
|
| 2019 |
+
"step": 2870
|
| 2020 |
+
},
|
| 2021 |
+
{
|
| 2022 |
+
"epoch": 8.0,
|
| 2023 |
+
"grad_norm": 17.360929489135742,
|
| 2024 |
+
"learning_rate": 6.361764705882353e-06,
|
| 2025 |
+
"loss": 3.9486,
|
| 2026 |
+
"step": 2880
|
| 2027 |
+
},
|
| 2028 |
+
{
|
| 2029 |
+
"epoch": 8.027777777777779,
|
| 2030 |
+
"grad_norm": 11.244579315185547,
|
| 2031 |
+
"learning_rate": 6.273529411764706e-06,
|
| 2032 |
+
"loss": 3.8162,
|
| 2033 |
+
"step": 2890
|
| 2034 |
+
},
|
| 2035 |
+
{
|
| 2036 |
+
"epoch": 8.055555555555555,
|
| 2037 |
+
"grad_norm": 16.21691131591797,
|
| 2038 |
+
"learning_rate": 6.185294117647059e-06,
|
| 2039 |
+
"loss": 3.8226,
|
| 2040 |
+
"step": 2900
|
| 2041 |
+
},
|
| 2042 |
+
{
|
| 2043 |
+
"epoch": 8.083333333333334,
|
| 2044 |
+
"grad_norm": 16.201114654541016,
|
| 2045 |
+
"learning_rate": 6.097058823529412e-06,
|
| 2046 |
+
"loss": 3.7959,
|
| 2047 |
+
"step": 2910
|
| 2048 |
+
},
|
| 2049 |
+
{
|
| 2050 |
+
"epoch": 8.11111111111111,
|
| 2051 |
+
"grad_norm": 17.415403366088867,
|
| 2052 |
+
"learning_rate": 6.008823529411765e-06,
|
| 2053 |
+
"loss": 3.7922,
|
| 2054 |
+
"step": 2920
|
| 2055 |
+
},
|
| 2056 |
+
{
|
| 2057 |
+
"epoch": 8.13888888888889,
|
| 2058 |
+
"grad_norm": 10.362262725830078,
|
| 2059 |
+
"learning_rate": 5.920588235294118e-06,
|
| 2060 |
+
"loss": 3.7602,
|
| 2061 |
+
"step": 2930
|
| 2062 |
+
},
|
| 2063 |
+
{
|
| 2064 |
+
"epoch": 8.166666666666666,
|
| 2065 |
+
"grad_norm": 12.301746368408203,
|
| 2066 |
+
"learning_rate": 5.8323529411764706e-06,
|
| 2067 |
+
"loss": 3.8236,
|
| 2068 |
+
"step": 2940
|
| 2069 |
+
},
|
| 2070 |
+
{
|
| 2071 |
+
"epoch": 8.194444444444445,
|
| 2072 |
+
"grad_norm": 15.563980102539062,
|
| 2073 |
+
"learning_rate": 5.744117647058824e-06,
|
| 2074 |
+
"loss": 3.7988,
|
| 2075 |
+
"step": 2950
|
| 2076 |
+
},
|
| 2077 |
+
{
|
| 2078 |
+
"epoch": 8.222222222222221,
|
| 2079 |
+
"grad_norm": 14.712089538574219,
|
| 2080 |
+
"learning_rate": 5.655882352941177e-06,
|
| 2081 |
+
"loss": 3.8601,
|
| 2082 |
+
"step": 2960
|
| 2083 |
+
},
|
| 2084 |
+
{
|
| 2085 |
+
"epoch": 8.25,
|
| 2086 |
+
"grad_norm": 16.73733901977539,
|
| 2087 |
+
"learning_rate": 5.5676470588235296e-06,
|
| 2088 |
+
"loss": 3.7956,
|
| 2089 |
+
"step": 2970
|
| 2090 |
+
},
|
| 2091 |
+
{
|
| 2092 |
+
"epoch": 8.277777777777779,
|
| 2093 |
+
"grad_norm": 15.459855079650879,
|
| 2094 |
+
"learning_rate": 5.479411764705883e-06,
|
| 2095 |
+
"loss": 3.7525,
|
| 2096 |
