Model upload
Browse files- config.json +52 -0
- model.safetensors +3 -0
- optimizer.pt +3 -0
- rng_state.pth +3 -0
- scaler.pt +3 -0
- scheduler.pt +3 -0
- tokenizer.json +0 -0
- tokenizer_config.json +23 -0
- trainer_state.json +2888 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,52 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"DebertaV2ForSequenceClassification"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.1,
|
| 6 |
+
"bos_token_id": 1,
|
| 7 |
+
"dtype": "float32",
|
| 8 |
+
"eos_token_id": 2,
|
| 9 |
+
"hidden_act": "gelu",
|
| 10 |
+
"hidden_dropout_prob": 0.1,
|
| 11 |
+
"hidden_size": 768,
|
| 12 |
+
"id2label": {
|
| 13 |
+
"0": "LABEL_0",
|
| 14 |
+
"1": "LABEL_1",
|
| 15 |
+
"2": "LABEL_2",
|
| 16 |
+
"3": "LABEL_3"
|
| 17 |
+
},
|
| 18 |
+
"initializer_range": 0.02,
|
| 19 |
+
"intermediate_size": 3072,
|
| 20 |
+
"label2id": {
|
| 21 |
+
"LABEL_0": 0,
|
| 22 |
+
"LABEL_1": 1,
|
| 23 |
+
"LABEL_2": 2,
|
| 24 |
+
"LABEL_3": 3
|
| 25 |
+
},
|
| 26 |
+
"layer_norm_eps": 1e-07,
|
| 27 |
+
"legacy": true,
|
| 28 |
+
"max_position_embeddings": 512,
|
| 29 |
+
"max_relative_positions": -1,
|
| 30 |
+
"model_type": "deberta-v2",
|
| 31 |
+
"norm_rel_ebd": "layer_norm",
|
| 32 |
+
"num_attention_heads": 12,
|
| 33 |
+
"num_hidden_layers": 12,
|
| 34 |
+
"pad_token_id": 0,
|
| 35 |
+
"pooler_dropout": 0,
|
| 36 |
+
"pooler_hidden_act": "gelu",
|
| 37 |
+
"pooler_hidden_size": 768,
|
| 38 |
+
"pos_att_type": [
|
| 39 |
+
"p2c",
|
| 40 |
+
"c2p"
|
| 41 |
+
],
|
| 42 |
+
"position_biased_input": false,
|
| 43 |
+
"position_buckets": 256,
|
| 44 |
+
"problem_type": "multi_label_classification",
|
| 45 |
+
"relative_attention": true,
|
| 46 |
+
"share_att_key": true,
|
| 47 |
+
"tie_word_embeddings": true,
|
| 48 |
+
"transformers_version": "5.1.0",
|
| 49 |
+
"type_vocab_size": 0,
|
| 50 |
+
"use_cache": false,
|
| 51 |
+
"vocab_size": 128100
|
| 52 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0b7f0cf9ba378b5b459fbe57101a27833ba67bd754ecc2f53597a2ea6f3c4e6a
|
| 3 |
+
size 737725408
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ac5a2318cb5cc89f98690dd8d9046fd5a205b4cb18f97db1d0582dcc95680328
|
| 3 |
+
size 1475570682
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1ba7f7f687dc30519ec7e8cdb55e38be9e06deab118555e5a3b929f34adc7684
|
| 3 |
+
size 14244
|
scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e17baf6d9219c5a08d7ac5cda72e5b9a8823233175e63f12778c2a21f955f06a
|
| 3 |
+
size 988
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0ec795293513fdac88e6b6abecf1aaf8089cfb6bfc9ad777c34ffb30a2508c7f
|
| 3 |
+
size 1064
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,23 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": true,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": "[CLS]",
|
| 5 |
+
"cls_token": "[CLS]",
|
| 6 |
+
"do_lower_case": false,
|
| 7 |
+
"eos_token": "[SEP]",
|
| 8 |
+
"extra_special_tokens": [
|
| 9 |
+
"[PAD]",
|
| 10 |
+
"[CLS]",
|
| 11 |
+
"[SEP]"
|
| 12 |
+
],
|
| 13 |
+
"is_local": false,
|
| 14 |
+
"mask_token": "[MASK]",
|
| 15 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 16 |
+
"pad_token": "[PAD]",
|
| 17 |
+
"sep_token": "[SEP]",
|
| 18 |
+
"split_by_punct": false,
|
| 19 |
+
"tokenizer_class": "DebertaV2Tokenizer",
|
| 20 |
+
"unk_id": 3,
|
| 21 |
+
"unk_token": "[UNK]",
|
| 22 |
+
"vocab_type": "spm"
|
| 23 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2888 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 4839,
|
| 3 |
+
"best_metric": 0.43726749573500223,
|
| 4 |
+
"best_model_checkpoint": "constellation_one_text/checkpoint-4839",
|
| 5 |
+
"epoch": 3.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 4839,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.00744301442084044,
|
| 14 |
+
"grad_norm": 16.79511070251465,
|
| 15 |
+
"learning_rate": 4.5454545454545457e-07,
|
| 16 |
+
"loss": 5.104981422424316,
|
| 17 |
+
"step": 12
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.01488602884168088,
|
| 21 |
+
"grad_norm": 14.2466402053833,
|
| 22 |
+
"learning_rate": 9.50413223140496e-07,
|
| 23 |
+
"loss": 4.855861345926921,
|
| 24 |
+
"step": 24
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.02232904326252132,
|
| 28 |
+
"grad_norm": 15.325632095336914,
|
| 29 |
+
"learning_rate": 1.4462809917355372e-06,
|
| 30 |
+
"loss": 4.62240473429362,
|
| 31 |
+
"step": 36
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.02977205768336176,
|
| 35 |
+
"grad_norm": 14.659135818481445,
|
| 36 |
+
"learning_rate": 1.9421487603305786e-06,
|
| 37 |
+
"loss": 4.279359499613444,
|
| 38 |
+
"step": 48
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.037215072104202204,
|
| 42 |
+
"grad_norm": 11.83539867401123,
|
| 43 |
+
"learning_rate": 2.43801652892562e-06,
|
| 44 |
+
"loss": 4.009869893391927,
|
| 45 |
+
"step": 60
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.04465808652504264,
|
| 49 |
+
"grad_norm": 12.459957122802734,
|
| 50 |
+
"learning_rate": 2.9338842975206615e-06,
|
| 51 |
+
"loss": 3.658400217692057,
|
| 52 |
+
"step": 72
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.052101100945883085,
|
| 56 |
+
"grad_norm": 10.960445404052734,
|
| 57 |
+
"learning_rate": 3.429752066115703e-06,
|
| 58 |
+
"loss": 3.3341188430786133,
|
| 59 |
+
"step": 84
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.05954411536672352,
|
| 63 |
+
"grad_norm": 10.394844055175781,
|
| 64 |
+
"learning_rate": 3.925619834710744e-06,
|
| 65 |
+
"loss": 2.9902642567952475,
|
| 66 |
+
"step": 96
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 0.06698712978756396,
|
| 70 |
+
"grad_norm": 10.080375671386719,
|
| 71 |
+
"learning_rate": 4.421487603305786e-06,
|
| 72 |
+
"loss": 2.7187296549479165,
|
| 73 |
+
"step": 108
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.07443014420840441,
|
| 77 |
+
"grad_norm": 11.83609676361084,
|
| 78 |
+
"learning_rate": 4.917355371900827e-06,
|
| 79 |
+
"loss": 2.4078760147094727,
|
| 80 |
+
"step": 120
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.08187315862924484,
|
| 84 |
+
"grad_norm": 20.040725708007812,
|
| 85 |
+
"learning_rate": 5.413223140495868e-06,
|
| 86 |
+
"loss": 2.1843010584513345,
|
| 87 |
+
"step": 132
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.08931617305008528,
|
| 91 |
+
"grad_norm": 10.70347785949707,
|
| 92 |
+
"learning_rate": 5.90909090909091e-06,
|
| 93 |
+
"loss": 1.9951588312784831,
|
| 94 |
+
"step": 144
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.09675918747092573,
|
| 98 |
+
"grad_norm": 13.84825611114502,
|
| 99 |
+
"learning_rate": 6.404958677685951e-06,
|
| 100 |
+
"loss": 1.7978707949320476,
|
| 101 |
+
"step": 156
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 0.10420220189176617,
|
| 105 |
+
"grad_norm": 8.921030044555664,
|
| 106 |
+
"learning_rate": 6.900826446280993e-06,
|
| 107 |
+
"loss": 1.6856780052185059,
|
| 108 |
+
"step": 168
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 0.1116452163126066,
|
| 112 |
+
"grad_norm": 7.919989585876465,
|
| 113 |
+
"learning_rate": 7.396694214876033e-06,
|
| 114 |
+
"loss": 1.497524897257487,
|
| 115 |
+
"step": 180
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.11908823073344704,
|
| 119 |
+
"grad_norm": 15.635968208312988,
|
| 120 |
+
"learning_rate": 7.892561983471076e-06,
|
| 121 |
+
"loss": 1.4976633389790852,
|
| 122 |
+
"step": 192
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.12653124515428749,
|
| 126 |
+
"grad_norm": 14.213494300842285,
|
| 127 |
+
"learning_rate": 8.388429752066116e-06,
|
| 128 |
+
"loss": 1.4405194918314617,
|
| 129 |
+
"step": 204
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.13397425957512793,
|
| 133 |
+
"grad_norm": 10.790483474731445,
|
| 134 |
+
"learning_rate": 8.884297520661158e-06,
|
| 135 |
+
"loss": 1.2696106433868408,
|
| 136 |
+
"step": 216
|
| 137 |
+
},
|
| 138 |
+
{
|
| 139 |
+
"epoch": 0.14141727399596837,
|
| 140 |
+
"grad_norm": 14.101875305175781,
|
| 141 |
+
"learning_rate": 9.3801652892562e-06,
|
| 142 |
+
"loss": 1.3300576210021973,
|
| 143 |
+
"step": 228
|
| 144 |
+
},
|
| 145 |
+
{
|
| 146 |
+
"epoch": 0.14886028841680882,
|
| 147 |
+
"grad_norm": 19.911815643310547,
|
| 148 |
+
"learning_rate": 9.876033057851241e-06,
|
| 149 |
+
"loss": 1.2497991720835369,
|
| 150 |
+
"step": 240
|
| 151 |
+
},
|
| 152 |
+
{
|
| 153 |
+
"epoch": 0.15630330283764926,
|
| 154 |
+
"grad_norm": 12.594736099243164,
|
| 155 |
+
"learning_rate": 1.0371900826446282e-05,
|
| 156 |
+
"loss": 1.20013427734375,
|
| 157 |
+
"step": 252
|
| 158 |
+
},
|
| 159 |
+
{
|
| 160 |
+
"epoch": 0.16374631725848968,
|
| 161 |
+
"grad_norm": 10.003790855407715,
|
| 162 |
+
"learning_rate": 1.0867768595041323e-05,
|
| 163 |
+
"loss": 1.1903626918792725,
|
| 164 |
+
"step": 264
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.17118933167933012,
|
| 168 |
+
"grad_norm": 19.644290924072266,
|
| 169 |
+
"learning_rate": 1.1363636363636366e-05,
|
| 170 |
+
"loss": 1.2084464232126872,
|
| 171 |
+
"step": 276
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.17863234610017056,
|
| 175 |
+
"grad_norm": 12.33438777923584,
|
| 176 |
+
"learning_rate": 1.1859504132231406e-05,
|
| 177 |
+
"loss": 1.1396081447601318,
|
| 178 |
+
"step": 288
|
| 179 |
+
},
|
| 180 |
+
{
|
| 181 |
+
"epoch": 0.186075360521011,
|
| 182 |
+
"grad_norm": 7.845709800720215,
|
| 183 |
+
"learning_rate": 1.2355371900826447e-05,
|
| 184 |
+
"loss": 1.0346049467722576,
|
| 185 |
+
"step": 300
|
| 186 |
+
},
|
| 187 |
+
{
|
| 188 |
+
"epoch": 0.19351837494185145,
|
| 189 |
+
"grad_norm": 12.355867385864258,
|
| 190 |
+
"learning_rate": 1.2851239669421488e-05,
|
| 191 |
+
"loss": 1.0486024220784504,
|
| 192 |
+
"step": 312
|
| 193 |
+
},
|
| 194 |
+
{
|
| 195 |
+
"epoch": 0.2009613893626919,
|
| 196 |
+
"grad_norm": 9.542502403259277,
|
| 197 |
+
"learning_rate": 1.3347107438016531e-05,
|
| 198 |
+
"loss": 1.1321392059326172,
|
| 199 |
+
"step": 324
|
| 200 |
+
},
|
| 201 |
+
{
|
| 202 |
+
"epoch": 0.20840440378353234,
|
| 203 |
+
"grad_norm": 171.94647216796875,
|
| 204 |
+
"learning_rate": 1.384297520661157e-05,
|
| 205 |
+
"loss": 0.9731620152791342,
|
| 206 |
+
"step": 336
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.21584741820437278,
|
| 210 |
+
"grad_norm": 14.012189865112305,
|
| 211 |
+
"learning_rate": 1.4338842975206612e-05,
|
| 212 |
+
"loss": 0.9310257434844971,
|
| 213 |
+
"step": 348
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.2232904326252132,
|
| 217 |
+
"grad_norm": 17.743682861328125,
|
| 218 |
+
"learning_rate": 1.4834710743801655e-05,
|
| 219 |
+
"loss": 0.9263285795847574,
|
| 220 |
+
"step": 360
|
| 221 |
+
},
|
| 222 |
+
{
|
| 223 |
+
"epoch": 0.23073344704605364,
|
| 224 |
+
"grad_norm": 29.65188217163086,
|
| 225 |
+
"learning_rate": 1.5330578512396693e-05,
|
| 226 |
+
"loss": 1.0049312114715576,
|
| 227 |
+
"step": 372
|
| 228 |
+
},
|
| 229 |
+
{
|
| 230 |
+
"epoch": 0.23817646146689408,
|
| 231 |
+
"grad_norm": 16.46782684326172,
|
| 232 |
+
"learning_rate": 1.5826446280991736e-05,
|
| 233 |
+
"loss": 1.078270673751831,
|
| 234 |
+
"step": 384
|
| 235 |
+
},
|
| 236 |
+
{
|
| 237 |
+
"epoch": 0.24561947588773453,
|
| 238 |
+
"grad_norm": 15.282443046569824,
|
| 239 |
+
"learning_rate": 1.632231404958678e-05,
|
| 240 |
+
"loss": 0.9908095200856527,
|
| 241 |
+
"step": 396
|
| 242 |
+
},
|
| 243 |
+
{
|
| 244 |
+
"epoch": 0.25306249030857497,
|
| 245 |
+
"grad_norm": 7.152077674865723,
|
| 246 |
+
"learning_rate": 1.681818181818182e-05,
|
| 247 |
+
"loss": 0.8867685794830322,
|
| 248 |
+
"step": 408
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.2605055047294154,
|
| 252 |
+
"grad_norm": 17.630233764648438,
|
| 253 |
+
"learning_rate": 1.731404958677686e-05,
|
| 254 |
+
"loss": 0.8261091709136963,
|
| 255 |
+
"step": 420
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.26794851915025586,
|
| 259 |
+
"grad_norm": 8.756381034851074,
|
| 260 |
+
"learning_rate": 1.78099173553719e-05,
|
| 261 |
+
"loss": 0.8141599496205648,
|
| 262 |
+
"step": 432
|
| 263 |
+
},
|
| 264 |
+
{
|
| 265 |
+
"epoch": 0.2753915335710963,
|
| 266 |
+
"grad_norm": 14.227313041687012,
|
| 267 |
+
"learning_rate": 1.8305785123966944e-05,
|
| 268 |
+
"loss": 0.8025492032368978,
|
| 269 |
+
"step": 444
|
| 270 |
+
},
|
| 271 |
+
{
|
| 272 |
+
"epoch": 0.28283454799193675,
|
| 273 |
+
"grad_norm": 6.028214931488037,
|
| 274 |
+
"learning_rate": 1.8801652892561987e-05,
|
| 275 |
+
"loss": 0.827876885732015,
|
| 276 |
+
"step": 456
|
| 277 |
+
},
|
| 278 |
+
{
|
| 279 |
+
"epoch": 0.2902775624127772,
|
| 280 |
+
"grad_norm": 9.791404724121094,
|
| 281 |
+
"learning_rate": 1.9297520661157026e-05,