+
"step": 2980
|
| 2097 |
+
},
|
| 2098 |
+
{
|
| 2099 |
+
"epoch": 8.305555555555555,
|
| 2100 |
+
"grad_norm": 16.97991943359375,
|
| 2101 |
+
"learning_rate": 5.391176470588236e-06,
|
| 2102 |
+
"loss": 3.7536,
|
| 2103 |
+
"step": 2990
|
| 2104 |
+
},
|
| 2105 |
+
{
|
| 2106 |
+
"epoch": 8.333333333333334,
|
| 2107 |
+
"grad_norm": 13.196775436401367,
|
| 2108 |
+
"learning_rate": 5.302941176470589e-06,
|
| 2109 |
+
"loss": 3.88,
|
| 2110 |
+
"step": 3000
|
| 2111 |
+
},
|
| 2112 |
+
{
|
| 2113 |
+
"epoch": 8.36111111111111,
|
| 2114 |
+
"grad_norm": 15.121062278747559,
|
| 2115 |
+
"learning_rate": 5.214705882352941e-06,
|
| 2116 |
+
"loss": 3.8269,
|
| 2117 |
+
"step": 3010
|
| 2118 |
+
},
|
| 2119 |
+
{
|
| 2120 |
+
"epoch": 8.38888888888889,
|
| 2121 |
+
"grad_norm": 17.251754760742188,
|
| 2122 |
+
"learning_rate": 5.1264705882352935e-06,
|
| 2123 |
+
"loss": 3.8579,
|
| 2124 |
+
"step": 3020
|
| 2125 |
+
},
|
| 2126 |
+
{
|
| 2127 |
+
"epoch": 8.416666666666666,
|
| 2128 |
+
"grad_norm": 14.106273651123047,
|
| 2129 |
+
"learning_rate": 5.038235294117647e-06,
|
| 2130 |
+
"loss": 3.8255,
|
| 2131 |
+
"step": 3030
|
| 2132 |
+
},
|
| 2133 |
+
{
|
| 2134 |
+
"epoch": 8.444444444444445,
|
| 2135 |
+
"grad_norm": 16.304101943969727,
|
| 2136 |
+
"learning_rate": 4.95e-06,
|
| 2137 |
+
"loss": 3.8521,
|
| 2138 |
+
"step": 3040
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 8.472222222222221,
|
| 2142 |
+
"grad_norm": 25.067447662353516,
|
| 2143 |
+
"learning_rate": 4.8617647058823525e-06,
|
| 2144 |
+
"loss": 3.8901,
|
| 2145 |
+
"step": 3050
|
| 2146 |
+
},
|
| 2147 |
+
{
|
| 2148 |
+
"epoch": 8.5,
|
| 2149 |
+
"grad_norm": 15.186288833618164,
|
| 2150 |
+
"learning_rate": 4.773529411764706e-06,
|
| 2151 |
+
"loss": 3.8275,
|
| 2152 |
+
"step": 3060
|
| 2153 |
+
},
|
| 2154 |
+
{
|
| 2155 |
+
"epoch": 8.527777777777779,
|
| 2156 |
+
"grad_norm": 14.004067420959473,
|
| 2157 |
+
"learning_rate": 4.685294117647059e-06,
|
| 2158 |
+
"loss": 3.9064,
|
| 2159 |
+
"step": 3070
|
| 2160 |
+
},
|
| 2161 |
+
{
|
| 2162 |
+
"epoch": 8.555555555555555,
|
| 2163 |
+
"grad_norm": 14.653326988220215,
|
| 2164 |
+
"learning_rate": 4.5970588235294115e-06,
|
| 2165 |
+
"loss": 3.7708,
|
| 2166 |
+
"step": 3080
|
| 2167 |
+
},
|
| 2168 |
+
{
|
| 2169 |
+
"epoch": 8.583333333333334,
|
| 2170 |
+
"grad_norm": 15.567551612854004,
|
| 2171 |
+
"learning_rate": 4.508823529411765e-06,
|
| 2172 |
+
"loss": 3.8827,
|
| 2173 |
+
"step": 3090