|
| 282 |
+
"loss": 0.8186439673105875,
|
| 283 |
+
"step": 468
|
| 284 |
+
},
|
| 285 |
+
{
|
| 286 |
+
"epoch": 0.29772057683361763,
|
| 287 |
+
"grad_norm": 19.028491973876953,
|
| 288 |
+
"learning_rate": 1.9793388429752066e-05,
|
| 289 |
+
"loss": 0.8027651309967041,
|
| 290 |
+
"step": 480
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.3051635912544581,
|
| 294 |
+
"grad_norm": 5.418436527252197,
|
| 295 |
+
"learning_rate": 1.996785304247991e-05,
|
| 296 |
+
"loss": 0.7800490061442057,
|
| 297 |
+
"step": 492
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.3126066056752985,
|
| 301 |
+
"grad_norm": 7.598865985870361,
|
| 302 |
+
"learning_rate": 1.9912743972445466e-05,
|
| 303 |
+
"loss": 0.7126566569010416,
|
| 304 |
+
"step": 504
|
| 305 |
+
},
|
| 306 |
+
{
|
| 307 |
+
"epoch": 0.3200496200961389,
|
| 308 |
+
"grad_norm": 7.867424011230469,
|
| 309 |
+
"learning_rate": 1.9857634902411024e-05,
|
| 310 |
+
"loss": 0.6536041895548502,
|
| 311 |
+
"step": 516
|
| 312 |
+
},
|
| 313 |
+
{
|
| 314 |
+
"epoch": 0.32749263451697935,
|
| 315 |
+
"grad_norm": 10.367350578308105,
|
| 316 |
+
"learning_rate": 1.980252583237658e-05,
|
| 317 |
+
"loss": 0.8624240557352701,
|
| 318 |
+
"step": 528
|
| 319 |
+
},
|
| 320 |
+
{
|
| 321 |
+
"epoch": 0.3349356489378198,
|
| 322 |
+
"grad_norm": 6.30031681060791,
|
| 323 |
+
"learning_rate": 1.9747416762342138e-05,
|
| 324 |
+
"loss": 0.8412895202636719,
|
| 325 |
+
"step": 540
|
| 326 |
+
},
|
| 327 |
+
{
|
| 328 |
+
"epoch": 0.34237866335866024,
|
| 329 |
+
"grad_norm": 15.809948921203613,
|
| 330 |
+
"learning_rate": 1.9692307692307696e-05,
|
| 331 |
+
"loss": 0.7370687325795492,
|
| 332 |
+
"step": 552
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.3498216777795007,
|
| 336 |
+
"grad_norm": 6.0920491218566895,
|
| 337 |
+
"learning_rate": 1.963719862227325e-05,
|
| 338 |
+
"loss": 0.7390193144480387,
|
| 339 |
+
"step": 564
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 0.3572646922003411,
|
| 343 |
+
"grad_norm": 11.583715438842773,
|
| 344 |
+
"learning_rate": 1.9582089552238807e-05,
|
| 345 |
+
"loss": 0.6651956637700399,
|
| 346 |
+
"step": 576
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 0.36470770662118157,
|
| 350 |
+
"grad_norm": 11.411588668823242,
|
| 351 |
+
"learning_rate": 1.9526980482204364e-05,
|
| 352 |
+
"loss": 0.7644002437591553,
|
| 353 |
+
"step": 588
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 0.372150721042022,
|
| 357 |
+
"grad_norm": 8.31484603881836,
|
| 358 |
+
"learning_rate": 1.947187141216992e-05,
|
| 359 |
+
"loss": 0.6794478893280029,
|
| 360 |
+
"step": 600
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 0.37959373546286246,
|
| 364 |
+
"grad_norm": 6.703721523284912,
|
| 365 |
+
"learning_rate": 1.941676234213548e-05,
|
| 366 |
+
"loss": 0.6266262531280518,
|
| 367 |
+
"step": 612
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 0.3870367498837029,
|
| 371 |
+
"grad_norm": 9.479427337646484,
|
| 372 |
+
"learning_rate": 1.9361653272101036e-05,
|
| 373 |
+
"loss": 0.6851427555084229,
|
| 374 |
+
"step": 624
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.39447976430454335,
|
| 378 |
+
"grad_norm": 7.663156032562256,
|
| 379 |
+
"learning_rate": 1.9306544202066593e-05,
|
| 380 |
+
"loss": 0.6938677628835043,
|
| 381 |
+
"step": 636
|
| 382 |
+
},
|
| 383 |
+
{
|
| 384 |
+
"epoch": 0.4019227787253838,
|
| 385 |
+
"grad_norm": 4.276080131530762,
|
| 386 |
+
"learning_rate": 1.9251435132032147e-05,
|
| 387 |
+
"loss": 0.76728622118632,
|
| 388 |
+
"step": 648
|
| 389 |
+
},
|
| 390 |
+
{
|
| 391 |
+
"epoch": 0.40936579314622423,
|
| 392 |
+
"grad_norm": 11.622859001159668,
|
| 393 |
+
"learning_rate": 1.9196326061997705e-05,
|
| 394 |
+
"loss": 0.7580918471018473,
|
| 395 |
+
"step": 660
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 0.4168088075670647,
|
| 399 |
+
"grad_norm": 13.203335762023926,
|
| 400 |
+
"learning_rate": 1.9141216991963262e-05,
|
| 401 |
+
"loss": 0.642679770787557,
|
| 402 |
+
"step": 672
|
| 403 |
+
},
|
| 404 |
+
{
|
| 405 |
+
"epoch": 0.4242518219879051,
|
| 406 |
+
"grad_norm": 8.963321685791016,
|
| 407 |
+
"learning_rate": 1.908610792192882e-05,
|
| 408 |
+
"loss": 0.6361099084218343,
|
| 409 |
+
"step": 684
|
| 410 |
+
},
|
| 411 |
+
{
|
| 412 |
+
"epoch": 0.43169483640874556,
|
| 413 |
+
"grad_norm": 8.1705904006958,
|
| 414 |
+
"learning_rate": 1.9030998851894377e-05,
|
| 415 |
+
"loss": 0.6898341178894043,
|
| 416 |
+
"step": 696
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.439137850829586,
|
| 420 |
+
"grad_norm": 3.9877262115478516,
|
| 421 |
+
"learning_rate": 1.8975889781859934e-05,
|
| 422 |
+
"loss": 0.6462088028589884,
|
| 423 |
+
"step": 708
|
| 424 |
+
},
|
| 425 |
+
{
|
| 426 |
+
"epoch": 0.4465808652504264,
|
| 427 |
+
"grad_norm": 12.81478500366211,
|
| 428 |
+
"learning_rate": 1.892078071182549e-05,
|
| 429 |
+
"loss": 0.6965091228485107,
|
| 430 |
+
"step": 720
|
| 431 |
+
},
|
| 432 |
+
{
|
| 433 |
+
"epoch": 0.45402387967126684,
|
| 434 |
+
"grad_norm": 7.810659885406494,
|
| 435 |
+
"learning_rate": 1.8865671641791045e-05,
|
| 436 |
+
"loss": 0.7788422902425131,
|
| 437 |
+
"step": 732
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 0.4614668940921073,
|
| 441 |
+
"grad_norm": 4.958326816558838,
|
| 442 |
+
"learning_rate": 1.8810562571756603e-05,
|
| 443 |
+
"loss": 0.7460188865661621,
|
| 444 |
+
"step": 744
|
| 445 |
+
},
|
| 446 |
+
{
|
| 447 |
+
"epoch": 0.4689099085129477,
|
| 448 |
+
"grad_norm": 9.091962814331055,
|
| 449 |
+
"learning_rate": 1.875545350172216e-05,
|
| 450 |
+
"loss": 0.6937299569447836,
|
| 451 |
+
"step": 756
|
| 452 |
+
},
|
| 453 |
+
{
|
| 454 |
+
"epoch": 0.47635292293378817,
|
| 455 |
+
"grad_norm": 7.729589939117432,
|
| 456 |
+
"learning_rate": 1.8700344431687717e-05,
|
| 457 |
+
"loss": 0.6188247601191202,
|
| 458 |
+
"step": 768
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.4837959373546286,
|
| 462 |
+
"grad_norm": 8.878933906555176,
|
| 463 |
+
"learning_rate": 1.8645235361653275e-05,
|
| 464 |
+
"loss": 0.7017858028411865,
|
| 465 |
+
"step": 780
|
| 466 |
+
},
|
| 467 |
+
{
|
| 468 |
+
"epoch": 0.49123895177546906,
|
| 469 |
+
"grad_norm": 23.914348602294922,
|
| 470 |
+
"learning_rate": 1.8590126291618832e-05,
|
| 471 |
+
"loss": 0.7923436164855957,
|
| 472 |
+
"step": 792
|
| 473 |
+
},
|
| 474 |
+
{
|
| 475 |
+
"epoch": 0.4986819661963095,
|
| 476 |
+
"grad_norm": 10.980387687683105,
|
| 477 |
+
"learning_rate": 1.853501722158439e-05,
|
| 478 |
+
"loss": 0.6881453990936279,
|
| 479 |
+
"step": 804
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 0.5061249806171499,
|
| 483 |
+
"grad_norm": 6.988458156585693,
|
| 484 |
+
"learning_rate": 1.8479908151549943e-05,
|
| 485 |
+
"loss": 0.683276891708374,
|
| 486 |
+
"step": 816
|
| 487 |
+
},
|
| 488 |
+
{
|
| 489 |
+
"epoch": 0.5135679950379903,
|
| 490 |
+
"grad_norm": 23.667926788330078,
|
| 491 |
+
"learning_rate": 1.84247990815155e-05,
|
| 492 |
+
"loss": 0.6124229431152344,
|
| 493 |
+
"step": 828
|
| 494 |
+
},
|
| 495 |
+
{
|
| 496 |
+
"epoch": 0.5210110094588308,
|
| 497 |
+
"grad_norm": 7.078935623168945,
|
| 498 |
+
"learning_rate": 1.8369690011481058e-05,
|
| 499 |
+
"loss": 0.7043429215749105,
|
| 500 |
+
"step": 840
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 0.5284540238796712,
|
| 504 |
+
"grad_norm": 9.82224178314209,
|
| 505 |
+
"learning_rate": 1.8314580941446615e-05,
|
| 506 |
+
"loss": 0.6555114189783732,
|
| 507 |
+
"step": 852
|
| 508 |
+
},
|
| 509 |
+
{
|
| 510 |
+
"epoch": 0.5358970383005117,
|
| 511 |
+
"grad_norm": 8.077360153198242,
|
| 512 |
+
"learning_rate": 1.8259471871412173e-05,
|
| 513 |
+
"loss": 0.6555444002151489,
|
| 514 |
+
"step": 864
|
| 515 |
+
},
|
| 516 |
+
{
|
| 517 |
+
"epoch": 0.5433400527213521,
|
| 518 |
+
"grad_norm": 3.6762046813964844,
|
| 519 |
+
"learning_rate": 1.820436280137773e-05,
|
| 520 |
+
"loss": 0.636172374089559,
|
| 521 |
+
"step": 876
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 0.5507830671421926,
|
| 525 |
+
"grad_norm": 3.8388607501983643,
|
| 526 |
+
"learning_rate": 1.8149253731343287e-05,
|
| 527 |
+
"loss": 0.6085333824157715,
|
| 528 |
+
"step": 888
|
| 529 |
+
},
|
| 530 |
+
{
|
| 531 |
+
"epoch": 0.558226081563033,
|
| 532 |
+
"grad_norm": 3.0353925228118896,
|
| 533 |
+
"learning_rate": 1.809414466130884e-05,
|
| 534 |
+
"loss": 0.58968718846639,
|
| 535 |
+
"step": 900
|
| 536 |
+
},
|
| 537 |
+
{
|
| 538 |
+
"epoch": 0.5656690959838735,
|
| 539 |
+
"grad_norm": 6.465055465698242,
|
| 540 |
+
"learning_rate": 1.80390355912744e-05,
|
| 541 |
+
"loss": 0.6078658103942871,
|
| 542 |
+
"step": 912
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 0.5731121104047139,
|
| 546 |
+
"grad_norm": 5.472475528717041,
|
| 547 |
+
"learning_rate": 1.7983926521239956e-05,
|
| 548 |
+
"loss": 0.6997927029927572,
|
| 549 |
+
"step": 924
|
| 550 |
+
},
|
| 551 |
+
{
|
| 552 |
+
"epoch": 0.5805551248255544,
|
| 553 |
+
"grad_norm": 15.40697193145752,
|
| 554 |
+
"learning_rate": 1.792881745120551e-05,
|
| 555 |
+
"loss": 0.6386371453603109,
|
| 556 |
+
"step": 936
|
| 557 |
+
},
|
| 558 |
+
{
|
| 559 |
+
"epoch": 0.5879981392463948,
|
| 560 |
+
"grad_norm": 6.439900875091553,
|
| 561 |
+
"learning_rate": 1.787370838117107e-05,
|
| 562 |
+
"loss": 0.6876135667165121,
|
| 563 |
+
"step": 948
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 0.5954411536672353,
|
| 567 |
+
"grad_norm": 10.793220520019531,
|
| 568 |
+
"learning_rate": 1.7818599311136628e-05,
|
| 569 |
+
"loss": 0.6237523953119913,
|
| 570 |
+
"step": 960
|
| 571 |
+
},
|
| 572 |
+
{
|
| 573 |
+
"epoch": 0.6028841680880757,
|
| 574 |
+
"grad_norm": 5.377976417541504,
|
| 575 |
+
"learning_rate": 1.7763490241102185e-05,
|
| 576 |
+
"loss": 0.614266554514567,
|
| 577 |
+
"step": 972
|
| 578 |
+
},
|
| 579 |
+
{
|
| 580 |
+
"epoch": 0.6103271825089162,
|
| 581 |
+
"grad_norm": 7.794371604919434,
|
| 582 |
+
"learning_rate": 1.770838117106774e-05,
|
| 583 |
+
"loss": 0.5918615261713663,
|
| 584 |
+
"step": 984
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 0.6177701969297565,
|
| 588 |
+
"grad_norm": 4.7419867515563965,
|
| 589 |
+
"learning_rate": 1.7653272101033296e-05,
|
| 590 |
+
"loss": 0.5848552385965983,
|
| 591 |
+
"step": 996
|
| 592 |
+
},
|
| 593 |
+
{
|
| 594 |
+
"epoch": 0.625213211350597,
|
| 595 |
+
"grad_norm": 14.705470085144043,
|
| 596 |
+
"learning_rate": 1.7598163030998854e-05,
|
| 597 |
+
"loss": 0.6608580350875854,
|
| 598 |
+
"step": 1008
|
| 599 |
+
},
|
| 600 |
+
{
|
| 601 |
+
"epoch": 0.6326562257714374,
|
| 602 |
+
"grad_norm": 6.041922092437744,
|
| 603 |
+
"learning_rate": 1.754305396096441e-05,
|
| 604 |
+
"loss": 0.549665609995524,
|
| 605 |
+
"step": 1020
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 0.6400992401922778,
|
| 609 |
+
"grad_norm": 5.13696813583374,
|
| 610 |
+
"learning_rate": 1.7487944890929965e-05,
|
| 611 |
+
"loss": 0.7017458279927572,
|
| 612 |
+
"step": 1032
|
| 613 |
+
},
|
| 614 |
+
{
|
| 615 |
+
"epoch": 0.6475422546131183,
|
| 616 |
+
"grad_norm": 6.016454696655273,
|
| 617 |
+
"learning_rate": 1.7432835820895522e-05,
|
| 618 |
+
"loss": 0.6309004227320353,
|
| 619 |
+
"step": 1044
|
| 620 |
+
},
|
| 621 |
+
{
|
| 622 |
+
"epoch": 0.6549852690339587,
|
| 623 |
+
"grad_norm": 9.331708908081055,
|
| 624 |
+
"learning_rate": 1.7377726750861083e-05,
|
| 625 |
+
"loss": 0.6831174691518148,
|
| 626 |
+
"step": 1056
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 0.6624282834547992,
|
| 630 |
+
"grad_norm": 9.878951072692871,
|
| 631 |
+
"learning_rate": 1.7322617680826637e-05,
|
| 632 |
+
"loss": 0.6587471961975098,
|
| 633 |
+
"step": 1068
|
| 634 |
+
},
|
| 635 |
+
{
|
| 636 |
+
"epoch": 0.6698712978756396,
|
| 637 |
+
"grad_norm": 5.033365726470947,
|
| 638 |
+
"learning_rate": 1.7267508610792194e-05,
|
| 639 |
+
"loss": 0.6370361646016439,
|
| 640 |
+
"step": 1080
|
| 641 |
+
},
|
| 642 |
+
{
|
| 643 |
+
"epoch": 0.6773143122964801,
|
| 644 |
+
"grad_norm": 18.762298583984375,
|
| 645 |
+
"learning_rate": 1.721239954075775e-05,
|
| 646 |
+
"loss": 0.5823976198832194,
|
| 647 |
+
"step": 1092
|
| 648 |
+
},
|
| 649 |
+
{
|
| 650 |
+
"epoch": 0.6847573267173205,
|
| 651 |
+
"grad_norm": 2.940394163131714,
|
| 652 |
+
"learning_rate": 1.715729047072331e-05,
|
| 653 |
+
"loss": 0.6264007488886515,
|
| 654 |
+
"step": 1104
|
| 655 |
+
},
|
| 656 |
+
{
|
| 657 |
+
"epoch": 0.692200341138161,
|
| 658 |
+
"grad_norm": 7.621018886566162,
|
| 659 |
+
"learning_rate": 1.7102181400688863e-05,
|
| 660 |
+
"loss": 0.5824793974558512,
|
| 661 |
+
"step": 1116
|
| 662 |
+
},
|