|
| 2174 |
+
},
|
| 2175 |
+
{
|
| 2176 |
+
"epoch": 8.61111111111111,
|
| 2177 |
+
"grad_norm": 15.521193504333496,
|
| 2178 |
+
"learning_rate": 4.420588235294117e-06,
|
| 2179 |
+
"loss": 3.7874,
|
| 2180 |
+
"step": 3100
|
| 2181 |
+
},
|
| 2182 |
+
{
|
| 2183 |
+
"epoch": 8.63888888888889,
|
| 2184 |
+
"grad_norm": 13.801298141479492,
|
| 2185 |
+
"learning_rate": 4.3323529411764705e-06,
|
| 2186 |
+
"loss": 3.8132,
|
| 2187 |
+
"step": 3110
|
| 2188 |
+
},
|
| 2189 |
+
{
|
| 2190 |
+
"epoch": 8.666666666666666,
|
| 2191 |
+
"grad_norm": 15.118513107299805,
|
| 2192 |
+
"learning_rate": 4.244117647058824e-06,
|
| 2193 |
+
"loss": 3.7819,
|
| 2194 |
+
"step": 3120
|
| 2195 |
+
},
|
| 2196 |
+
{
|
| 2197 |
+
"epoch": 8.694444444444445,
|
| 2198 |
+
"grad_norm": 14.111237525939941,
|
| 2199 |
+
"learning_rate": 4.155882352941176e-06,
|
| 2200 |
+
"loss": 3.8452,
|
| 2201 |
+
"step": 3130
|
| 2202 |
+
},
|
| 2203 |
+
{
|
| 2204 |
+
"epoch": 8.722222222222221,
|
| 2205 |
+
"grad_norm": 13.785816192626953,
|
| 2206 |
+
"learning_rate": 4.0676470588235295e-06,
|
| 2207 |
+
"loss": 3.8541,
|
| 2208 |
+
"step": 3140
|
| 2209 |
+
},
|
| 2210 |
+
{
|
| 2211 |
+
"epoch": 8.75,
|
| 2212 |
+
"grad_norm": 13.524584770202637,
|
| 2213 |
+
"learning_rate": 3.979411764705883e-06,
|
| 2214 |
+
"loss": 3.8521,
|
| 2215 |
+
"step": 3150
|
| 2216 |
+
},
|
| 2217 |
+
{
|
| 2218 |
+
"epoch": 8.777777777777779,
|
| 2219 |
+
"grad_norm": 15.52719497680664,
|
| 2220 |
+
"learning_rate": 3.891176470588235e-06,
|
| 2221 |
+
"loss": 3.8354,
|
| 2222 |
+
"step": 3160
|
| 2223 |
+
},
|
| 2224 |
+
{
|
| 2225 |
+
"epoch": 8.805555555555555,
|
| 2226 |
+
"grad_norm": 17.874053955078125,
|
| 2227 |
+
"learning_rate": 3.8029411764705885e-06,
|
| 2228 |
+
"loss": 3.794,
|
| 2229 |
+
"step": 3170
|
| 2230 |
+
},
|
| 2231 |
+
{
|
| 2232 |
+
"epoch": 8.833333333333334,
|
| 2233 |
+
"grad_norm": 15.086000442504883,
|
| 2234 |
+
"learning_rate": 3.7147058823529414e-06,
|
| 2235 |
+
"loss": 3.7661,
|
| 2236 |
+
"step": 3180
|
| 2237 |
+
},
|
| 2238 |
+
{
|
| 2239 |
+
"epoch": 8.86111111111111,
|
| 2240 |
+
"grad_norm": 18.586685180664062,
|
| 2241 |
+
"learning_rate": 3.6264705882352943e-06,
|
| 2242 |
+
"loss": 3.8781,
|
| 2243 |
+
"step": 3190
|
| 2244 |
+
},
|
| 2245 |
+
{
|
| 2246 |
+
"epoch": 8.88888888888889,
|
| 2247 |
+
"grad_norm": 16.09383201599121,
|
| 2248 |
+
"learning_rate": 3.538235294117647e-06,
|
| 2249 |
+
"loss": 3.8242,
|
| 2250 |
+
"step": 3200