| 663 |
+
{
|
| 664 |
+
"epoch": 0.6996433555590014,
|
| 665 |
+
"grad_norm": 3.141854763031006,
|
| 666 |
+
"learning_rate": 1.704707233065442e-05,
|
| 667 |
+
"loss": 0.5842764774958292,
|
| 668 |
+
"step": 1128
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 0.7070863699798419,
|
| 672 |
+
"grad_norm": 5.849940776824951,
|
| 673 |
+
"learning_rate": 1.6991963260619978e-05,
|
| 674 |
+
"loss": 0.5304047664006551,
|
| 675 |
+
"step": 1140
|
| 676 |
+
},
|
| 677 |
+
{
|
| 678 |
+
"epoch": 0.7145293844006823,
|
| 679 |
+
"grad_norm": 7.9883551597595215,
|
| 680 |
+
"learning_rate": 1.6936854190585535e-05,
|
| 681 |
+
"loss": 0.5599017937978109,
|
| 682 |
+
"step": 1152
|
| 683 |
+
},
|
| 684 |
+
{
|
| 685 |
+
"epoch": 0.7219723988215228,
|
| 686 |
+
"grad_norm": 11.370931625366211,
|
| 687 |
+
"learning_rate": 1.6881745120551092e-05,
|
| 688 |
+
"loss": 0.5798830588658651,
|
| 689 |
+
"step": 1164
|
| 690 |
+
},
|
| 691 |
+
{
|
| 692 |
+
"epoch": 0.7294154132423631,
|
| 693 |
+
"grad_norm": 3.5065290927886963,
|
| 694 |
+
"learning_rate": 1.682663605051665e-05,
|
| 695 |
+
"loss": 0.6167506376902262,
|
| 696 |
+
"step": 1176
|
| 697 |
+
},
|
| 698 |
+
{
|
| 699 |
+
"epoch": 0.7368584276632036,
|
| 700 |
+
"grad_norm": 5.930673599243164,
|
| 701 |
+
"learning_rate": 1.6771526980482207e-05,
|
| 702 |
+
"loss": 0.5873833497365316,
|
| 703 |
+
"step": 1188
|
| 704 |
+
},
|
| 705 |
+
{
|
| 706 |
+
"epoch": 0.744301442084044,
|
| 707 |
+
"grad_norm": 6.102614402770996,
|
| 708 |
+
"learning_rate": 1.671641791044776e-05,
|
| 709 |
+
"loss": 0.6477183898289999,
|
| 710 |
+
"step": 1200
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 0.7517444565048845,
|
| 714 |
+
"grad_norm": 4.337888717651367,
|
| 715 |
+
"learning_rate": 1.6661308840413318e-05,
|
| 716 |
+
"loss": 0.5860347350438436,
|
| 717 |
+
"step": 1212
|
| 718 |
+
},
|
| 719 |
+
{
|
| 720 |
+
"epoch": 0.7591874709257249,
|
| 721 |
+
"grad_norm": 4.841605186462402,
|
| 722 |
+
"learning_rate": 1.6606199770378875e-05,
|
| 723 |
+
"loss": 0.6613442897796631,
|
| 724 |
+
"step": 1224
|
| 725 |
+
},
|
| 726 |
+
{
|
| 727 |
+
"epoch": 0.7666304853465653,
|
| 728 |
+
"grad_norm": 14.614047050476074,
|
| 729 |
+
"learning_rate": 1.6551090700344433e-05,
|
| 730 |
+
"loss": 0.6218246618906657,
|
| 731 |
+
"step": 1236
|
| 732 |
+
},
|
| 733 |
+
{
|
| 734 |
+
"epoch": 0.7740734997674058,
|
| 735 |
+
"grad_norm": 8.036581039428711,
|
| 736 |
+
"learning_rate": 1.649598163030999e-05,
|
| 737 |
+
"loss": 0.5646830002466837,
|
| 738 |
+
"step": 1248
|
| 739 |
+
},
|
| 740 |
+
{
|
| 741 |
+
"epoch": 0.7815165141882462,
|
| 742 |
+
"grad_norm": 3.943291664123535,
|
| 743 |
+
"learning_rate": 1.6440872560275547e-05,
|
| 744 |
+
"loss": 0.6018180449803671,
|
| 745 |
+
"step": 1260
|
| 746 |
+
},
|
| 747 |
+
{
|
| 748 |
+
"epoch": 0.7889595286090867,
|
| 749 |
+
"grad_norm": 12.51102352142334,
|
| 750 |
+
"learning_rate": 1.6385763490241105e-05,
|
| 751 |
+
"loss": 0.6140671968460083,
|
| 752 |
+
"step": 1272
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 0.7964025430299271,
|
| 756 |
+
"grad_norm": 3.718653678894043,
|
| 757 |
+
"learning_rate": 1.633065442020666e-05,
|
| 758 |
+
"loss": 0.5359119176864624,
|
| 759 |
+
"step": 1284
|
| 760 |
+
},
|
| 761 |
+
{
|
| 762 |
+
"epoch": 0.8038455574507676,
|
| 763 |
+
"grad_norm": 2.8353357315063477,
|
| 764 |
+
"learning_rate": 1.6275545350172216e-05,
|
| 765 |
+
"loss": 0.502113143603007,
|
| 766 |
+
"step": 1296
|
| 767 |
+
},
|
| 768 |
+
{
|
| 769 |
+
"epoch": 0.811288571871608,
|
| 770 |
+
"grad_norm": 4.345269203186035,
|
| 771 |
+
"learning_rate": 1.6220436280137773e-05,
|
| 772 |
+
"loss": 0.5975545644760132,
|
| 773 |
+
"step": 1308
|
| 774 |
+
},
|
| 775 |
+
{
|
| 776 |
+
"epoch": 0.8187315862924485,
|
| 777 |
+
"grad_norm": 6.92914342880249,
|
| 778 |
+
"learning_rate": 1.616532721010333e-05,
|
| 779 |
+
"loss": 0.6587652762730917,
|
| 780 |
+
"step": 1320
|
| 781 |
+
},
|
| 782 |
+
{
|
| 783 |
+
"epoch": 0.8261746007132889,
|
| 784 |
+
"grad_norm": 4.188693046569824,
|
| 785 |
+
"learning_rate": 1.6110218140068888e-05,
|
| 786 |
+
"loss": 0.6142017841339111,
|
| 787 |
+
"step": 1332
|
| 788 |
+
},
|
| 789 |
+
{
|
| 790 |
+
"epoch": 0.8336176151341294,
|
| 791 |
+
"grad_norm": 9.596400260925293,
|
| 792 |
+
"learning_rate": 1.6055109070034445e-05,
|
| 793 |
+
"loss": 0.5469466845194498,
|
| 794 |
+
"step": 1344
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 0.8410606295549697,
|
| 798 |
+
"grad_norm": 4.810947895050049,
|
| 799 |
+
"learning_rate": 1.6000000000000003e-05,
|
| 800 |
+
"loss": 0.5744484265645345,
|
| 801 |
+
"step": 1356
|
| 802 |
+
},
|
| 803 |
+
{
|
| 804 |
+
"epoch": 0.8485036439758102,
|
| 805 |
+
"grad_norm": 3.5819036960601807,
|
| 806 |
+
"learning_rate": 1.5944890929965557e-05,
|
| 807 |
+
"loss": 0.5045839150746664,
|
| 808 |
+
"step": 1368
|
| 809 |
+
},
|
| 810 |
+
{
|
| 811 |
+
"epoch": 0.8559466583966506,
|
| 812 |
+
"grad_norm": 4.879307746887207,
|
| 813 |
+
"learning_rate": 1.5889781859931114e-05,
|
| 814 |
+
"loss": 0.6669184366861979,
|
| 815 |
+
"step": 1380
|
| 816 |
+
},
|
| 817 |
+
{
|
| 818 |
+
"epoch": 0.8633896728174911,
|
| 819 |
+
"grad_norm": 6.7210693359375,
|
| 820 |
+
"learning_rate": 1.583467278989667e-05,
|
| 821 |
+
"loss": 0.5324758291244507,
|
| 822 |
+
"step": 1392
|
| 823 |
+
},
|
| 824 |
+
{
|
| 825 |
+
"epoch": 0.8708326872383315,
|
| 826 |
+
"grad_norm": 4.653786659240723,
|
| 827 |
+
"learning_rate": 1.577956371986223e-05,
|
| 828 |
+
"loss": 0.543891986211141,
|
| 829 |
+
"step": 1404
|
| 830 |
+
},
|
| 831 |
+
{
|
| 832 |
+
"epoch": 0.878275701659172,
|
| 833 |
+
"grad_norm": 6.386638641357422,
|
| 834 |
+
"learning_rate": 1.5724454649827786e-05,
|
| 835 |
+
"loss": 0.5688877105712891,
|
| 836 |
+
"step": 1416
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 0.8857187160800124,
|
| 840 |
+
"grad_norm": 14.5455322265625,
|
| 841 |
+
"learning_rate": 1.5669345579793343e-05,
|
| 842 |
+
"loss": 0.5081936915715536,
|
| 843 |
+
"step": 1428
|
| 844 |
+
},
|
| 845 |
+
{
|
| 846 |
+
"epoch": 0.8931617305008528,
|
| 847 |
+
"grad_norm": 13.621335983276367,
|
| 848 |
+
"learning_rate": 1.56142365097589e-05,
|
| 849 |
+
"loss": 0.5466565688451132,
|
| 850 |
+
"step": 1440
|
| 851 |
+
},
|
| 852 |
+
{
|
| 853 |
+
"epoch": 0.9006047449216933,
|
| 854 |
+
"grad_norm": 7.791660308837891,
|
| 855 |
+
"learning_rate": 1.5559127439724455e-05,
|
| 856 |
+
"loss": 0.5543188651402792,
|
| 857 |
+
"step": 1452
|
| 858 |
+
},
|
| 859 |
+
{
|
| 860 |
+
"epoch": 0.9080477593425337,
|
| 861 |
+
"grad_norm": 4.039332866668701,
|
| 862 |
+
"learning_rate": 1.5504018369690012e-05,
|
| 863 |
+
"loss": 0.564227819442749,
|
| 864 |
+
"step": 1464
|
| 865 |
+
},
|
| 866 |
+
{
|
| 867 |
+
"epoch": 0.9154907737633742,
|
| 868 |
+
"grad_norm": 5.744030475616455,
|
| 869 |
+
"learning_rate": 1.544890929965557e-05,
|
| 870 |
+
"loss": 0.5645032723744711,
|
| 871 |
+
"step": 1476
|
| 872 |
+
},
|
| 873 |
+
{
|
| 874 |
+
"epoch": 0.9229337881842146,
|
| 875 |
+
"grad_norm": 7.17244815826416,
|
| 876 |
+
"learning_rate": 1.5393800229621126e-05,
|
| 877 |
+
"loss": 0.6025459369023641,
|
| 878 |
+
"step": 1488
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 0.9303768026050551,
|
| 882 |
+
"grad_norm": 9.460329055786133,
|
| 883 |
+
"learning_rate": 1.5338691159586684e-05,
|
| 884 |
+
"loss": 0.5522710482279459,
|
| 885 |
+
"step": 1500
|
| 886 |
+
},
|
| 887 |
+
{
|
| 888 |
+
"epoch": 0.9378198170258955,
|
| 889 |
+
"grad_norm": 8.257369995117188,
|
| 890 |
+
"learning_rate": 1.528358208955224e-05,
|
| 891 |
+
"loss": 0.5696142514546713,
|
| 892 |
+
"step": 1512
|
| 893 |
+
},
|
| 894 |
+
{
|
| 895 |
+
"epoch": 0.945262831446736,
|
| 896 |
+
"grad_norm": 14.734770774841309,
|
| 897 |
+
"learning_rate": 1.5228473019517798e-05,
|
| 898 |
+
"loss": 0.60454261302948,
|
| 899 |
+
"step": 1524
|
| 900 |
+
},
|
| 901 |
+
{
|
| 902 |
+
"epoch": 0.9527058458675763,
|
| 903 |
+
"grad_norm": 4.352370738983154,
|
| 904 |
+
"learning_rate": 1.5173363949483352e-05,
|
| 905 |
+
"loss": 0.48172632853190106,
|
| 906 |
+
"step": 1536
|
| 907 |
+
},
|
| 908 |
+
{
|
| 909 |
+
"epoch": 0.9601488602884168,
|
| 910 |
+
"grad_norm": 2.388683557510376,
|
| 911 |
+
"learning_rate": 1.511825487944891e-05,
|
| 912 |
+
"loss": 0.5889216661453247,
|
| 913 |
+
"step": 1548
|
| 914 |
+
},
|
| 915 |
+
{
|
| 916 |
+
"epoch": 0.9675918747092572,
|
| 917 |
+
"grad_norm": 9.910285949707031,
|
| 918 |
+
"learning_rate": 1.5063145809414467e-05,
|
| 919 |
+
"loss": 0.5621689558029175,
|
| 920 |
+
"step": 1560
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 0.9750348891300977,
|
| 924 |
+
"grad_norm": 5.445796966552734,
|
| 925 |
+
"learning_rate": 1.5008036739380026e-05,
|
| 926 |
+
"loss": 0.5526663859685262,
|
| 927 |
+
"step": 1572
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 0.9824779035509381,
|
| 931 |
+
"grad_norm": 5.242825984954834,
|
| 932 |
+
"learning_rate": 1.495292766934558e-05,
|
| 933 |
+
"loss": 0.5673882563908895,
|
| 934 |
+
"step": 1584
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 0.9899209179717786,
|
| 938 |
+
"grad_norm": 10.1865234375,
|
| 939 |
+
"learning_rate": 1.4897818599311137e-05,
|
| 940 |
+
"loss": 0.5648102362950643,
|
| 941 |
+
"step": 1596
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 0.997363932392619,
|
| 945 |
+
"grad_norm": 5.388990879058838,
|
| 946 |
+
"learning_rate": 1.4842709529276695e-05,
|
| 947 |
+
"loss": 0.5376612345377604,
|
| 948 |
+
"step": 1608
|
| 949 |
+
},
|
| 950 |
+
{
|
| 951 |
+
"epoch": 1.0,
|
| 952 |
+
"eval_f1": 0.4322638779516363,
|
| 953 |
+
"eval_loss": 0.13781657814979553,
|
| 954 |
+
"eval_precision": 0.3967545697112817,
|
| 955 |
+
"eval_recall": 0.4884485429972486,
|
| 956 |
+
"eval_runtime": 583.7374,
|
| 957 |
+
"eval_samples_per_second": 66.278,
|
| 958 |
+
"eval_steps_per_second": 1.382,
|
| 959 |
+
"step": 1613
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"epoch": 1.004341758412157,
|
| 963 |
+
"grad_norm": 11.303878784179688,
|
| 964 |
+
"learning_rate": 1.478760045924225e-05,
|
| 965 |
+
"loss": 0.46324888865152997,
|
| 966 |
+
"step": 1620
|
| 967 |
+
},
|
| 968 |
+
{
|
| 969 |
+
"epoch": 1.0117847728329974,
|
| 970 |
+
"grad_norm": 4.389431476593018,
|
| 971 |
+
"learning_rate": 1.4732491389207808e-05,
|
| 972 |
+
"loss": 0.48095786571502686,
|
| 973 |
+
"step": 1632
|
| 974 |
+
},
|
| 975 |
+
{
|
| 976 |
+
"epoch": 1.0192277872538378,
|
| 977 |
+
"grad_norm": 2.298799514770508,
|
| 978 |
+
"learning_rate": 1.4677382319173365e-05,
|
| 979 |
+
"loss": 0.5406383275985718,
|
| 980 |
+
"step": 1644
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 1.0266708016746782,
|
| 984 |
+
"grad_norm": 4.433741092681885,
|
| 985 |
+
"learning_rate": 1.4622273249138922e-05,
|
| 986 |
+
"loss": 0.4697510798772176,
|
| 987 |
+
"step": 1656
|
| 988 |
+
},
|
| 989 |
+
{
|
| 990 |
+
"epoch": 1.0341138160955188,
|
| 991 |
+
"grad_norm": 4.704965591430664,
|
| 992 |
+
"learning_rate": 1.4567164179104478e-05,
|
| 993 |
+
"loss": 0.5180115699768066,
|
| 994 |
+
"step": 1668
|
| 995 |
+
},
|
| 996 |
+
{
|
| 997 |
+
"epoch": 1.0415568305163592,
|
| 998 |
+
"grad_norm": 5.16159725189209,
|
| 999 |
+
"learning_rate": 1.4512055109070035e-05,
|
| 1000 |
+
"loss": 0.49386584758758545,
|
| 1001 |
+
"step": 1680
|
| 1002 |
+
},
|
| 1003 |
+
{
|
| 1004 |
+
"epoch": 1.0489998449371996,
|
| 1005 |
+
"grad_norm": 2.5488502979278564,
|
| 1006 |
+
"learning_rate": 1.4456946039035593e-05,
|
| 1007 |
+
"loss": 0.41516109307607013,
|
| 1008 |
+
"step": 1692
|
| 1009 |
+
},
|
| 1010 |
+
{
|
| 1011 |
+
"epoch": 1.05644285935804,
|
| 1012 |
+
"grad_norm": 12.81408405303955,
|
| 1013 |
+
"learning_rate": 1.4401836969001148e-05,
|
| 1014 |
+
"loss": 0.5269262790679932,
|
| 1015 |
+
"step": 1704
|
| 1016 |
+
},
|
| 1017 |
+
{
|
| 1018 |
+
"epoch": 1.0638858737788803,
|
| 1019 |
+
"grad_norm": 2.8521316051483154,
|
| 1020 |
+
"learning_rate": 1.4346727898966706e-05,
|
| 1021 |
+
"loss": 0.45834481716156006,
|
| 1022 |
+
"step": 1716
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 1.071328888199721,
|
| 1026 |
+
"grad_norm": 5.517307758331299,
|
| 1027 |
+
"learning_rate": 1.4291618828932263e-05,
|
| 1028 |
+
"loss": 0.573523203531901,
|
| 1029 |
+
"step": 1728
|
| 1030 |
+
},
|
| 1031 |
+
{
|
| 1032 |
+
"epoch": 1.0787719026205613,
|
| 1033 |
+
"grad_norm": 2.4321818351745605,
|
| 1034 |
+
"learning_rate": 1.423650975889782e-05,
|
| 1035 |
+
"loss": 0.4625085194905599,
|
| 1036 |
+
"step": 1740
|
| 1037 |
+
},
|
| 1038 |
+
{
|
| 1039 |
+
"epoch": 1.0862149170414017,
|
| 1040 |
+
"grad_norm": 15.532980918884277,
|
| 1041 |
+
"learning_rate": 1.4181400688863376e-05,
|
| 1042 |
+
"loss": 0.5057009855906168,
|
| 1043 |
+
"step": 1752
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"epoch": 1.093657931462242,
|
| 1047 |
+
"grad_norm": 4.501278877258301,
|
| 1048 |
+
"learning_rate": 1.4126291618828933e-05,
|
| 1049 |
+
"loss": 0.4823911984761556,
|
| 1050 |
+
"step": 1764
|
| 1051 |
+
},
|
| 1052 |
+
{
|
| 1053 |
+
"epoch": 1.1011009458830827,
|
| 1054 |
+
"grad_norm": 6.726215362548828,
|
| 1055 |
+
"learning_rate": 1.407118254879449e-05,
|
| 1056 |
+
"loss": 0.42187273502349854,
|
| 1057 |
+
"step": 1776
|
| 1058 |
+
},
|
| 1059 |
+
{
|
| 1060 |
+
"epoch": 1.108543960303923,
|
| 1061 |
+
"grad_norm": 14.170055389404297,
|
| 1062 |
+
"learning_rate": 1.4016073478760046e-05,
|
| 1063 |
+
"loss": 0.5301618576049805,
|
| 1064 |
+
"step": 1788
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 1.1159869747247635,
|
| 1068 |
+
"grad_norm": 2.776092767715454,
|
| 1069 |
+
"learning_rate": 1.3960964408725603e-05,
|
| 1070 |
+
"loss": 0.4935903151830037,
|
| 1071 |
+
"step": 1800
|
| 1072 |
+
},
|
| 1073 |
+
{
|
| 1074 |
+
"epoch": 1.1234299891456039,
|
| 1075 |
+
"grad_norm": 7.334898948669434,
|
| 1076 |
+
"learning_rate": 1.390585533869116e-05,
|
| 1077 |
+
"loss": 0.5331637859344482,
|
| 1078 |
+
"step": 1812
|
| 1079 |
+
},
|
| 1080 |
+
{
|
| 1081 |
+
"epoch": 1.1308730035664445,
|
| 1082 |
+
"grad_norm": 4.995052337646484,
|
| 1083 |
+
"learning_rate": 1.3850746268656718e-05,
|
| 1084 |
+
"loss": 0.4663925568262736,
|
| 1085 |
+
"step": 1824
|
| 1086 |
+
},
|
| 1087 |
+
{
|
| 1088 |
+
"epoch": 1.1383160179872849,
|
| 1089 |
+
"grad_norm": 9.281367301940918,
|
| 1090 |
+
"learning_rate": 1.3795637198622274e-05,
|
| 1091 |
+
"loss": 0.44923396905263263,
|
| 1092 |
+
"step": 1836
|
| 1093 |
+
},
|
| 1094 |
+
{
|
| 1095 |
+
"epoch": 1.1457590324081253,
|
| 1096 |
+
"grad_norm": 5.095090866088867,
|
| 1097 |
+
"learning_rate": 1.3740528128587831e-05,
|
| 1098 |
+
"loss": 0.5650514364242554,
|
| 1099 |
+
"step": 1848
|
| 1100 |
+
},
|
| 1101 |
+
{
|
| 1102 |
+
"epoch": 1.1532020468289657,
|
| 1103 |
+
"grad_norm": 2.299600839614868,
|
| 1104 |
+
"learning_rate": 1.3685419058553388e-05,
|
| 1105 |
+
"loss": 0.48252185185750324,
|
| 1106 |
+
"step": 1860
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 1.1606450612498063,
|
| 1110 |
+
"grad_norm": 6.702273368835449,
|
| 1111 |
+
"learning_rate": 1.3630309988518944e-05,
|
| 1112 |
+
"loss": 0.5192966063817342,
|
| 1113 |
+
"step": 1872
|
| 1114 |
+
},
|
| 1115 |
+
{
|
| 1116 |
+
"epoch": 1.1680880756706467,
|
| 1117 |
+
"grad_norm": 10.89989948272705,
|
| 1118 |
+
"learning_rate": 1.3575200918484501e-05,
|
| 1119 |
+
"loss": 0.48262282212575275,
|
| 1120 |
+
"step": 1884
|
| 1121 |
+
},
|
| 1122 |
+
{
|
| 1123 |
+
"epoch": 1.175531090091487,
|
| 1124 |
+
"grad_norm": 15.075289726257324,
|
| 1125 |
+
"learning_rate": 1.3520091848450059e-05,
|
| 1126 |
+
"loss": 0.45538806915283203,
|
| 1127 |
+
"step": 1896
|
| 1128 |
+
},
|
| 1129 |
+
{
|
| 1130 |
+
"epoch": 1.1829741045123274,
|
| 1131 |
+
"grad_norm": 3.0880722999572754,
|
| 1132 |
+
"learning_rate": 1.3464982778415616e-05,
|
| 1133 |
+
"loss": 0.46872226397196454,
|
| 1134 |
+
"step": 1908
|
| 1135 |
+
},
|
| 1136 |
+
{
|
| 1137 |
+
"epoch": 1.1904171189331678,
|
| 1138 |
+
"grad_norm": 8.533724784851074,
|
| 1139 |
+
"learning_rate": 1.3409873708381172e-05,
|
| 1140 |
+
"loss": 0.4827297528584798,
|
| 1141 |
+
"step": 1920
|
| 1142 |
+
},
|
| 1143 |
+
{
|
| 1144 |
+
"epoch": 1.1978601333540084,
|
| 1145 |
+
"grad_norm": 3.070657968521118,
|
| 1146 |
+
"learning_rate": 1.3354764638346729e-05,
|
| 1147 |
+
"loss": 0.48583118120829266,
|
| 1148 |
+
"step": 1932
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 1.2053031477748488,
|
| 1152 |
+
"grad_norm": 3.7270054817199707,
|
| 1153 |
+
"learning_rate": 1.3299655568312286e-05,
|
| 1154 |
+
"loss": 0.505421002705892,
|
| 1155 |
+
"step": 1944
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 1.2127461621956892,
|
| 1159 |
+
"grad_norm": 9.997303009033203,
|
| 1160 |
+
"learning_rate": 1.3244546498277842e-05,
|
| 1161 |
+
"loss": 0.4140005111694336,
|
| 1162 |
+
"step": 1956
|
| 1163 |
+
},
|
| 1164 |
+
{
|
| 1165 |
+
"epoch": 1.2201891766165298,
|
| 1166 |
+
"grad_norm": 11.578160285949707,
|
| 1167 |
+
"learning_rate": 1.31894374282434e-05,
|
| 1168 |
+
"loss": 0.44274091720581055,
|
| 1169 |
+
"step": 1968
|
| 1170 |
+
},
|
| 1171 |
+
{
|
| 1172 |
+
"epoch": 1.2276321910373702,
|
| 1173 |
+
"grad_norm": 9.199183464050293,
|
| 1174 |
+
"learning_rate": 1.3134328358208957e-05,
|
| 1175 |
+
"loss": 0.5600036780039469,
|
| 1176 |
+
"step": 1980
|
| 1177 |
+
},
|
| 1178 |
+
{
|
| 1179 |
+
"epoch": 1.2350752054582106,
|
| 1180 |
+
"grad_norm": 7.212144374847412,
|
| 1181 |
+
"learning_rate": 1.3079219288174514e-05,
|
| 1182 |
+
"loss": 0.494090994199117,
|
| 1183 |
+
"step": 1992
|
| 1184 |
+
},
|
| 1185 |
+
{
|
| 1186 |
+
"epoch": 1.242518219879051,
|
| 1187 |
+
"grad_norm": 3.4123635292053223,
|
| 1188 |
+
"learning_rate": 1.302411021814007e-05,
|
| 1189 |
+
"loss": 0.4909547170003255,
|
| 1190 |
+
"step": 2004
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 1.2499612342998914,
|
| 1194 |
+
"grad_norm": 7.941708087921143,
|
| 1195 |
+
"learning_rate": 1.2969001148105627e-05,
|
| 1196 |
+
"loss": 0.47832663853963214,
|
| 1197 |
+
"step": 2016
|
| 1198 |
+
},
|
| 1199 |
+
{
|
| 1200 |
+
"epoch": 1.257404248720732,
|
| 1201 |
+
"grad_norm": 2.4799387454986572,
|
| 1202 |
+
"learning_rate": 1.2913892078071184e-05,
|
| 1203 |
+
"loss": 0.49106045564015705,
|
| 1204 |
+
"step": 2028
|
| 1205 |
+
},
|
| 1206 |
+
{
|
| 1207 |
+
"epoch": 1.2648472631415724,
|
| 1208 |
+
"grad_norm": 5.136545658111572,
|
| 1209 |
+
"learning_rate": 1.2858783008036742e-05,
|
| 1210 |
+
"loss": 0.4738738536834717,
|
| 1211 |
+
"step": 2040
|
| 1212 |
+
},
|
| 1213 |
+
{
|
| 1214 |
+
"epoch": 1.2722902775624128,
|
| 1215 |
+
"grad_norm": 4.9489240646362305,
|
| 1216 |
+
"learning_rate": 1.2803673938002297e-05,
|
| 1217 |
+
"loss": 0.4953068097432454,
|
| 1218 |
+
"step": 2052
|
| 1219 |
+
},
|
| 1220 |
+
{
|
| 1221 |
+
"epoch": 1.2797332919832531,
|
| 1222 |
+
"grad_norm": 6.822914123535156,
|
| 1223 |
+
"learning_rate": 1.2748564867967855e-05,
|
| 1224 |
+
"loss": 0.46026841799418133,
|
| 1225 |
+
"step": 2064
|
| 1226 |
+
},
|
| 1227 |
+
{
|
| 1228 |
+
"epoch": 1.2871763064040938,
|
| 1229 |
+
"grad_norm": 6.177013874053955,
|
| 1230 |
+
"learning_rate": 1.2693455797933412e-05,
|
| 1231 |
+
"loss": 0.494237224260966,
|
| 1232 |
+
"step": 2076
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 1.2946193208249341,
|
| 1236 |
+
"grad_norm": 2.4243626594543457,
|
| 1237 |
+
"learning_rate": 1.2638346727898967e-05,
|
| 1238 |
+
"loss": 0.5003351370493571,
|
| 1239 |
+
"step": 2088
|
| 1240 |
+
},
|
| 1241 |
+
{
|
| 1242 |
+
"epoch": 1.3020623352457745,
|
| 1243 |
+
"grad_norm": 18.99603843688965,
|
| 1244 |
+
"learning_rate": 1.2583237657864525e-05,
|
| 1245 |
+
"loss": 0.5109163920084635,
|
| 1246 |
+
"step": 2100
|
| 1247 |
+
},
|
| 1248 |
+
{
|
| 1249 |
+
"epoch": 1.309505349666615,
|
| 1250 |
+
"grad_norm": 2.4371707439422607,
|
| 1251 |
+
"learning_rate": 1.2528128587830082e-05,
|
| 1252 |
+
"loss": 0.41310568650563556,
|
| 1253 |
+
"step": 2112
|
| 1254 |
+
},
|
| 1255 |
+
{
|
| 1256 |
+
"epoch": 1.3169483640874553,
|
| 1257 |
+
"grad_norm": 3.7665302753448486,
|
| 1258 |
+
"learning_rate": 1.247301951779564e-05,
|
| 1259 |
+
"loss": 0.45848862330118817,
|
| 1260 |
+
"step": 2124
|
| 1261 |
+
},
|
| 1262 |
+
{
|
| 1263 |
+
"epoch": 1.324391378508296,
|
| 1264 |
+
"grad_norm": 12.537642478942871,
|
| 1265 |
+
"learning_rate": 1.2417910447761195e-05,
|
| 1266 |
+
"loss": 0.523716410001119,
|
| 1267 |
+
"step": 2136
|
| 1268 |
+
},
|
| 1269 |
+
{
|
| 1270 |
+
"epoch": 1.3318343929291363,
|
| 1271 |
+
"grad_norm": 2.882084846496582,
|
| 1272 |
+
"learning_rate": 1.2362801377726752e-05,
|
| 1273 |
+
"loss": 0.47608526547749835,
|
| 1274 |
+
"step": 2148
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 1.3392774073499767,
|
| 1278 |
+
"grad_norm": 2.86336612701416,
|
| 1279 |
+
"learning_rate": 1.230769230769231e-05,
|
| 1280 |
+
"loss": 0.4347230593363444,
|
| 1281 |
+
"step": 2160
|
| 1282 |
+
},
|
| 1283 |
+
{
|
| 1284 |
+
"epoch": 1.3467204217708173,
|
| 1285 |
+
"grad_norm": 3.1628830432891846,
|
| 1286 |
+
"learning_rate": 1.2252583237657865e-05,
|
| 1287 |
+
"loss": 0.46674203872680664,
|
| 1288 |
+
"step": 2172
|
| 1289 |
+
},
|
| 1290 |
+
{
|
| 1291 |
+
"epoch": 1.3541634361916577,
|
| 1292 |
+
"grad_norm": 11.767653465270996,
|
| 1293 |
+
"learning_rate": 1.2197474167623423e-05,
|
| 1294 |
+
"loss": 0.47306569417317706,
|
| 1295 |
+
"step": 2184
|
| 1296 |
+
},
|
| 1297 |
+
{
|
| 1298 |
+
"epoch": 1.361606450612498,
|
| 1299 |
+
"grad_norm": 11.81271743774414,
|
| 1300 |
+
"learning_rate": 1.214236509758898e-05,
|
| 1301 |
+
"loss": 0.4672517776489258,
|
| 1302 |
+
"step": 2196
|
| 1303 |
+
},
|
| 1304 |
+
{
|
| 1305 |
+
"epoch": 1.3690494650333385,
|
| 1306 |
+
"grad_norm": 3.6157212257385254,
|
| 1307 |
+
"learning_rate": 1.2087256027554537e-05,
|
| 1308 |
+
"loss": 0.4465065797170003,
|
| 1309 |
+
"step": 2208
|
| 1310 |
+
},
|
| 1311 |
+
{
|
| 1312 |
+
"epoch": 1.3764924794541789,
|
| 1313 |
+
"grad_norm": 3.778449773788452,
|
| 1314 |
+
"learning_rate": 1.2032146957520093e-05,
|
| 1315 |
+
"loss": 0.5149937868118286,
|
| 1316 |
+
"step": 2220
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 1.3839354938750195,
|
| 1320 |
+
"grad_norm": 2.5120906829833984,
|
| 1321 |
+
"learning_rate": 1.197703788748565e-05,
|
| 1322 |
+
"loss": 0.45879046122233075,
|
| 1323 |
+
"step": 2232
|
| 1324 |
+
},
|
| 1325 |
+
{
|
| 1326 |
+
"epoch": 1.3913785082958599,
|
| 1327 |
+
"grad_norm": 17.704999923706055,
|
| 1328 |
+
"learning_rate": 1.1921928817451208e-05,
|
| 1329 |
+
"loss": 0.5167669057846069,
|
| 1330 |
+
"step": 2244
|
| 1331 |
+
},
|
| 1332 |
+
{
|
| 1333 |
+
"epoch": 1.3988215227167002,
|
| 1334 |
+
"grad_norm": 11.8012113571167,
|
| 1335 |
+
"learning_rate": 1.1866819747416762e-05,
|
| 1336 |
+
"loss": 0.496524175008138,
|
| 1337 |
+
"step": 2256
|
| 1338 |
+
},
|
| 1339 |
+
{
|
| 1340 |
+
"epoch": 1.4062645371375406,
|
| 1341 |
+
"grad_norm": 13.236916542053223,
|
| 1342 |
+
"learning_rate": 1.181171067738232e-05,
|
| 1343 |
+
"loss": 0.47164463996887207,
|
| 1344 |
+
"step": 2268
|
| 1345 |
+
},
|
| 1346 |
+
{
|
| 1347 |
+
"epoch": 1.4137075515583812,
|
| 1348 |
+
"grad_norm": 3.6107146739959717,
|
| 1349 |
+
"learning_rate": 1.1756601607347878e-05,
|
| 1350 |
+
"loss": 0.4411802689234416,
|
| 1351 |
+
"step": 2280
|
| 1352 |
+
},
|
| 1353 |
+
{
|
| 1354 |
+
"epoch": 1.4211505659792216,
|
| 1355 |
+
"grad_norm": 3.5400538444519043,
|
| 1356 |
+
"learning_rate": 1.1701492537313435e-05,
|
| 1357 |
+
"loss": 0.44078512986501056,
|
| 1358 |
+
"step": 2292
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 1.428593580400062,
|
| 1362 |
+
"grad_norm": 3.386744260787964,
|
| 1363 |
+
"learning_rate": 1.164638346727899e-05,
|
| 1364 |
+
"loss": 0.44522058963775635,
|
| 1365 |
+
"step": 2304
|
| 1366 |
+
},
|
| 1367 |
+
{
|
| 1368 |
+
"epoch": 1.4360365948209024,
|
| 1369 |
+
"grad_norm": 7.451818466186523,
|
| 1370 |
+
"learning_rate": 1.1591274397244548e-05,
|
| 1371 |
+
"loss": 0.4643220106760661,
|
| 1372 |
+
"step": 2316
|
| 1373 |
+
},
|
| 1374 |
+
{
|
| 1375 |
+
"epoch": 1.4434796092417428,
|
| 1376 |
+
"grad_norm": 3.741562843322754,
|
| 1377 |
+
"learning_rate": 1.1536165327210106e-05,
|
| 1378 |
+
"loss": 0.4557652473449707,
|
| 1379 |
+
"step": 2328
|
| 1380 |
+
},
|
| 1381 |
+
{
|
| 1382 |
+
"epoch": 1.4509226236625834,
|
| 1383 |
+
"grad_norm": 2.767171621322632,
|
| 1384 |
+
"learning_rate": 1.148105625717566e-05,
|
| 1385 |
+
"loss": 0.4677225748697917,
|
| 1386 |
+
"step": 2340
|
| 1387 |
+
},
|
| 1388 |
+
{
|
| 1389 |
+
"epoch": 1.4583656380834238,
|
| 1390 |
+
"grad_norm": 5.696690559387207,
|
| 1391 |
+
"learning_rate": 1.1425947187141217e-05,
|
| 1392 |
+
"loss": 0.42428747812906903,
|
| 1393 |
+
"step": 2352
|
| 1394 |
+
},
|
| 1395 |
+
{
|
| 1396 |
+
"epoch": 1.4658086525042642,
|
| 1397 |
+
"grad_norm": 6.44115686416626,
|
| 1398 |
+
"learning_rate": 1.1370838117106774e-05,
|
| 1399 |
+
"loss": 0.4969560702641805,
|
| 1400 |
+
"step": 2364
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 1.4732516669251048,
|
| 1404 |
+
"grad_norm": 6.7684831619262695,
|
| 1405 |
+
"learning_rate": 1.1315729047072333e-05,
|
| 1406 |
+
"loss": 0.5301390091578165,
|
| 1407 |
+
"step": 2376
|
| 1408 |
+
},
|
| 1409 |
+
{
|
| 1410 |
+
"epoch": 1.4806946813459452,