|
| 2251 |
+
},
|
| 2252 |
+
{
|
| 2253 |
+
"epoch": 8.916666666666666,
|
| 2254 |
+
"grad_norm": 13.737457275390625,
|
| 2255 |
+
"learning_rate": 3.4500000000000004e-06,
|
| 2256 |
+
"loss": 3.8439,
|
| 2257 |
+
"step": 3210
|
| 2258 |
+
},
|
| 2259 |
+
{
|
| 2260 |
+
"epoch": 8.944444444444445,
|
| 2261 |
+
"grad_norm": 14.633013725280762,
|
| 2262 |
+
"learning_rate": 3.361764705882353e-06,
|
| 2263 |
+
"loss": 3.8111,
|
| 2264 |
+
"step": 3220
|
| 2265 |
+
},
|
| 2266 |
+
{
|
| 2267 |
+
"epoch": 8.972222222222221,
|
| 2268 |
+
"grad_norm": 12.145355224609375,
|
| 2269 |
+
"learning_rate": 3.2735294117647057e-06,
|
| 2270 |
+
"loss": 3.8591,
|
| 2271 |
+
"step": 3230
|
| 2272 |
+
},
|
| 2273 |
+
{
|
| 2274 |
+
"epoch": 9.0,
|
| 2275 |
+
"grad_norm": 15.101285934448242,
|
| 2276 |
+
"learning_rate": 3.1852941176470586e-06,
|
| 2277 |
+
"loss": 3.8327,
|
| 2278 |
+
"step": 3240
|
| 2279 |
+
},
|
| 2280 |
+
{
|
| 2281 |
+
"epoch": 9.027777777777779,
|
| 2282 |
+
"grad_norm": 12.12878131866455,
|
| 2283 |
+
"learning_rate": 3.097058823529412e-06,
|
| 2284 |
+
"loss": 3.7079,
|
| 2285 |
+
"step": 3250
|
| 2286 |
+
},
|
| 2287 |
+
{
|
| 2288 |
+
"epoch": 9.055555555555555,
|
| 2289 |
+
"grad_norm": 11.098604202270508,
|
| 2290 |
+
"learning_rate": 3.0088235294117647e-06,
|
| 2291 |
+
"loss": 3.6859,
|
| 2292 |
+
"step": 3260
|
| 2293 |
+
},
|
| 2294 |
+
{
|
| 2295 |
+
"epoch": 9.083333333333334,
|
| 2296 |
+
"grad_norm": 11.60837459564209,
|
| 2297 |
+
"learning_rate": 2.9205882352941176e-06,
|
| 2298 |
+
"loss": 3.6906,
|
| 2299 |
+
"step": 3270
|
| 2300 |
+
},
|
| 2301 |
+
{
|
| 2302 |
+
"epoch": 9.11111111111111,
|
| 2303 |
+
"grad_norm": 11.516129493713379,
|
| 2304 |
+
"learning_rate": 2.8323529411764705e-06,
|
| 2305 |
+
"loss": 3.7629,
|
| 2306 |
+
"step": 3280
|
| 2307 |
+
},
|
| 2308 |
+
{
|
| 2309 |
+
"epoch": 9.13888888888889,
|
| 2310 |
+
"grad_norm": 11.439606666564941,
|
| 2311 |
+
"learning_rate": 2.7441176470588238e-06,
|
| 2312 |
+
"loss": 3.7087,
|
| 2313 |
+
"step": 3290
|
| 2314 |
+
},
|
| 2315 |
+
{
|
| 2316 |
+
"epoch": 9.166666666666666,
|
| 2317 |
+
"grad_norm": 12.872289657592773,
|
| 2318 |
+
"learning_rate": 2.6558823529411766e-06,
|
| 2319 |
+
"loss": 3.754,
|
| 2320 |
+
"step": 3300
|
| 2321 |
+
},
|
| 2322 |
+
{
|
| 2323 |
+
"epoch": 9.194444444444445,
|
| 2324 |
+
"grad_norm": 13.252586364746094,
|
| 2325 |
+
"learning_rate": 2.5676470588235295e-06,
|
| 2326 |
+
"loss": 3.7477,
|
| 2327 |
+