|
| 1411 |
+
"grad_norm": 2.761455774307251,
|
| 1412 |
+
"learning_rate": 1.1260619977037887e-05,
|
| 1413 |
+
"loss": 0.4755421082178752,
|
| 1414 |
+
"step": 2388
|
| 1415 |
+
},
|
| 1416 |
+
{
|
| 1417 |
+
"epoch": 1.4881376957667856,
|
| 1418 |
+
"grad_norm": 7.615389347076416,
|
| 1419 |
+
"learning_rate": 1.1205510907003444e-05,
|
| 1420 |
+
"loss": 0.4676011800765991,
|
| 1421 |
+
"step": 2400
|
| 1422 |
+
},
|
| 1423 |
+
{
|
| 1424 |
+
"epoch": 1.495580710187626,
|
| 1425 |
+
"grad_norm": 3.118619680404663,
|
| 1426 |
+
"learning_rate": 1.1150401836969002e-05,
|
| 1427 |
+
"loss": 0.4575995206832886,
|
| 1428 |
+
"step": 2412
|
| 1429 |
+
},
|
| 1430 |
+
{
|
| 1431 |
+
"epoch": 1.5030237246084663,
|
| 1432 |
+
"grad_norm": 4.179815769195557,
|
| 1433 |
+
"learning_rate": 1.1095292766934557e-05,
|
| 1434 |
+
"loss": 0.5326940615971884,
|
| 1435 |
+
"step": 2424
|
| 1436 |
+
},
|
| 1437 |
+
{
|
| 1438 |
+
"epoch": 1.5104667390293067,
|
| 1439 |
+
"grad_norm": 3.128330945968628,
|
| 1440 |
+
"learning_rate": 1.1040183696900115e-05,
|
| 1441 |
+
"loss": 0.45927361647288006,
|
| 1442 |
+
"step": 2436
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 1.5179097534501473,
|
| 1446 |
+
"grad_norm": 3.6722943782806396,
|
| 1447 |
+
"learning_rate": 1.0985074626865672e-05,
|
| 1448 |
+
"loss": 0.5232657591501871,
|
| 1449 |
+
"step": 2448
|
| 1450 |
+
},
|
| 1451 |
+
{
|
| 1452 |
+
"epoch": 1.5253527678709877,
|
| 1453 |
+
"grad_norm": 8.696102142333984,
|
| 1454 |
+
"learning_rate": 1.092996555683123e-05,
|
| 1455 |
+
"loss": 0.5253320535024008,
|
| 1456 |
+
"step": 2460
|
| 1457 |
+
},
|
| 1458 |
+
{
|
| 1459 |
+
"epoch": 1.5327957822918283,
|
| 1460 |
+
"grad_norm": 6.030095100402832,
|
| 1461 |
+
"learning_rate": 1.0874856486796785e-05,
|
| 1462 |
+
"loss": 0.4725768566131592,
|
| 1463 |
+
"step": 2472
|
| 1464 |
+
},
|
| 1465 |
+
{
|
| 1466 |
+
"epoch": 1.5402387967126687,
|
| 1467 |
+
"grad_norm": 8.892803192138672,
|
| 1468 |
+
"learning_rate": 1.0819747416762342e-05,
|
| 1469 |
+
"loss": 0.44700531164805096,
|
| 1470 |
+
"step": 2484
|
| 1471 |
+
},
|
| 1472 |
+
{
|
| 1473 |
+
"epoch": 1.5476818111335091,
|
| 1474 |
+
"grad_norm": 15.271442413330078,
|
| 1475 |
+
"learning_rate": 1.07646383467279e-05,
|
| 1476 |
+
"loss": 0.44845902919769287,
|
| 1477 |
+
"step": 2496
|
| 1478 |
+
},
|
| 1479 |
+
{
|
| 1480 |
+
"epoch": 1.5551248255543495,
|
| 1481 |
+
"grad_norm": 5.234111785888672,
|
| 1482 |
+
"learning_rate": 1.0709529276693457e-05,
|
| 1483 |
+
"loss": 0.5186563730239868,
|
| 1484 |
+
"step": 2508
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 1.5625678399751899,
|
| 1488 |
+
"grad_norm": 6.541170597076416,
|
| 1489 |
+
"learning_rate": 1.0654420206659013e-05,
|
| 1490 |
+
"loss": 0.4690740505854289,
|
| 1491 |
+
"step": 2520
|
| 1492 |
+
},
|
| 1493 |
+
{
|
| 1494 |
+
"epoch": 1.5700108543960303,
|
| 1495 |
+
"grad_norm": 2.7548892498016357,
|
| 1496 |
+
"learning_rate": 1.059931113662457e-05,
|
| 1497 |
+
"loss": 0.43329620361328125,
|
| 1498 |
+
"step": 2532
|
| 1499 |
+
},
|
| 1500 |
+
{
|
| 1501 |
+
"epoch": 1.5774538688168709,
|
| 1502 |
+
"grad_norm": 12.428861618041992,
|
| 1503 |
+
"learning_rate": 1.0544202066590127e-05,
|
| 1504 |
+
"loss": 0.43588805198669434,
|
| 1505 |
+
"step": 2544
|
| 1506 |
+
},
|
| 1507 |
+
{
|
| 1508 |
+
"epoch": 1.5848968832377113,
|
| 1509 |
+
"grad_norm": 9.76059627532959,
|
| 1510 |
+
"learning_rate": 1.0489092996555683e-05,
|
| 1511 |
+
"loss": 0.4283796151479085,
|
| 1512 |
+
"step": 2556
|
| 1513 |
+
},
|
| 1514 |
+
{
|
| 1515 |
+
"epoch": 1.5923398976585517,
|
| 1516 |
+
"grad_norm": 10.960260391235352,
|
| 1517 |
+
"learning_rate": 1.043398392652124e-05,
|
| 1518 |
+
"loss": 0.4565364519755046,
|
| 1519 |
+
"step": 2568
|
| 1520 |
+
},
|
| 1521 |
+
{
|
| 1522 |
+
"epoch": 1.5997829120793923,
|
| 1523 |
+
"grad_norm": 6.568747043609619,
|
| 1524 |
+
"learning_rate": 1.0378874856486798e-05,
|
| 1525 |
+
"loss": 0.41670429706573486,
|
| 1526 |
+
"step": 2580
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 1.6072259265002327,
|
| 1530 |
+
"grad_norm": 2.755124568939209,
|
| 1531 |
+
"learning_rate": 1.0323765786452355e-05,
|
| 1532 |
+
"loss": 0.4691346486409505,
|
| 1533 |
+
"step": 2592
|
| 1534 |
+
},
|
| 1535 |
+
{
|
| 1536 |
+
"epoch": 1.614668940921073,
|
| 1537 |
+
"grad_norm": 21.070772171020508,
|
| 1538 |
+
"learning_rate": 1.026865671641791e-05,
|
| 1539 |
+
"loss": 0.4186259905497233,
|
| 1540 |
+
"step": 2604
|
| 1541 |
+
},
|
| 1542 |
+
{
|
| 1543 |
+
"epoch": 1.6221119553419134,
|
| 1544 |
+
"grad_norm": 5.272284507751465,
|
| 1545 |
+
"learning_rate": 1.0213547646383468e-05,
|
| 1546 |
+
"loss": 0.4942372639973958,
|
| 1547 |
+
"step": 2616
|
| 1548 |
+
},
|
| 1549 |
+
{
|
| 1550 |
+
"epoch": 1.6295549697627538,
|
| 1551 |
+
"grad_norm": 8.858941078186035,
|
| 1552 |
+
"learning_rate": 1.0158438576349025e-05,
|
| 1553 |
+
"loss": 0.4842514594395955,
|
| 1554 |
+
"step": 2628
|
| 1555 |
+
},
|
| 1556 |
+
{
|
| 1557 |
+
"epoch": 1.6369979841835942,
|
| 1558 |
+
"grad_norm": 4.663693428039551,
|
| 1559 |
+
"learning_rate": 1.010332950631458e-05,
|
| 1560 |
+
"loss": 0.49429325262705487,
|
| 1561 |
+
"step": 2640
|
| 1562 |
+
},
|
| 1563 |
+
{
|
| 1564 |
+
"epoch": 1.6444409986044348,
|
| 1565 |
+
"grad_norm": 14.864917755126953,
|
| 1566 |
+
"learning_rate": 1.0048220436280138e-05,
|
| 1567 |
+
"loss": 0.46838700771331787,
|
| 1568 |
+
"step": 2652
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 1.6518840130252752,
|
| 1572 |
+
"grad_norm": 2.5411393642425537,
|
| 1573 |
+
"learning_rate": 9.993111366245695e-06,
|
| 1574 |
+
"loss": 0.4521595239639282,
|
| 1575 |
+
"step": 2664
|
| 1576 |
+
},
|
| 1577 |
+
{
|
| 1578 |
+
"epoch": 1.6593270274461158,
|
| 1579 |
+
"grad_norm": 3.005941152572632,
|
| 1580 |
+
"learning_rate": 9.938002296211253e-06,
|
| 1581 |
+
"loss": 0.48365652561187744,
|
| 1582 |
+
"step": 2676
|
| 1583 |
+
},
|
| 1584 |
+
{
|
| 1585 |
+
"epoch": 1.6667700418669562,
|
| 1586 |
+
"grad_norm": 5.7398552894592285,
|
| 1587 |
+
"learning_rate": 9.88289322617681e-06,
|
| 1588 |
+
"loss": 0.4695123831431071,
|
| 1589 |
+
"step": 2688
|
| 1590 |
+
},
|
| 1591 |
+
{
|
| 1592 |
+
"epoch": 1.6742130562877966,
|
| 1593 |
+
"grad_norm": 4.946065902709961,
|
| 1594 |
+
"learning_rate": 9.827784156142366e-06,
|
| 1595 |
+
"loss": 0.4761979579925537,
|
| 1596 |
+
"step": 2700
|
| 1597 |
+
},
|
| 1598 |
+
{
|
| 1599 |
+
"epoch": 1.681656070708637,
|
| 1600 |
+
"grad_norm": 7.703652858734131,
|
| 1601 |
+
"learning_rate": 9.772675086107923e-06,
|
| 1602 |
+
"loss": 0.49780480066935223,
|
| 1603 |
+
"step": 2712
|
| 1604 |
+
},
|
| 1605 |
+
{
|
| 1606 |
+
"epoch": 1.6890990851294774,
|
| 1607 |
+
"grad_norm": 8.237687110900879,
|
| 1608 |
+
"learning_rate": 9.71756601607348e-06,
|
| 1609 |
+
"loss": 0.4623022476832072,
|
| 1610 |
+
"step": 2724
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 1.6965420995503178,
|
| 1614 |
+
"grad_norm": 2.87007474899292,
|
| 1615 |
+
"learning_rate": 9.662456946039036e-06,
|
| 1616 |
+
"loss": 0.41221630573272705,
|
| 1617 |
+
"step": 2736
|
| 1618 |
+
},
|
| 1619 |
+
{
|
| 1620 |
+
"epoch": 1.7039851139711584,
|
| 1621 |
+
"grad_norm": 4.247465133666992,
|
| 1622 |
+
"learning_rate": 9.607347876004593e-06,
|
| 1623 |
+
"loss": 0.4721166690190633,
|
| 1624 |
+
"step": 2748
|
| 1625 |
+
},
|
| 1626 |
+
{
|
| 1627 |
+
"epoch": 1.7114281283919988,
|
| 1628 |
+
"grad_norm": 4.022077560424805,
|
| 1629 |
+
"learning_rate": 9.552238805970149e-06,
|
| 1630 |
+
"loss": 0.47880788644154865,
|
| 1631 |
+
"step": 2760
|
| 1632 |
+
},
|
| 1633 |
+
{
|
| 1634 |
+
"epoch": 1.7188711428128391,
|
| 1635 |
+
"grad_norm": 5.686273574829102,
|
| 1636 |
+
"learning_rate": 9.497129735935708e-06,
|
| 1637 |
+
"loss": 0.486567219098409,
|
| 1638 |
+
"step": 2772
|
| 1639 |
+
},
|
| 1640 |
+
{
|
| 1641 |
+
"epoch": 1.7263141572336798,
|
| 1642 |
+
"grad_norm": 4.733608245849609,
|
| 1643 |
+
"learning_rate": 9.442020665901264e-06,
|
| 1644 |
+
"loss": 0.4696682294209798,
|
| 1645 |
+
"step": 2784
|
| 1646 |
+
},
|
| 1647 |
+
{
|
| 1648 |
+
"epoch": 1.7337571716545201,
|
| 1649 |
+
"grad_norm": 3.8102357387542725,
|
| 1650 |
+
"learning_rate": 9.38691159586682e-06,
|
| 1651 |
+
"loss": 0.4944278796513875,
|
| 1652 |
+
"step": 2796
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 1.7412001860753605,
|
| 1656 |
+
"grad_norm": 5.343743801116943,
|
| 1657 |
+
"learning_rate": 9.331802525832377e-06,
|
| 1658 |
+
"loss": 0.45073699951171875,
|
| 1659 |
+
"step": 2808
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 1.748643200496201,
|
| 1663 |
+
"grad_norm": 8.939608573913574,
|
| 1664 |
+
"learning_rate": 9.276693455797934e-06,
|
| 1665 |
+
"loss": 0.5150019327799479,
|
| 1666 |
+
"step": 2820
|
| 1667 |
+
},
|
| 1668 |
+
{
|
| 1669 |
+
"epoch": 1.7560862149170413,
|
| 1670 |
+
"grad_norm": 9.984607696533203,
|
| 1671 |
+
"learning_rate": 9.221584385763491e-06,
|
| 1672 |
+
"loss": 0.49051181475321454,
|
| 1673 |
+
"step": 2832
|
| 1674 |
+
},
|
| 1675 |
+
{
|
| 1676 |
+
"epoch": 1.7635292293378817,
|
| 1677 |
+
"grad_norm": 4.297845840454102,
|
| 1678 |
+
"learning_rate": 9.166475315729047e-06,
|
| 1679 |
+
"loss": 0.43834813435872394,
|
| 1680 |
+
"step": 2844
|
| 1681 |
+
},
|
| 1682 |
+
{
|
| 1683 |
+
"epoch": 1.7709722437587223,
|
| 1684 |
+
"grad_norm": 4.738193035125732,
|
| 1685 |
+
"learning_rate": 9.111366245694604e-06,
|
| 1686 |
+
"loss": 0.48496174812316895,
|
| 1687 |
+
"step": 2856
|
| 1688 |
+
},
|
| 1689 |
+
{
|
| 1690 |
+
"epoch": 1.7784152581795627,
|
| 1691 |
+
"grad_norm": 6.950840473175049,
|
| 1692 |
+
"learning_rate": 9.056257175660162e-06,
|
| 1693 |
+
"loss": 0.4803895950317383,
|
| 1694 |
+
"step": 2868
|
| 1695 |
+
},
|
| 1696 |
+
{
|
| 1697 |
+
"epoch": 1.7858582726004033,
|
| 1698 |
+
"grad_norm": 2.9567737579345703,
|
| 1699 |
+
"learning_rate": 9.001148105625719e-06,
|
| 1700 |
+
"loss": 0.47137478987375897,
|
| 1701 |
+
"step": 2880
|
| 1702 |
+
},
|
| 1703 |
+
{
|
| 1704 |
+
"epoch": 1.7933012870212437,
|
| 1705 |
+
"grad_norm": 21.629295349121094,
|
| 1706 |
+
"learning_rate": 8.946039035591275e-06,
|
| 1707 |
+
"loss": 0.5382961829503378,
|
| 1708 |
+
"step": 2892
|
| 1709 |
+
},
|
| 1710 |
+
{
|
| 1711 |
+
"epoch": 1.800744301442084,
|
| 1712 |
+
"grad_norm": 4.054839611053467,
|
| 1713 |
+
"learning_rate": 8.890929965556832e-06,
|
| 1714 |
+
"loss": 0.429937203725179,
|
| 1715 |
+
"step": 2904
|
| 1716 |
+
},
|
| 1717 |
+
{
|
| 1718 |
+
"epoch": 1.8081873158629245,
|
| 1719 |
+
"grad_norm": 8.124676704406738,
|
| 1720 |
+
"learning_rate": 8.83582089552239e-06,
|
| 1721 |
+
"loss": 0.46343564987182617,
|
| 1722 |
+
"step": 2916
|
| 1723 |
+
},
|
| 1724 |
+
{
|
| 1725 |
+
"epoch": 1.8156303302837649,
|
| 1726 |
+
"grad_norm": 6.405475616455078,
|
| 1727 |
+
"learning_rate": 8.780711825487945e-06,
|
| 1728 |
+
"loss": 0.47476502259572345,
|
| 1729 |
+
"step": 2928
|
| 1730 |
+
},
|
| 1731 |
+
{
|
| 1732 |
+
"epoch": 1.8230733447046052,
|
| 1733 |
+
"grad_norm": 3.4982993602752686,
|
| 1734 |
+
"learning_rate": 8.725602755453502e-06,
|
| 1735 |
+
"loss": 0.42661325136820477,
|
| 1736 |
+
"step": 2940
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"epoch": 1.8305163591254459,
|
| 1740 |
+
"grad_norm": 5.036385536193848,
|
| 1741 |
+
"learning_rate": 8.67049368541906e-06,
|
| 1742 |
+
"loss": 0.42475831508636475,
|
| 1743 |
+
"step": 2952
|
| 1744 |
+
},
|
| 1745 |
+
{
|
| 1746 |
+
"epoch": 1.8379593735462862,
|
| 1747 |
+
"grad_norm": 9.453807830810547,
|
| 1748 |
+
"learning_rate": 8.615384615384617e-06,
|
| 1749 |
+
"loss": 0.4522843360900879,
|
| 1750 |
+
"step": 2964
|
| 1751 |
+
},
|
| 1752 |
+
{
|
| 1753 |
+
"epoch": 1.8454023879671266,
|
| 1754 |
+
"grad_norm": 7.572172164916992,
|
| 1755 |
+
"learning_rate": 8.560275545350172e-06,
|
| 1756 |
+
"loss": 0.5405757427215576,
|
| 1757 |
+
"step": 2976
|
| 1758 |
+
},
|
| 1759 |
+
{
|
| 1760 |
+
"epoch": 1.8528454023879672,
|
| 1761 |
+
"grad_norm": 3.8509397506713867,
|
| 1762 |
+
"learning_rate": 8.50516647531573e-06,
|
| 1763 |
+
"loss": 0.4206368128458659,
|
| 1764 |
+
"step": 2988
|
| 1765 |
+
},
|
| 1766 |
+
{
|
| 1767 |
+
"epoch": 1.8602884168088076,
|
| 1768 |
+
"grad_norm": 3.8660781383514404,
|
| 1769 |
+
"learning_rate": 8.450057405281287e-06,
|
| 1770 |
+
"loss": 0.4278140465418498,
|
| 1771 |
+
"step": 3000
|
| 1772 |
+
},
|
| 1773 |
+
{
|
| 1774 |
+
"epoch": 1.867731431229648,
|
| 1775 |
+
"grad_norm": 13.179638862609863,
|
| 1776 |
+
"learning_rate": 8.394948335246843e-06,
|
| 1777 |
+
"loss": 0.45146167278289795,
|
| 1778 |
+
"step": 3012
|
| 1779 |
+
},
|
| 1780 |
+
{
|
| 1781 |
+
"epoch": 1.8751744456504884,
|
| 1782 |
+
"grad_norm": 2.5003507137298584,
|
| 1783 |
+
"learning_rate": 8.3398392652124e-06,
|
| 1784 |
+
"loss": 0.5010615189870199,
|
| 1785 |
+
"step": 3024
|
| 1786 |
+
},
|
| 1787 |
+
{
|
| 1788 |
+
"epoch": 1.8826174600713288,
|
| 1789 |
+
"grad_norm": 6.336158752441406,
|
| 1790 |
+
"learning_rate": 8.284730195177957e-06,
|
| 1791 |
+
"loss": 0.48331379890441895,
|
| 1792 |
+
"step": 3036
|
| 1793 |
+
},
|
| 1794 |
+
{
|
| 1795 |
+
"epoch": 1.8900604744921692,
|
| 1796 |
+
"grad_norm": 3.9048869609832764,
|
| 1797 |
+
"learning_rate": 8.229621125143515e-06,
|
| 1798 |
+
"loss": 0.4964629014333089,
|
| 1799 |
+
"step": 3048
|
| 1800 |
+
},
|
| 1801 |
+
{
|
| 1802 |
+
"epoch": 1.8975034889130098,
|
| 1803 |
+
"grad_norm": 4.851749897003174,
|
| 1804 |
+
"learning_rate": 8.17451205510907e-06,
|
| 1805 |
+
"loss": 0.4605306386947632,
|
| 1806 |
+
"step": 3060
|
| 1807 |
+
},
|
| 1808 |
+
{
|
| 1809 |
+
"epoch": 1.9049465033338502,
|
| 1810 |
+
"grad_norm": 2.5984604358673096,
|
| 1811 |
+
"learning_rate": 8.119402985074628e-06,
|
| 1812 |
+
"loss": 0.42377761999766034,
|
| 1813 |
+
"step": 3072
|
| 1814 |
+
},
|
| 1815 |
+
{
|
| 1816 |
+
"epoch": 1.9123895177546908,
|
| 1817 |
+
"grad_norm": 14.330255508422852,
|
| 1818 |
+
"learning_rate": 8.064293915040185e-06,
|
| 1819 |
+
"loss": 0.4586070378621419,
|
| 1820 |
+
"step": 3084
|
| 1821 |
+
},
|
| 1822 |
+
{
|
| 1823 |
+
"epoch": 1.9198325321755312,
|
| 1824 |
+
"grad_norm": 5.363494873046875,
|
| 1825 |
+
"learning_rate": 8.00918484500574e-06,
|
| 1826 |
+
"loss": 0.4935295581817627,
|
| 1827 |
+
"step": 3096
|
| 1828 |
+
},
|
| 1829 |
+
{
|
| 1830 |
+
"epoch": 1.9272755465963716,
|
| 1831 |
+
"grad_norm": 5.703904151916504,
|
| 1832 |
+
"learning_rate": 7.954075774971298e-06,
|
| 1833 |
+
"loss": 0.44021427631378174,
|
| 1834 |
+
"step": 3108
|
| 1835 |
+
},
|
| 1836 |
+
{
|
| 1837 |
+
"epoch": 1.934718561017212,
|
| 1838 |
+
"grad_norm": 5.600277423858643,
|
| 1839 |
+
"learning_rate": 7.898966704936855e-06,
|
| 1840 |
+
"loss": 0.48560158411661786,
|
| 1841 |
+
"step": 3120
|
| 1842 |
+
},
|
| 1843 |
+
{
|
| 1844 |
+
"epoch": 1.9421615754380523,
|
| 1845 |
+
"grad_norm": 11.074832916259766,
|
| 1846 |
+
"learning_rate": 7.843857634902413e-06,
|
| 1847 |
+
"loss": 0.4312416712443034,
|
| 1848 |
+
"step": 3132
|
| 1849 |
+
},
|
| 1850 |
+
{
|
| 1851 |
+
"epoch": 1.9496045898588927,
|
| 1852 |
+
"grad_norm": 3.4356892108917236,
|
| 1853 |
+
"learning_rate": 7.788748564867968e-06,
|
| 1854 |
+
"loss": 0.4442025025685628,
|
| 1855 |
+
"step": 3144
|
| 1856 |
+
},
|
| 1857 |
+
{
|
| 1858 |
+
"epoch": 1.9570476042797333,
|
| 1859 |
+
"grad_norm": 3.7474091053009033,
|
| 1860 |
+
"learning_rate": 7.733639494833526e-06,
|
| 1861 |
+
"loss": 0.5241368214289347,
|
| 1862 |
+
"step": 3156
|
| 1863 |
+
},
|
| 1864 |
+
{
|
| 1865 |
+
"epoch": 1.9644906187005737,
|
| 1866 |
+
"grad_norm": 4.750489234924316,
|
| 1867 |
+
"learning_rate": 7.678530424799083e-06,
|
| 1868 |
+
"loss": 0.4401020606358846,
|
| 1869 |
+
"step": 3168
|
| 1870 |
+
},
|
| 1871 |
+
{
|
| 1872 |
+
"epoch": 1.9719336331214141,
|
| 1873 |
+
"grad_norm": 22.131851196289062,
|
| 1874 |
+
"learning_rate": 7.6234213547646386e-06,
|
| 1875 |
+
"loss": 0.5134913126627604,
|
| 1876 |
+
"step": 3180
|
| 1877 |
+
},
|
| 1878 |
+
{
|
| 1879 |
+
"epoch": 1.9793766475422547,
|
| 1880 |
+
"grad_norm": 4.812230587005615,
|
| 1881 |
+
"learning_rate": 7.568312284730196e-06,
|
| 1882 |
+
"loss": 0.5479523340861002,
|
| 1883 |
+
"step": 3192
|
| 1884 |
+
},
|
| 1885 |
+
{
|
| 1886 |
+
"epoch": 1.9868196619630951,
|
| 1887 |
+
"grad_norm": 6.560222625732422,
|
| 1888 |
+
"learning_rate": 7.513203214695752e-06,
|
| 1889 |
+
"loss": 0.4738404353459676,
|
| 1890 |
+
"step": 3204
|
| 1891 |
+
},
|
| 1892 |
+
{
|
| 1893 |
+
"epoch": 1.9942626763839355,
|
| 1894 |
+
"grad_norm": 5.240246772766113,
|
| 1895 |
+
"learning_rate": 7.45809414466131e-06,
|
| 1896 |
+
"loss": 0.4475013017654419,
|
| 1897 |
+
"step": 3216
|
| 1898 |
+
},
|
| 1899 |
+
{
|
| 1900 |
+
"epoch": 2.0,
|
| 1901 |
+
"eval_f1": 0.43079906968624254,
|
| 1902 |
+
"eval_loss": 0.11952196806669235,
|
| 1903 |
+
"eval_precision": 0.391528709389682,
|
| 1904 |
+
"eval_recall": 0.4931553870446119,
|
| 1905 |
+
"eval_runtime": 585.0453,
|
| 1906 |
+
"eval_samples_per_second": 66.13,
|
| 1907 |
+
"eval_steps_per_second": 1.379,
|
| 1908 |
+
"step": 3226
|
| 1909 |
+
},
|
| 1910 |
+
{
|
| 1911 |
+
"epoch": 2.0012405024034736,
|
| 1912 |
+
"grad_norm": 4.430677890777588,
|
| 1913 |
+
"learning_rate": 7.402985074626866e-06,
|
| 1914 |
+
"loss": 0.4009953737258911,
|
| 1915 |
+
"step": 3228
|
| 1916 |
+
},
|
| 1917 |
+
{
|
| 1918 |
+
"epoch": 2.008683516824314,
|
| 1919 |
+
"grad_norm": 10.324471473693848,
|
| 1920 |
+
"learning_rate": 7.3478760045924235e-06,
|
| 1921 |
+
"loss": 0.4711928367614746,
|
| 1922 |
+
"step": 3240
|
| 1923 |
+
},
|
| 1924 |
+
{
|
| 1925 |
+
"epoch": 2.0161265312451544,
|
| 1926 |
+
"grad_norm": 11.249197006225586,
|
| 1927 |
+
"learning_rate": 7.29276693455798e-06,
|
| 1928 |
+
"loss": 0.4341440995534261,
|
| 1929 |
+
"step": 3252
|
| 1930 |
+
},
|
| 1931 |
+
{
|
| 1932 |
+
"epoch": 2.023569545665995,
|
| 1933 |
+
"grad_norm": 2.7949812412261963,
|
| 1934 |
+
"learning_rate": 7.2376578645235365e-06,
|
| 1935 |
+
"loss": 0.3914073705673218,
|
| 1936 |
+
"step": 3264
|
| 1937 |
+
},
|
| 1938 |
+
{
|
| 1939 |
+
"epoch": 2.031012560086835,
|
| 1940 |
+
"grad_norm": 10.501336097717285,
|
| 1941 |
+
"learning_rate": 7.182548794489094e-06,
|
| 1942 |
+
"loss": 0.3871670166651408,
|
| 1943 |
+
"step": 3276
|
| 1944 |
+
},
|
| 1945 |
+
{
|
| 1946 |
+
"epoch": 2.0384555745076756,
|
| 1947 |
+
"grad_norm": 11.492402076721191,
|
| 1948 |
+
"learning_rate": 7.12743972445465e-06,
|
| 1949 |
+
"loss": 0.44295652707417804,
|
| 1950 |
+
"step": 3288
|
| 1951 |
+
},
|
| 1952 |
+
{
|
| 1953 |
+
"epoch": 2.045898588928516,
|
| 1954 |
+
"grad_norm": 8.688313484191895,
|
| 1955 |
+
"learning_rate": 7.072330654420208e-06,
|
| 1956 |
+
"loss": 0.4092850685119629,
|
| 1957 |
+
"step": 3300
|
| 1958 |
+
},
|
| 1959 |
+
{
|
| 1960 |
+
"epoch": 2.0533416033493563,
|
| 1961 |
+
"grad_norm": 5.402098178863525,
|
| 1962 |
+
"learning_rate": 7.017221584385764e-06,
|
| 1963 |
+
"loss": 0.41869743665059406,
|
| 1964 |
+
"step": 3312
|
| 1965 |
+
},
|
| 1966 |
+
{
|
| 1967 |
+
"epoch": 2.0607846177701967,
|
| 1968 |
+
"grad_norm": 3.6429481506347656,
|
| 1969 |
+
"learning_rate": 6.962112514351321e-06,
|
| 1970 |
+
"loss": 0.3916611671447754,
|
| 1971 |
+
"step": 3324
|
| 1972 |
+
},
|
| 1973 |
+
{
|
| 1974 |
+
"epoch": 2.0682276321910376,
|
| 1975 |
+
"grad_norm": 4.778937339782715,
|
| 1976 |
+
"learning_rate": 6.907003444316878e-06,
|
| 1977 |
+
"loss": 0.3913481632868449,
|
| 1978 |
+
"step": 3336
|
| 1979 |
+
},
|
| 1980 |
+
{
|
| 1981 |
+
"epoch": 2.075670646611878,
|
| 1982 |
+
"grad_norm": 4.281859874725342,
|
| 1983 |
+
"learning_rate": 6.851894374282435e-06,
|
| 1984 |
+
"loss": 0.380032738049825,
|
| 1985 |
+
"step": 3348
|
| 1986 |
+
},
|
| 1987 |
+
{
|
| 1988 |
+
"epoch": 2.0831136610327183,
|
| 1989 |
+
"grad_norm": 7.385513782501221,
|
| 1990 |
+
"learning_rate": 6.796785304247992e-06,
|
| 1991 |
+
"loss": 0.3545822699864705,
|
| 1992 |
+
"step": 3360
|
| 1993 |
+
},
|
| 1994 |
+
{
|
| 1995 |
+
"epoch": 2.0905566754535587,
|
| 1996 |
+
"grad_norm": 2.9248600006103516,
|
| 1997 |
+
"learning_rate": 6.741676234213548e-06,
|
| 1998 |
+
"loss": 0.419588565826416,
|
| 1999 |
+
"step": 3372
|
| 2000 |
+
},
|
| 2001 |
+
{
|
| 2002 |
+
"epoch": 2.097999689874399,
|
| 2003 |
+
"grad_norm": 3.0418336391448975,
|
| 2004 |
+
"learning_rate": 6.6865671641791055e-06,
|
| 2005 |
+
"loss": 0.4189613262812297,
|
| 2006 |
+
"step": 3384
|
| 2007 |
+
},
|
| 2008 |
+
{
|
| 2009 |
+
"epoch": 2.1054427042952395,
|
| 2010 |
+
"grad_norm": 4.628702640533447,
|
| 2011 |
+
"learning_rate": 6.631458094144662e-06,
|
| 2012 |
+
"loss": 0.38280495007832843,
|
| 2013 |
+
"step": 3396
|
| 2014 |
+
},
|
| 2015 |
+
{
|
| 2016 |
+
"epoch": 2.11288571871608,
|
| 2017 |
+
"grad_norm": 2.931917667388916,
|
| 2018 |
+
"learning_rate": 6.576349024110219e-06,
|
| 2019 |
+
"loss": 0.40134119987487793,
|
| 2020 |
+
"step": 3408
|
| 2021 |
+
},
|
| 2022 |
+
{
|
| 2023 |
+
"epoch": 2.1203287331369203,
|
| 2024 |
+
"grad_norm": 5.4905853271484375,
|
| 2025 |
+
"learning_rate": 6.521239954075776e-06,
|
| 2026 |
+
"loss": 0.3685312271118164,
|
| 2027 |
+
"step": 3420
|
| 2028 |
+
},
|
| 2029 |
+
{
|
| 2030 |
+
"epoch": 2.1277717475577607,
|
| 2031 |
+
"grad_norm": 2.9753782749176025,
|
| 2032 |
+
"learning_rate": 6.466130884041333e-06,
|
| 2033 |
+
"loss": 0.3878607749938965,
|
| 2034 |
+
"step": 3432
|
| 2035 |
+
},
|
| 2036 |
+
{
|
| 2037 |
+
"epoch": 2.1352147619786015,
|
| 2038 |
+
"grad_norm": 7.17921257019043,
|
| 2039 |
+
"learning_rate": 6.411021814006889e-06,
|
| 2040 |
+
"loss": 0.41369112332661945,
|
| 2041 |
+
"step": 3444
|
| 2042 |
+
},
|
| 2043 |
+
{
|
| 2044 |
+
"epoch": 2.142657776399442,
|
| 2045 |
+
"grad_norm": 13.806902885437012,
|
| 2046 |
+
"learning_rate": 6.355912743972445e-06,
|
| 2047 |
+
"loss": 0.43599124749501544,
|
| 2048 |
+
"step": 3456
|
| 2049 |
+
},
|
| 2050 |
+
{
|
| 2051 |
+
"epoch": 2.1501007908202823,
|
| 2052 |
+
"grad_norm": 3.4916634559631348,
|
| 2053 |
+
"learning_rate": 6.3008036739380026e-06,
|
| 2054 |
+
"loss": 0.3406885862350464,
|
| 2055 |
+
"step": 3468
|
| 2056 |
+
},
|
| 2057 |
+
{
|
| 2058 |
+
"epoch": 2.1575438052411227,
|
| 2059 |
+
"grad_norm": 6.193579196929932,
|
| 2060 |
+
"learning_rate": 6.245694603903559e-06,
|
| 2061 |
+
"loss": 0.3558163642883301,
|
| 2062 |
+
"step": 3480
|
| 2063 |
+
},
|
| 2064 |
+
{
|
| 2065 |
+
"epoch": 2.164986819661963,
|
| 2066 |
+
"grad_norm": 6.37896203994751,
|
| 2067 |
+
"learning_rate": 6.190585533869116e-06,
|
| 2068 |
+
"loss": 0.35776766141255695,
|
| 2069 |
+
"step": 3492
|
| 2070 |
+
},
|
| 2071 |
+
{
|
| 2072 |
+
"epoch": 2.1724298340828034,
|
| 2073 |
+
"grad_norm": 12.731496810913086,
|
| 2074 |
+
"learning_rate": 6.135476463834673e-06,
|
| 2075 |
+
"loss": 0.37972402572631836,
|
| 2076 |
+
"step": 3504
|
| 2077 |
+
},
|
| 2078 |
+
{
|
| 2079 |
+
"epoch": 2.179872848503644,
|
| 2080 |
+
"grad_norm": 19.98930549621582,
|
| 2081 |
+
"learning_rate": 6.08036739380023e-06,
|
| 2082 |
+
"loss": 0.42111217975616455,
|
| 2083 |
+
"step": 3516
|
| 2084 |
+
},
|
| 2085 |
+
{
|
| 2086 |
+
"epoch": 2.187315862924484,
|
| 2087 |
+
"grad_norm": 6.11861515045166,
|
| 2088 |
+
"learning_rate": 6.025258323765787e-06,
|
| 2089 |
+
"loss": 0.3672644297281901,
|
| 2090 |
+
"step": 3528
|
| 2091 |
+
},
|
| 2092 |
+
{
|
| 2093 |
+
"epoch": 2.194758877345325,
|
| 2094 |
+
"grad_norm": 11.929699897766113,
|
| 2095 |
+
"learning_rate": 5.970149253731343e-06,
|
| 2096 |
+
"loss": 0.4023996591567993,
|
| 2097 |
+
"step": 3540
|
| 2098 |
+
},
|
| 2099 |
+
{
|
| 2100 |
+
"epoch": 2.2022018917661654,
|
| 2101 |
+
"grad_norm": 17.26346206665039,
|
| 2102 |
+
"learning_rate": 5.9150401836969005e-06,
|
| 2103 |
+
"loss": 0.38841597239176434,
|
| 2104 |
+
"step": 3552
|
| 2105 |
+
},
|
| 2106 |
+
{
|
| 2107 |
+
"epoch": 2.209644906187006,
|
| 2108 |
+
"grad_norm": 9.183552742004395,
|
| 2109 |
+
"learning_rate": 5.859931113662457e-06,
|
| 2110 |
+
"loss": 0.42536401748657227,
|
| 2111 |
+
"step": 3564
|
| 2112 |
+
},
|
| 2113 |
+
{
|
| 2114 |
+
"epoch": 2.217087920607846,
|
| 2115 |
+
"grad_norm": 2.3118231296539307,
|
| 2116 |
+
"learning_rate": 5.804822043628014e-06,
|
| 2117 |
+
"loss": 0.4157342513402303,
|
| 2118 |
+
"step": 3576
|
| 2119 |
+
},
|
| 2120 |
+
{
|
| 2121 |
+
"epoch": 2.2245309350286866,
|
| 2122 |
+
"grad_norm": 6.309724807739258,
|
| 2123 |
+
"learning_rate": 5.749712973593571e-06,
|
| 2124 |
+
"loss": 0.4599275191624959,
|
| 2125 |
+
"step": 3588
|
| 2126 |
+
},
|
| 2127 |
+
{
|
| 2128 |
+
"epoch": 2.231973949449527,
|
| 2129 |
+
"grad_norm": 2.892469882965088,
|
| 2130 |
+
"learning_rate": 5.694603903559128e-06,
|
| 2131 |
+
"loss": 0.441303292910258,
|
| 2132 |
+
"step": 3600
|
| 2133 |
+
},
|
| 2134 |
+
{
|
| 2135 |
+
"epoch": 2.2394169638703674,
|
| 2136 |
+
"grad_norm": 6.523403167724609,
|
| 2137 |
+
"learning_rate": 5.6394948335246846e-06,
|
| 2138 |
+
"loss": 0.41275028387705487,
|
| 2139 |
+
"step": 3612
|
| 2140 |
+
},
|
| 2141 |
+
{
|
| 2142 |
+
"epoch": 2.2468599782912078,
|
| 2143 |
+
"grad_norm": 2.6101267337799072,
|
| 2144 |
+
"learning_rate": 5.584385763490242e-06,
|
| 2145 |
+
"loss": 0.41505225499471027,
|
| 2146 |
+
"step": 3624
|
| 2147 |
+
},
|
| 2148 |
+
{
|
| 2149 |
+
"epoch": 2.2543029927120486,
|
| 2150 |
+
"grad_norm": 5.343144416809082,