"step": 3310
|
| 2328 |
+
},
|
| 2329 |
+
{
|
| 2330 |
+
"epoch": 9.222222222222221,
|
| 2331 |
+
"grad_norm": 14.191280364990234,
|
| 2332 |
+
"learning_rate": 2.4794117647058824e-06,
|
| 2333 |
+
"loss": 3.7703,
|
| 2334 |
+
"step": 3320
|
| 2335 |
+
},
|
| 2336 |
+
{
|
| 2337 |
+
"epoch": 9.25,
|
| 2338 |
+
"grad_norm": 17.2321720123291,
|
| 2339 |
+
"learning_rate": 2.3911764705882356e-06,
|
| 2340 |
+
"loss": 3.7569,
|
| 2341 |
+
"step": 3330
|
| 2342 |
+
},
|
| 2343 |
+
{
|
| 2344 |
+
"epoch": 9.277777777777779,
|
| 2345 |
+
"grad_norm": 15.821660995483398,
|
| 2346 |
+
"learning_rate": 2.3029411764705885e-06,
|
| 2347 |
+
"loss": 3.7321,
|
| 2348 |
+
"step": 3340
|
| 2349 |
+
},
|
| 2350 |
+
{
|
| 2351 |
+
"epoch": 9.305555555555555,
|
| 2352 |
+
"grad_norm": 10.092562675476074,
|
| 2353 |
+
"learning_rate": 2.2147058823529414e-06,
|
| 2354 |
+
"loss": 3.7033,
|
| 2355 |
+
"step": 3350
|
| 2356 |
+
},
|
| 2357 |
+
{
|
| 2358 |
+
"epoch": 9.333333333333334,
|
| 2359 |
+
"grad_norm": 19.321290969848633,
|
| 2360 |
+
"learning_rate": 2.1264705882352942e-06,
|
| 2361 |
+
"loss": 3.7683,
|
| 2362 |
+
"step": 3360
|
| 2363 |
+
},
|
| 2364 |
+
{
|
| 2365 |
+
"epoch": 9.36111111111111,
|
| 2366 |
+
"grad_norm": 15.781285285949707,
|
| 2367 |
+
"learning_rate": 2.038235294117647e-06,
|
| 2368 |
+
"loss": 3.7304,
|
| 2369 |
+
"step": 3370
|
| 2370 |
+
},
|
| 2371 |
+
{
|
| 2372 |
+
"epoch": 9.38888888888889,
|
| 2373 |
+
"grad_norm": 13.418021202087402,
|
| 2374 |
+
"learning_rate": 1.95e-06,
|
| 2375 |
+
"loss": 3.7117,
|
| 2376 |
+
"step": 3380
|
| 2377 |
+
},
|
| 2378 |
+
{
|
| 2379 |
+
"epoch": 9.416666666666666,
|
| 2380 |
+
"grad_norm": 14.020906448364258,
|
| 2381 |
+
"learning_rate": 1.861764705882353e-06,
|
| 2382 |
+
"loss": 3.7131,
|
| 2383 |
+
"step": 3390
|
| 2384 |
+
},
|
| 2385 |
+
{
|
| 2386 |
+
"epoch": 9.444444444444445,
|
| 2387 |
+
"grad_norm": 13.270752906799316,
|
| 2388 |
+
"learning_rate": 1.773529411764706e-06,
|
| 2389 |
+
"loss": 3.6744,
|
| 2390 |
+
"step": 3400
|
| 2391 |
+
},
|
| 2392 |
+
{
|
| 2393 |
+
"epoch": 9.472222222222221,
|
| 2394 |
+
"grad_norm": 12.589795112609863,
|
| 2395 |
+
"learning_rate": 1.6852941176470588e-06,
|
| 2396 |
+
"loss": 3.7544,
|
| 2397 |
+
"step": 3410
|
| 2398 |
+
},
|
| 2399 |
+
{
|
| 2400 |
+
"epoch": 9.5,
|
| 2401 |
+
"grad_norm": 13.454065322875977,
|
| 2402 |
+
"learning_rate": 1.5970588235294118e-06,
|
| 2403 |
+
"loss": 3.7782,
|
| 2404 |
+
"step": 3420
|
| 2405 |
+
},
|
| 2406 |
+
{
|
| 2407 |
+
"epoch": 9.527777777777779,
|
| 2408 |
+
"grad_norm": 14.777534484863281,
|
| 2409 |
+
"learning_rate": 1.5088235294117647e-06,
|
| 2410 |
+
"loss": 3.7203,
|
| 2411 |
+
"step": 3430
|
| 2412 |
+
},
|
| 2413 |
+
{
|
| 2414 |
+
"epoch": 9.555555555555555,
|
| 2415 |
+
"grad_norm": 14.142292022705078,
|
| 2416 |
+
"learning_rate": 1.4205882352941178e-06,
|
| 2417 |
+
"loss": 3.7331,
|
| 2418 |
+
"step": 3440
|
| 2419 |
+
},
|
| 2420 |
+
{
|
| 2421 |
+
"epoch": 9.583333333333334,
|
| 2422 |
+
"grad_norm": 17.070104598999023,
|
| 2423 |
+
"learning_rate": 1.3323529411764706e-06,
|
| 2424 |
+
"loss": 3.7318,
|
| 2425 |
+
"step": 3450
|
| 2426 |
+
},
|
| 2427 |
+
{
|
| 2428 |
+
"epoch": 9.61111111111111,
|
| 2429 |
+
"grad_norm": 13.819690704345703,
|
| 2430 |
+
"learning_rate": 1.2441176470588237e-06,
|
| 2431 |
+
"loss": 3.7379,
|
| 2432 |
+
"step": 3460
|
| 2433 |
+
},
|
| 2434 |
+
{
|
| 2435 |
+
"epoch": 9.63888888888889,
|
| 2436 |
+
"grad_norm": 13.859333038330078,
|
| 2437 |
+
"learning_rate": 1.1558823529411764e-06,
|
| 2438 |
+
"loss": 3.6641,
|
| 2439 |
+
"step": 3470
|
| 2440 |
+
},
|
| 2441 |
+
{
|
| 2442 |
+
"epoch": 9.666666666666666,
|
| 2443 |
+
"grad_norm": 12.934626579284668,
|
| 2444 |
+
"learning_rate": 1.0676470588235295e-06,
|
| 2445 |
+
"loss": 3.7164,
|
| 2446 |
+
"step": 3480
|
| 2447 |
+
},
|
| 2448 |
+
{
|
| 2449 |
+
"epoch": 9.694444444444445,
|
| 2450 |
+
"grad_norm": 13.10987377166748,
|
| 2451 |
+
"learning_rate": 9.794117647058823e-07,
|
| 2452 |
+
"loss": 3.7427,
|
| 2453 |
+
"step": 3490
|
| 2454 |
+
},
|
| 2455 |
+
{
|
| 2456 |
+
"epoch": 9.722222222222221,
|
| 2457 |
+
"grad_norm": 13.990032196044922,
|
| 2458 |
+
"learning_rate": 8.911764705882353e-07,
|
| 2459 |
+
"loss": 3.7195,
|
| 2460 |
+
"step": 3500
|
| 2461 |
+
},
|
| 2462 |
+
{
|
| 2463 |
+
"epoch": 9.75,
|
| 2464 |
+
"grad_norm": 13.015247344970703,
|
| 2465 |
+
"learning_rate": 8.029411764705883e-07,
|
| 2466 |
+
"loss": 3.7378,
|
| 2467 |
+
"step": 3510
|
| 2468 |
+
},
|
| 2469 |
+
{
|
| 2470 |
+
"epoch": 9.777777777777779,
|
| 2471 |
+
"grad_norm": 13.674323081970215,
|
| 2472 |
+
"learning_rate": 7.147058823529411e-07,
|
| 2473 |
+
"loss": 3.7118,
|
| 2474 |
+
"step": 3520
|
| 2475 |
+
},
|
| 2476 |
+
{
|
| 2477 |
+
"epoch": 9.805555555555555,
|
| 2478 |
+
"grad_norm": 11.697880744934082,
|
| 2479 |
+
"learning_rate": 6.264705882352941e-07,
|
| 2480 |
+
"loss": 3.7387,
|
| 2481 |
+
"step": 3530
|