|
| 2151 |
+
"learning_rate": 5.529276693455798e-06,
|
| 2152 |
+
"loss": 0.38965781529744464,
|
| 2153 |
+
"step": 3636
|
| 2154 |
+
},
|
| 2155 |
+
{
|
| 2156 |
+
"epoch": 2.261746007132889,
|
| 2157 |
+
"grad_norm": 4.3300395011901855,
|
| 2158 |
+
"learning_rate": 5.474167623421355e-06,
|
| 2159 |
+
"loss": 0.4278339942296346,
|
| 2160 |
+
"step": 3648
|
| 2161 |
+
},
|
| 2162 |
+
{
|
| 2163 |
+
"epoch": 2.2691890215537294,
|
| 2164 |
+
"grad_norm": 5.109958171844482,
|
| 2165 |
+
"learning_rate": 5.419058553386912e-06,
|
| 2166 |
+
"loss": 0.366446574529012,
|
| 2167 |
+
"step": 3660
|
| 2168 |
+
},
|
| 2169 |
+
{
|
| 2170 |
+
"epoch": 2.2766320359745698,
|
| 2171 |
+
"grad_norm": 3.8399014472961426,
|
| 2172 |
+
"learning_rate": 5.363949483352469e-06,
|
| 2173 |
+
"loss": 0.3991047541300456,
|
| 2174 |
+
"step": 3672
|
| 2175 |
+
},
|
| 2176 |
+
{
|
| 2177 |
+
"epoch": 2.28407505039541,
|
| 2178 |
+
"grad_norm": 6.625537872314453,
|
| 2179 |
+
"learning_rate": 5.308840413318026e-06,
|
| 2180 |
+
"loss": 0.3346426486968994,
|
| 2181 |
+
"step": 3684
|
| 2182 |
+
},
|
| 2183 |
+
{
|
| 2184 |
+
"epoch": 2.2915180648162505,
|
| 2185 |
+
"grad_norm": 11.645654678344727,
|
| 2186 |
+
"learning_rate": 5.2537313432835825e-06,
|
| 2187 |
+
"loss": 0.3985482454299927,
|
| 2188 |
+
"step": 3696
|
| 2189 |
+
},
|
| 2190 |
+
{
|
| 2191 |
+
"epoch": 2.298961079237091,
|
| 2192 |
+
"grad_norm": 5.67885684967041,
|
| 2193 |
+
"learning_rate": 5.19862227324914e-06,
|
| 2194 |
+
"loss": 0.3815650939941406,
|
| 2195 |
+
"step": 3708
|
| 2196 |
+
},
|
| 2197 |
+
{
|
| 2198 |
+
"epoch": 2.3064040936579313,
|
| 2199 |
+
"grad_norm": 4.548233985900879,
|
| 2200 |
+
"learning_rate": 5.143513203214696e-06,
|
| 2201 |
+
"loss": 0.39840646584828693,
|
| 2202 |
+
"step": 3720
|
| 2203 |
+
},
|
| 2204 |
+
{
|
| 2205 |
+
"epoch": 2.3138471080787717,
|
| 2206 |
+
"grad_norm": 3.8364691734313965,
|
| 2207 |
+
"learning_rate": 5.088404133180253e-06,
|
| 2208 |
+
"loss": 0.4081765413284302,
|
| 2209 |
+
"step": 3732
|
| 2210 |
+
},
|
| 2211 |
+
{
|
| 2212 |
+
"epoch": 2.3212901224996125,
|
| 2213 |
+
"grad_norm": 2.5266079902648926,
|
| 2214 |
+
"learning_rate": 5.03329506314581e-06,
|
| 2215 |
+
"loss": 0.3613650401433309,
|
| 2216 |
+
"step": 3744
|
| 2217 |
+
},
|
| 2218 |
+
{
|
| 2219 |
+
"epoch": 2.328733136920453,
|
| 2220 |
+
"grad_norm": 7.049173831939697,
|
| 2221 |
+
"learning_rate": 4.9781859931113666e-06,
|
| 2222 |
+
"loss": 0.4112436771392822,
|
| 2223 |
+
"step": 3756
|
| 2224 |
+
},
|
| 2225 |
+
{
|
| 2226 |
+
"epoch": 2.3361761513412933,
|
| 2227 |
+
"grad_norm": 7.23855447769165,
|
| 2228 |
+
"learning_rate": 4.923076923076924e-06,
|
| 2229 |
+
"loss": 0.4015626907348633,
|
| 2230 |
+
"step": 3768
|
| 2231 |
+
},
|
| 2232 |
+
{
|
| 2233 |
+
"epoch": 2.3436191657621337,
|
| 2234 |
+
"grad_norm": 7.326627731323242,
|
| 2235 |
+
"learning_rate": 4.86796785304248e-06,
|
| 2236 |
+
"loss": 0.389956792195638,
|
| 2237 |
+
"step": 3780
|
| 2238 |
+
},
|
| 2239 |
+
{
|
| 2240 |
+
"epoch": 2.351062180182974,
|
| 2241 |
+
"grad_norm": 11.426876068115234,
|
| 2242 |
+
"learning_rate": 4.812858783008037e-06,
|
| 2243 |
+
"loss": 0.392941157023112,
|
| 2244 |
+
"step": 3792
|
| 2245 |
+
},
|
| 2246 |
+
{
|
| 2247 |
+
"epoch": 2.3585051946038145,
|
| 2248 |
+
"grad_norm": 5.058406352996826,
|
| 2249 |
+
"learning_rate": 4.757749712973594e-06,
|
| 2250 |
+
"loss": 0.388182799021403,
|
| 2251 |
+
"step": 3804
|
| 2252 |
+
},
|
| 2253 |
+
{
|
| 2254 |
+
"epoch": 2.365948209024655,
|
| 2255 |
+
"grad_norm": 7.783097267150879,
|
| 2256 |
+
"learning_rate": 4.702640642939151e-06,
|
| 2257 |
+
"loss": 0.4082544247309367,
|
| 2258 |
+
"step": 3816
|
| 2259 |
+
},
|
| 2260 |
+
{
|
| 2261 |
+
"epoch": 2.3733912234454952,
|
| 2262 |
+
"grad_norm": 4.8967084884643555,
|
| 2263 |
+
"learning_rate": 4.647531572904708e-06,
|
| 2264 |
+
"loss": 0.40780651569366455,
|
| 2265 |
+
"step": 3828
|
| 2266 |
+
},
|
| 2267 |
+
{
|
| 2268 |
+
"epoch": 2.3808342378663356,
|
| 2269 |
+
"grad_norm": 5.760252952575684,
|
| 2270 |
+
"learning_rate": 4.5924225028702645e-06,
|
| 2271 |
+
"loss": 0.4002196391423543,
|
| 2272 |
+
"step": 3840
|
| 2273 |
+
},
|
| 2274 |
+
{
|
| 2275 |
+
"epoch": 2.3882772522871765,
|
| 2276 |
+
"grad_norm": 4.79511022567749,
|
| 2277 |
+
"learning_rate": 4.537313432835822e-06,
|
| 2278 |
+
"loss": 0.3828426996866862,
|
| 2279 |
+
"step": 3852
|
| 2280 |
+
},
|
| 2281 |
+
{
|
| 2282 |
+
"epoch": 2.395720266708017,
|
| 2283 |
+
"grad_norm": 3.2499914169311523,
|
| 2284 |
+
"learning_rate": 4.4822043628013774e-06,
|
| 2285 |
+
"loss": 0.3649975061416626,
|
| 2286 |
+
"step": 3864
|
| 2287 |
+
},
|
| 2288 |
+
{
|
| 2289 |
+
"epoch": 2.4031632811288572,
|
| 2290 |
+
"grad_norm": 5.451921463012695,
|
| 2291 |
+
"learning_rate": 4.427095292766935e-06,
|
| 2292 |
+
"loss": 0.3998970588048299,
|
| 2293 |
+
"step": 3876
|
| 2294 |
+
},
|
| 2295 |
+
{
|
| 2296 |
+
"epoch": 2.4106062955496976,
|
| 2297 |
+
"grad_norm": 3.8105506896972656,
|
| 2298 |
+
"learning_rate": 4.371986222732491e-06,
|
| 2299 |
+
"loss": 0.45681726932525635,
|
| 2300 |
+
"step": 3888
|
| 2301 |
+
},
|
| 2302 |
+
{
|
| 2303 |
+
"epoch": 2.418049309970538,
|
| 2304 |
+
"grad_norm": 3.690845012664795,
|
| 2305 |
+
"learning_rate": 4.3168771526980486e-06,
|
| 2306 |
+
"loss": 0.3797287543614705,
|
| 2307 |
+
"step": 3900
|
| 2308 |
+
},
|
| 2309 |
+
{
|
| 2310 |
+
"epoch": 2.4254923243913784,
|
| 2311 |
+
"grad_norm": 12.44582748413086,
|
| 2312 |
+
"learning_rate": 4.261768082663605e-06,
|
| 2313 |
+
"loss": 0.47908584276835126,
|
| 2314 |
+
"step": 3912
|
| 2315 |
+
},
|
| 2316 |
+
{
|
| 2317 |
+
"epoch": 2.432935338812219,
|
| 2318 |
+
"grad_norm": 3.862395763397217,
|
| 2319 |
+
"learning_rate": 4.206659012629162e-06,
|
| 2320 |
+
"loss": 0.4127648671468099,
|
| 2321 |
+
"step": 3924
|
| 2322 |
+
},
|
| 2323 |
+
{
|
| 2324 |
+
"epoch": 2.4403783532330596,
|
| 2325 |
+
"grad_norm": 11.71980094909668,
|
| 2326 |
+
"learning_rate": 4.151549942594719e-06,
|
| 2327 |
+
"loss": 0.33937788009643555,
|
| 2328 |
+
"step": 3936
|
| 2329 |
+
},
|
| 2330 |
+
{
|
| 2331 |
+
"epoch": 2.4478213676539,
|
| 2332 |
+
"grad_norm": 4.254403591156006,
|
| 2333 |
+
"learning_rate": 4.096440872560276e-06,
|
| 2334 |
+
"loss": 0.3548990885416667,
|
| 2335 |
+
"step": 3948
|
| 2336 |
+
},
|
| 2337 |
+
{
|
| 2338 |
+
"epoch": 2.4552643820747404,
|
| 2339 |
+
"grad_norm": 5.00128173828125,
|
| 2340 |
+
"learning_rate": 4.041331802525833e-06,
|
| 2341 |
+
"loss": 0.4270055294036865,
|
| 2342 |
+
"step": 3960
|
| 2343 |
+
},
|
| 2344 |
+
{
|
| 2345 |
+
"epoch": 2.462707396495581,
|
| 2346 |
+
"grad_norm": 3.918459892272949,
|
| 2347 |
+
"learning_rate": 3.986222732491389e-06,
|
| 2348 |
+
"loss": 0.3760935465494792,
|
| 2349 |
+
"step": 3972
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"epoch": 2.470150410916421,
|
| 2353 |
+
"grad_norm": 11.43891716003418,
|
| 2354 |
+
"learning_rate": 3.9311136624569465e-06,
|
| 2355 |
+
"loss": 0.4183223644892375,
|
| 2356 |
+
"step": 3984
|
| 2357 |
+
},
|
| 2358 |
+
{
|
| 2359 |
+
"epoch": 2.4775934253372616,
|
| 2360 |
+
"grad_norm": 16.374967575073242,
|
| 2361 |
+
"learning_rate": 3.876004592422503e-06,
|
| 2362 |
+
"loss": 0.36837557951609295,
|
| 2363 |
+
"step": 3996
|
| 2364 |
+
},
|
| 2365 |
+
{
|
| 2366 |
+
"epoch": 2.485036439758102,
|
| 2367 |
+
"grad_norm": 4.490777015686035,
|
| 2368 |
+
"learning_rate": 3.82089552238806e-06,
|
| 2369 |
+
"loss": 0.4069160620371501,
|
| 2370 |
+
"step": 4008
|
| 2371 |
+
},
|
| 2372 |
+
{
|
| 2373 |
+
"epoch": 2.4924794541789423,
|
| 2374 |
+
"grad_norm": 8.420413970947266,
|
| 2375 |
+
"learning_rate": 3.7657864523536168e-06,
|
| 2376 |
+
"loss": 0.4271164337793986,
|
| 2377 |
+
"step": 4020
|
| 2378 |
+
},
|
| 2379 |
+
{
|
| 2380 |
+
"epoch": 2.4999224685997827,
|
| 2381 |
+
"grad_norm": 8.309126853942871,
|
| 2382 |
+
"learning_rate": 3.7106773823191737e-06,
|
| 2383 |
+
"loss": 0.3547343810399373,
|
| 2384 |
+
"step": 4032
|
| 2385 |
+
},
|
| 2386 |
+
{
|
| 2387 |
+
"epoch": 2.5073654830206236,
|
| 2388 |
+
"grad_norm": 14.98065185546875,
|
| 2389 |
+
"learning_rate": 3.6555683122847306e-06,
|
| 2390 |
+
"loss": 0.40314682324727374,
|
| 2391 |
+
"step": 4044
|
| 2392 |
+
},
|
| 2393 |
+
{
|
| 2394 |
+
"epoch": 2.514808497441464,
|
| 2395 |
+
"grad_norm": 16.558191299438477,
|
| 2396 |
+
"learning_rate": 3.600459242250287e-06,
|
| 2397 |
+
"loss": 0.36269084612528485,
|
| 2398 |
+
"step": 4056
|
| 2399 |
+
},
|
| 2400 |
+
{
|
| 2401 |
+
"epoch": 2.5222515118623043,
|
| 2402 |
+
"grad_norm": 6.547549724578857,
|
| 2403 |
+
"learning_rate": 3.545350172215844e-06,
|
| 2404 |
+
"loss": 0.36424537499745685,
|
| 2405 |
+
"step": 4068
|
| 2406 |
+
},
|
| 2407 |
+
{
|
| 2408 |
+
"epoch": 2.5296945262831447,
|
| 2409 |
+
"grad_norm": 4.773808002471924,
|
| 2410 |
+
"learning_rate": 3.490241102181401e-06,
|
| 2411 |
+
"loss": 0.37531224886576336,
|
| 2412 |
+
"step": 4080
|
| 2413 |
+
},
|
| 2414 |
+
{
|
| 2415 |
+
"epoch": 2.537137540703985,
|
| 2416 |
+
"grad_norm": 4.01258659362793,
|
| 2417 |
+
"learning_rate": 3.4351320321469578e-06,
|
| 2418 |
+
"loss": 0.36545733610788983,
|
| 2419 |
+
"step": 4092
|
| 2420 |
+
},
|
| 2421 |
+
{
|
| 2422 |
+
"epoch": 2.5445805551248255,
|
| 2423 |
+
"grad_norm": 10.372180938720703,
|
| 2424 |
+
"learning_rate": 3.3800229621125147e-06,
|
| 2425 |
+
"loss": 0.4671864112218221,
|
| 2426 |
+
"step": 4104
|
| 2427 |
+
},
|
| 2428 |
+
{
|
| 2429 |
+
"epoch": 2.552023569545666,
|
| 2430 |
+
"grad_norm": 3.3598952293395996,
|
| 2431 |
+
"learning_rate": 3.3249138920780716e-06,
|
| 2432 |
+
"loss": 0.3458172082901001,
|
| 2433 |
+
"step": 4116
|
| 2434 |
+
},
|
| 2435 |
+
{
|
| 2436 |
+
"epoch": 2.5594665839665063,
|
| 2437 |
+
"grad_norm": 11.469687461853027,
|
| 2438 |
+
"learning_rate": 3.2698048220436285e-06,
|
| 2439 |
+
"loss": 0.39522536595662433,
|
| 2440 |
+
"step": 4128
|
| 2441 |
+
},
|
| 2442 |
+
{
|
| 2443 |
+
"epoch": 2.5669095983873467,
|
| 2444 |
+
"grad_norm": 3.848041534423828,
|
| 2445 |
+
"learning_rate": 3.2146957520091854e-06,
|
| 2446 |
+
"loss": 0.41400329271952313,
|
| 2447 |
+
"step": 4140
|
| 2448 |
+
},
|
| 2449 |
+
{
|
| 2450 |
+
"epoch": 2.5743526128081875,
|
| 2451 |
+
"grad_norm": 4.791919231414795,
|
| 2452 |
+
"learning_rate": 3.159586681974742e-06,
|
| 2453 |
+
"loss": 0.393940011660258,
|
| 2454 |
+
"step": 4152
|
| 2455 |
+
},
|
| 2456 |
+
{
|
| 2457 |
+
"epoch": 2.581795627229028,
|
| 2458 |
+
"grad_norm": 21.486618041992188,
|
| 2459 |
+
"learning_rate": 3.1044776119402988e-06,
|
| 2460 |
+
"loss": 0.4398730993270874,
|
| 2461 |
+
"step": 4164
|
| 2462 |
+
},
|
| 2463 |
+
{
|
| 2464 |
+
"epoch": 2.5892386416498683,
|
| 2465 |
+
"grad_norm": 5.638022422790527,
|
| 2466 |
+
"learning_rate": 3.0493685419058557e-06,
|
| 2467 |
+
"loss": 0.3547349770863851,
|
| 2468 |
+
"step": 4176
|
| 2469 |
+
},
|
| 2470 |
+
{
|
| 2471 |
+
"epoch": 2.5966816560707087,
|
| 2472 |
+
"grad_norm": 7.414913177490234,
|
| 2473 |
+
"learning_rate": 2.9942594718714126e-06,
|
| 2474 |
+
"loss": 0.38705146312713623,
|
| 2475 |
+
"step": 4188
|
| 2476 |
+
},
|
| 2477 |
+
{
|
| 2478 |
+
"epoch": 2.604124670491549,
|
| 2479 |
+
"grad_norm": 6.696681976318359,
|
| 2480 |
+
"learning_rate": 2.9391504018369695e-06,
|
| 2481 |
+
"loss": 0.36440642674763996,
|
| 2482 |
+
"step": 4200
|
| 2483 |
+
},
|
| 2484 |
+
{
|
| 2485 |
+
"epoch": 2.6115676849123894,
|
| 2486 |
+
"grad_norm": 4.02039098739624,
|
| 2487 |
+
"learning_rate": 2.8840413318025264e-06,
|
| 2488 |
+
"loss": 0.39015217622121173,
|
| 2489 |
+
"step": 4212
|
| 2490 |
+
},
|
| 2491 |
+
{
|
| 2492 |
+
"epoch": 2.61901069933323,
|
| 2493 |
+
"grad_norm": 3.370777130126953,
|
| 2494 |
+
"learning_rate": 2.8289322617680833e-06,
|
| 2495 |
+
"loss": 0.4275425275166829,
|
| 2496 |
+
"step": 4224
|
| 2497 |
+
},
|
| 2498 |
+
{
|
| 2499 |
+
"epoch": 2.6264537137540707,
|
| 2500 |
+
"grad_norm": 8.47400951385498,
|
| 2501 |
+
"learning_rate": 2.7738231917336393e-06,
|
| 2502 |
+
"loss": 0.3559015194574992,
|
| 2503 |
+
"step": 4236
|
| 2504 |
+
},
|
| 2505 |
+
{
|
| 2506 |
+
"epoch": 2.6338967281749106,
|
| 2507 |
+
"grad_norm": 11.06500244140625,
|
| 2508 |
+
"learning_rate": 2.7187141216991963e-06,
|
| 2509 |
+
"loss": 0.3683815002441406,
|
| 2510 |
+
"step": 4248
|
| 2511 |
+
},
|
| 2512 |
+
{
|
| 2513 |
+
"epoch": 2.6413397425957514,
|
| 2514 |
+
"grad_norm": 3.4861528873443604,
|
| 2515 |
+
"learning_rate": 2.663605051664753e-06,
|
| 2516 |
+
"loss": 0.44681187470753986,
|
| 2517 |
+
"step": 4260
|
| 2518 |
+
},
|
| 2519 |
+
{
|
| 2520 |
+
"epoch": 2.648782757016592,
|
| 2521 |
+
"grad_norm": 10.642603874206543,
|
| 2522 |
+
"learning_rate": 2.60849598163031e-06,
|
| 2523 |
+
"loss": 0.4434703588485718,
|
| 2524 |
+
"step": 4272
|
| 2525 |
+
},
|
| 2526 |
+
{
|
| 2527 |
+
"epoch": 2.656225771437432,
|
| 2528 |
+
"grad_norm": 2.501110315322876,
|
| 2529 |
+
"learning_rate": 2.553386911595867e-06,
|
| 2530 |
+
"loss": 0.3525495131810506,
|
| 2531 |
+
"step": 4284
|
| 2532 |
+
},
|
| 2533 |
+
{
|
| 2534 |
+
"epoch": 2.6636687858582726,
|
| 2535 |
+
"grad_norm": 5.691764831542969,
|
| 2536 |
+
"learning_rate": 2.498277841561424e-06,
|
| 2537 |
+
"loss": 0.3853313128153483,
|
| 2538 |
+
"step": 4296
|
| 2539 |
+
},
|
| 2540 |
+
{
|
| 2541 |
+
"epoch": 2.671111800279113,
|
| 2542 |
+
"grad_norm": 4.1908135414123535,
|
| 2543 |
+
"learning_rate": 2.4431687715269808e-06,
|
| 2544 |
+
"loss": 0.38127346833546955,
|
| 2545 |
+
"step": 4308
|
| 2546 |
+
},
|
| 2547 |
+
{
|
| 2548 |
+
"epoch": 2.6785548146999534,
|
| 2549 |
+
"grad_norm": 9.538026809692383,
|
| 2550 |
+
"learning_rate": 2.3880597014925373e-06,
|
| 2551 |
+
"loss": 0.39995817343393963,
|
| 2552 |
+
"step": 4320
|
| 2553 |
+
},
|
| 2554 |
+
{
|
| 2555 |
+
"epoch": 2.6859978291207938,
|
| 2556 |
+
"grad_norm": 8.436595916748047,
|
| 2557 |
+
"learning_rate": 2.332950631458094e-06,
|
| 2558 |
+
"loss": 0.3635564645131429,
|
| 2559 |
+
"step": 4332
|
| 2560 |
+
},
|
| 2561 |
+
{
|
| 2562 |
+
"epoch": 2.6934408435416346,
|
| 2563 |
+
"grad_norm": 2.5905513763427734,
|
| 2564 |
+
"learning_rate": 2.277841561423651e-06,
|
| 2565 |
+
"loss": 0.46339670817057294,
|
| 2566 |
+
"step": 4344
|
| 2567 |
+
},
|
| 2568 |
+
{
|
| 2569 |
+
"epoch": 2.7008838579624745,
|
| 2570 |
+
"grad_norm": 6.738951206207275,
|
| 2571 |
+
"learning_rate": 2.222732491389208e-06,
|
| 2572 |
+
"loss": 0.3373739719390869,
|
| 2573 |
+
"step": 4356
|
| 2574 |
+
},
|
| 2575 |
+
{
|
| 2576 |
+
"epoch": 2.7083268723833154,
|
| 2577 |
+
"grad_norm": 5.625753402709961,
|
| 2578 |
+
"learning_rate": 2.167623421354765e-06,
|
| 2579 |
+
"loss": 0.3713107109069824,
|
| 2580 |
+
"step": 4368
|
| 2581 |
+
},
|
| 2582 |
+
{
|
| 2583 |
+
"epoch": 2.7157698868041558,
|
| 2584 |
+
"grad_norm": 3.6908581256866455,
|
| 2585 |
+
"learning_rate": 2.1125143513203218e-06,
|
| 2586 |
+
"loss": 0.3845006227493286,
|
| 2587 |
+
"step": 4380
|
| 2588 |
+
},
|
| 2589 |
+
{
|
| 2590 |
+
"epoch": 2.723212901224996,
|
| 2591 |
+
"grad_norm": 5.123325824737549,
|
| 2592 |
+
"learning_rate": 2.0574052812858787e-06,
|
| 2593 |
+
"loss": 0.3693963686625163,
|
| 2594 |
+
"step": 4392
|
| 2595 |
+
},
|
| 2596 |
+
{
|
| 2597 |
+
"epoch": 2.7306559156458365,
|
| 2598 |
+
"grad_norm": 5.600500583648682,
|
| 2599 |
+
"learning_rate": 2.002296211251435e-06,
|
| 2600 |
+
"loss": 0.4005578358968099,
|
| 2601 |
+
"step": 4404
|
| 2602 |
+
},
|
| 2603 |
+
{
|
| 2604 |
+
"epoch": 2.738098930066677,
|
| 2605 |
+
"grad_norm": 4.9075775146484375,
|
| 2606 |
+
"learning_rate": 1.947187141216992e-06,
|
| 2607 |
+
"loss": 0.44304617245992023,
|
| 2608 |
+
"step": 4416
|
| 2609 |
+
},
|
| 2610 |
+
{
|
| 2611 |
+
"epoch": 2.7455419444875173,
|
| 2612 |
+
"grad_norm": 2.535568952560425,
|
| 2613 |
+
"learning_rate": 1.892078071182549e-06,
|
| 2614 |
+
"loss": 0.36018415292104083,
|
| 2615 |
+
"step": 4428
|
| 2616 |
+
},
|
| 2617 |
+
{
|
| 2618 |
+
"epoch": 2.7529849589083577,
|
| 2619 |
+
"grad_norm": 3.863154888153076,
|
| 2620 |
+
"learning_rate": 1.8369690011481059e-06,
|
| 2621 |
+
"loss": 0.3833086093266805,
|
| 2622 |
+
"step": 4440
|
| 2623 |
+
},
|
| 2624 |
+
{
|
| 2625 |
+
"epoch": 2.7604279733291985,
|
| 2626 |
+
"grad_norm": 3.38565731048584,
|
| 2627 |
+
"learning_rate": 1.7818599311136626e-06,
|
| 2628 |
+
"loss": 0.36296629905700684,
|
| 2629 |
+
"step": 4452
|
| 2630 |
+
},
|
| 2631 |
+
{
|
| 2632 |
+
"epoch": 2.767870987750039,
|
| 2633 |
+
"grad_norm": 3.979094982147217,
|
| 2634 |
+
"learning_rate": 1.7267508610792195e-06,
|
| 2635 |
+
"loss": 0.400799036026001,
|
| 2636 |
+
"step": 4464
|
| 2637 |
+
},
|
| 2638 |
+
{
|
| 2639 |
+
"epoch": 2.7753140021708793,
|
| 2640 |
+
"grad_norm": 3.6006662845611572,
|
| 2641 |
+
"learning_rate": 1.6716417910447764e-06,
|
| 2642 |
+
"loss": 0.38404210408528644,
|
| 2643 |
+
"step": 4476
|
| 2644 |
+
},
|
| 2645 |
+
{
|
| 2646 |
+
"epoch": 2.7827570165917197,
|
| 2647 |
+
"grad_norm": 9.927759170532227,
|
| 2648 |
+
"learning_rate": 1.6165327210103333e-06,
|
| 2649 |
+
"loss": 0.47922762235005695,
|
| 2650 |
+
"step": 4488
|
| 2651 |
+
},
|
| 2652 |
+
{
|
| 2653 |
+
"epoch": 2.79020003101256,
|
| 2654 |
+
"grad_norm": 4.767171859741211,
|
| 2655 |
+
"learning_rate": 1.5614236509758898e-06,
|
| 2656 |
+
"loss": 0.40151556332906085,
|
| 2657 |
+
"step": 4500
|
| 2658 |
+
},
|
| 2659 |
+
{
|
| 2660 |
+
"epoch": 2.7976430454334005,
|
| 2661 |
+
"grad_norm": 5.649435043334961,
|
| 2662 |
+
"learning_rate": 1.5063145809414467e-06,
|
| 2663 |
+
"loss": 0.3603046735127767,
|
| 2664 |
+
"step": 4512
|
| 2665 |
+
},
|
| 2666 |
+
{
|
| 2667 |
+
"epoch": 2.805086059854241,
|
| 2668 |
+
"grad_norm": 11.296677589416504,
|
| 2669 |
+
"learning_rate": 1.4512055109070036e-06,
|
| 2670 |
+
"loss": 0.38084761301676434,
|
| 2671 |
+
"step": 4524
|
| 2672 |
+
},
|
| 2673 |
+
{
|
| 2674 |
+
"epoch": 2.8125290742750813,
|
| 2675 |
+
"grad_norm": 2.71022629737854,
|
| 2676 |
+
"learning_rate": 1.3960964408725605e-06,
|
| 2677 |
+
"loss": 0.3726603190104167,
|
| 2678 |
+
"step": 4536
|
| 2679 |
+
},
|
| 2680 |
+
{
|
| 2681 |
+
"epoch": 2.8199720886959216,
|
| 2682 |
+
"grad_norm": 3.849479913711548,
|
| 2683 |
+
"learning_rate": 1.3409873708381172e-06,
|
| 2684 |
+
"loss": 0.3995700279871623,
|
| 2685 |
+
"step": 4548
|
| 2686 |
+
},
|
| 2687 |
+
{
|
| 2688 |
+
"epoch": 2.8274151031167625,
|
| 2689 |
+
"grad_norm": 14.668109893798828,
|
| 2690 |
+
"learning_rate": 1.285878300803674e-06,
|
| 2691 |
+
"loss": 0.39227835337320965,
|
| 2692 |
+
"step": 4560
|
| 2693 |
+
},
|
| 2694 |
+
{
|
| 2695 |
+
"epoch": 2.834858117537603,
|
| 2696 |
+
"grad_norm": 3.9545083045959473,
|
| 2697 |
+
"learning_rate": 1.230769230769231e-06,
|
| 2698 |
+
"loss": 0.42009902000427246,
|
| 2699 |
+
"step": 4572
|
| 2700 |
+
},
|
| 2701 |
+
{
|
| 2702 |
+
"epoch": 2.8423011319584432,
|
| 2703 |
+
"grad_norm": 5.8148298263549805,
|
| 2704 |
+
"learning_rate": 1.1756601607347877e-06,
|
| 2705 |
+
"loss": 0.39560989538828534,
|
| 2706 |
+
"step": 4584
|
| 2707 |
+
},
|
| 2708 |
+
{
|
| 2709 |
+
"epoch": 2.8497441463792836,
|
| 2710 |
+
"grad_norm": 6.249505996704102,
|
| 2711 |
+
"learning_rate": 1.1205510907003444e-06,
|
| 2712 |
+
"loss": 0.42494750022888184,
|
| 2713 |
+
"step": 4596
|
| 2714 |
+
},
|
| 2715 |
+
{
|
| 2716 |
+
"epoch": 2.857187160800124,
|
| 2717 |
+
"grad_norm": 4.1339921951293945,
|
| 2718 |
+
"learning_rate": 1.0654420206659013e-06,
|
| 2719 |
+
"loss": 0.5030697584152222,
|
| 2720 |
+
"step": 4608
|
| 2721 |
+
},
|
| 2722 |
+
{
|
| 2723 |
+
"epoch": 2.8646301752209644,
|
| 2724 |
+
"grad_norm": 13.68895435333252,
|
| 2725 |
+
"learning_rate": 1.0103329506314582e-06,
|
| 2726 |
+
"loss": 0.36397520701090497,
|
| 2727 |
+
"step": 4620
|
| 2728 |
+
},
|
| 2729 |
+
{
|
| 2730 |
+
"epoch": 2.872073189641805,
|
| 2731 |
+
"grad_norm": 2.826042890548706,
|
| 2732 |
+
"learning_rate": 9.55223880597015e-07,
|
| 2733 |
+
"loss": 0.3502591848373413,
|
| 2734 |
+
"step": 4632
|
| 2735 |
+
},
|
| 2736 |
+
{
|
| 2737 |
+
"epoch": 2.8795162040626456,
|
| 2738 |
+
"grad_norm": 6.833806991577148,
|
| 2739 |
+
"learning_rate": 9.001148105625718e-07,
|
| 2740 |
+
"loss": 0.3613890012105306,
|
| 2741 |
+
"step": 4644
|
| 2742 |
+
},
|
| 2743 |
+
{
|
| 2744 |
+
"epoch": 2.8869592184834856,
|
| 2745 |
+
"grad_norm": 4.942678451538086,
|
| 2746 |
+
"learning_rate": 8.450057405281287e-07,
|
| 2747 |
+
"loss": 0.39194099108378094,
|
| 2748 |
+
"step": 4656
|
| 2749 |
+
},
|
| 2750 |
+
{
|
| 2751 |
+
"epoch": 2.8944022329043264,
|
| 2752 |
+
"grad_norm": 4.509676456451416,
|
| 2753 |
+
"learning_rate": 7.898966704936855e-07,
|
| 2754 |
+
"loss": 0.351750651995341,
|
| 2755 |
+
"step": 4668
|
| 2756 |
+
},
|
| 2757 |
+
{
|
| 2758 |
+
"epoch": 2.901845247325167,
|
| 2759 |
+
"grad_norm": 8.305526733398438,
|
| 2760 |
+
"learning_rate": 7.347876004592424e-07,
|
| 2761 |
+
"loss": 0.40360478560129803,
|
| 2762 |
+
"step": 4680
|
| 2763 |
+
},
|
| 2764 |
+
{
|
| 2765 |
+
"epoch": 2.909288261746007,
|
| 2766 |
+
"grad_norm": 4.9328765869140625,
|
| 2767 |
+
"learning_rate": 6.796785304247991e-07,
|
| 2768 |
+
"loss": 0.33100277185440063,
|
| 2769 |
+
"step": 4692
|
| 2770 |
+
},
|
| 2771 |
+
{
|
| 2772 |
+
"epoch": 2.9167312761668476,
|
| 2773 |
+
"grad_norm": 4.945671558380127,
|
| 2774 |
+
"learning_rate": 6.24569460390356e-07,
|
| 2775 |
+
"loss": 0.39974749088287354,
|
| 2776 |
+
"step": 4704
|
| 2777 |
+
},
|
| 2778 |
+
{
|
| 2779 |
+
"epoch": 2.924174290587688,
|
| 2780 |
+
"grad_norm": 9.925528526306152,
|
| 2781 |
+
"learning_rate": 5.694603903559128e-07,
|
| 2782 |
+
"loss": 0.4116141001383464,
|
| 2783 |
+
"step": 4716
|
| 2784 |
+
},
|
| 2785 |
+
{
|
| 2786 |
+
"epoch": 2.9316173050085284,
|
| 2787 |
+
"grad_norm": 4.063233375549316,
|
| 2788 |
+
"learning_rate": 5.143513203214697e-07,
|
| 2789 |
+
"loss": 0.3659325838088989,
|
| 2790 |
+
"step": 4728
|
| 2791 |
+
},
|
| 2792 |
+
{
|
| 2793 |
+
"epoch": 2.9390603194293687,
|
| 2794 |
+
"grad_norm": 3.5343589782714844,
|
| 2795 |
+
"learning_rate": 4.5924225028702647e-07,
|
| 2796 |
+
"loss": 0.3983626365661621,
|
| 2797 |
+
"step": 4740
|
| 2798 |
+
},
|
| 2799 |
+
{
|
| 2800 |
+
"epoch": 2.9465033338502096,
|
| 2801 |
+
"grad_norm": 6.534095764160156,
|
| 2802 |
+
"learning_rate": 4.041331802525833e-07,
|
| 2803 |
+
"loss": 0.393149733543396,
|
| 2804 |
+
"step": 4752
|
| 2805 |
+
},
|
| 2806 |
+
{
|
| 2807 |
+
"epoch": 2.9539463482710495,
|
| 2808 |
+
"grad_norm": 3.4787096977233887,
|
| 2809 |
+
"learning_rate": 3.490241102181401e-07,
|
| 2810 |
+
"loss": 0.3340187867482503,
|
| 2811 |
+
"step": 4764
|
| 2812 |
+
},
|
| 2813 |
+
{
|
| 2814 |
+
"epoch": 2.9613893626918903,
|
| 2815 |
+
"grad_norm": 5.42100191116333,
|
| 2816 |
+
"learning_rate": 2.939150401836969e-07,
|
| 2817 |
+
"loss": 0.3814918597539266,
|
| 2818 |
+
"step": 4776
|
| 2819 |
+
},
|
| 2820 |
+
{
|
| 2821 |
+
"epoch": 2.9688323771127307,
|
| 2822 |
+
"grad_norm": 4.148738861083984,
|
| 2823 |
+
"learning_rate": 2.3880597014925377e-07,
|
| 2824 |
+
"loss": 0.4039960702260335,
|
| 2825 |
+
"step": 4788
|
| 2826 |
+
},
|
| 2827 |
+
{
|
| 2828 |
+
"epoch": 2.976275391533571,
|
| 2829 |
+
"grad_norm": 4.3285746574401855,
|
| 2830 |
+
"learning_rate": 1.836969001148106e-07,
|
| 2831 |
+
"loss": 0.34236987431844074,
|
| 2832 |
+
"step": 4800
|
| 2833 |
+
},
|
| 2834 |
+
{
|
| 2835 |
+
"epoch": 2.9837184059544115,
|
| 2836 |
+
"grad_norm": 2.8112664222717285,
|
| 2837 |
+
"learning_rate": 1.2858783008036742e-07,
|
| 2838 |
+
"loss": 0.3349067767461141,
|
| 2839 |
+
"step": 4812
|
| 2840 |
+
},
|
| 2841 |
+
{
|
| 2842 |
+
"epoch": 2.991161420375252,
|
| 2843 |
+
"grad_norm": 4.724297523498535,
|
| 2844 |
+
"learning_rate": 7.347876004592423e-08,
|
| 2845 |
+
"loss": 0.38507378101348877,
|
| 2846 |
+
"step": 4824
|
| 2847 |
+
},
|
| 2848 |
+
{
|
| 2849 |
+
"epoch": 2.9986044347960923,
|
| 2850 |
+
"grad_norm": 7.1218132972717285,
|
| 2851 |
+
"learning_rate": 1.8369690011481057e-08,
|
| 2852 |
+
"loss": 0.34174474080403644,
|
| 2853 |
+
"step": 4836
|
| 2854 |
+
},
|
| 2855 |
+
{
|
| 2856 |
+
"epoch": 3.0,
|
| 2857 |
+
"eval_f1": 0.43726749573500223,
|
| 2858 |
+
"eval_loss": 0.12126699090003967,
|
| 2859 |
+
"eval_precision": 0.4012637195169362,
|
| 2860 |
+
"eval_recall": 0.4913673269074057,
|
| 2861 |
+
"eval_runtime": 522.9994,
|
| 2862 |
+
"eval_samples_per_second": 73.975,
|
| 2863 |
+
"eval_steps_per_second": 1.543,
|
| 2864 |
+
"step": 4839
|
| 2865 |
+
}
|
| 2866 |
+
],
|
| 2867 |
+
"logging_steps": 12,
|
| 2868 |
+
"max_steps": 4839,
|
| 2869 |
+
"num_input_tokens_seen": 0,
|
| 2870 |
+
"num_train_epochs": 3,
|
| 2871 |
+
"save_steps": 500,
|
| 2872 |
+
"stateful_callbacks": {
|
| 2873 |
+
"TrainerControl": {
|
| 2874 |
+
"args": {
|
| 2875 |
+
"should_epoch_stop": false,
|
| 2876 |
+
"should_evaluate": false,
|
| 2877 |
+
"should_log": false,
|
| 2878 |
+
"should_save": true,
|
| 2879 |
+
"should_training_stop": true
|
| 2880 |
+
},
|
| 2881 |
+
"attributes": {}
|
| 2882 |
+
}
|
| 2883 |
+
},
|
| 2884 |
+
"total_flos": 1.2215684896188826e+17,
|
| 2885 |
+
"train_batch_size": 48,
|
| 2886 |
+
"trial_name": null,
|
| 2887 |
+
"trial_params": null
|
| 2888 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4a3e9b6857a14b78f40f49b62af19e176a243af623343fdde91f131fdf81cfbe
|
| 3 |
+
size 4792
|