| 2482 |
+
},
|
| 2483 |
+
{
|
| 2484 |
+
"epoch": 9.833333333333334,
|
| 2485 |
+
"grad_norm": 13.14771842956543,
|
| 2486 |
+
"learning_rate": 5.382352941176471e-07,
|
| 2487 |
+
"loss": 3.6967,
|
| 2488 |
+
"step": 3540
|
| 2489 |
+
},
|
| 2490 |
+
{
|
| 2491 |
+
"epoch": 9.86111111111111,
|
| 2492 |
+
"grad_norm": 12.617823600769043,
|
| 2493 |
+
"learning_rate": 4.5e-07,
|
| 2494 |
+
"loss": 3.7122,
|
| 2495 |
+
"step": 3550
|
| 2496 |
+
},
|
| 2497 |
+
{
|
| 2498 |
+
"epoch": 9.88888888888889,
|
| 2499 |
+
"grad_norm": 14.002346992492676,
|
| 2500 |
+
"learning_rate": 3.6176470588235295e-07,
|
| 2501 |
+
"loss": 3.7429,
|
| 2502 |
+
"step": 3560
|
| 2503 |
+
},
|
| 2504 |
+
{
|
| 2505 |
+
"epoch": 9.916666666666666,
|
| 2506 |
+
"grad_norm": 14.497004508972168,
|
| 2507 |
+
"learning_rate": 2.7352941176470587e-07,
|
| 2508 |
+
"loss": 3.7298,
|
| 2509 |
+
"step": 3570
|
| 2510 |
+
},
|
| 2511 |
+
{
|
| 2512 |
+
"epoch": 9.944444444444445,
|
| 2513 |
+
"grad_norm": 12.17514419555664,
|
| 2514 |
+
"learning_rate": 1.852941176470588e-07,
|
| 2515 |
+
"loss": 3.7254,
|
| 2516 |
+
"step": 3580
|
| 2517 |
+
},
|
| 2518 |
+
{
|
| 2519 |
+
"epoch": 9.972222222222221,
|
| 2520 |
+
"grad_norm": 17.909778594970703,
|
| 2521 |
+
"learning_rate": 9.705882352941177e-08,
|
| 2522 |
+
"loss": 3.7097,
|
| 2523 |
+
"step": 3590
|
| 2524 |
+
},
|
| 2525 |
+
{
|
| 2526 |
+
"epoch": 10.0,
|
| 2527 |
+
"grad_norm": 13.537755012512207,
|
| 2528 |
+
"learning_rate": 8.823529411764706e-09,
|
| 2529 |
+
"loss": 3.7612,
|
| 2530 |
+
"step": 3600
|
| 2531 |
+
}
|
| 2532 |
+
],
|
| 2533 |
+
"logging_steps": 10,
|
| 2534 |
+
"max_steps": 3600,
|
| 2535 |
+
"num_input_tokens_seen": 0,
|
| 2536 |
+
"num_train_epochs": 10,
|
| 2537 |
+
"save_steps": 500,
|
| 2538 |
+
"stateful_callbacks": {
|
| 2539 |
+
"TrainerControl": {
|
| 2540 |
+
"args": {
|
| 2541 |
+
"should_epoch_stop": false,
|
| 2542 |
+
"should_evaluate": false,
|
| 2543 |
+
"should_log": false,
|
| 2544 |
+
"should_save": true,
|
| 2545 |
+
"should_training_stop": true
|
| 2546 |
+
},
|
| 2547 |
+
"attributes": {}
|
| 2548 |
+
}
|
| 2549 |
+
},
|
| 2550 |
+
"total_flos": 4403097501696000.0,
|
| 2551 |
+
"train_batch_size": 4,
|
| 2552 |
+
"trial_name": null,
|
| 2553 |
+
"trial_params": null
|
| 2554 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c95faaaa0df18cd2102701ab39d413b8c034a1d3bb64489833f2711e6286cb70
|
| 3 |
+
size 5969
|
vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|