Gaduh Hartawan
commited on
Commit
•
6a48c5e
1
Parent(s):
310b105
first commit
Browse files- config.json +68 -0
- generation_config.json +16 -0
- model.safetensors +3 -0
- optimizer.pt +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- trainer_state.json +3381 -0
- training_args.bin +3 -0
config.json
ADDED
|
@@ -0,0 +1,68 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "facebook/bart-large-cnn",
|
| 3 |
+
"_num_labels": 3,
|
| 4 |
+
"activation_dropout": 0,
|
| 5 |
+
"activation_function": "gelu",
|
| 6 |
+
"add_final_layer_norm": false,
|
| 7 |
+
"architectures": ["BartForConditionalGeneration"],
|
| 8 |
+
"attention_dropout": 0,
|
| 9 |
+
"bos_token_id": 0,
|
| 10 |
+
"classif_dropout": 0,
|
| 11 |
+
"classifier_dropout": 0,
|
| 12 |
+
"d_model": 1024,
|
| 13 |
+
"decoder_attention_heads": 16,
|
| 14 |
+
"decoder_ffn_dim": 4096,
|
| 15 |
+
"decoder_layerdrop": 0,
|
| 16 |
+
"decoder_layers": 12,
|
| 17 |
+
"decoder_start_token_id": 2,
|
| 18 |
+
"dropout": 0.1,
|
| 19 |
+
"early_stopping": true,
|
| 20 |
+
"encoder_attention_heads": 16,
|
| 21 |
+
"encoder_ffn_dim": 4096,
|
| 22 |
+
"encoder_layerdrop": 0,
|
| 23 |
+
"encoder_layers": 12,
|
| 24 |
+
"eos_token_id": 2,
|
| 25 |
+
"force_bos_token_to_be_generated": true,
|
| 26 |
+
"forced_bos_token_id": 0,
|
| 27 |
+
"forced_eos_token_id": 2,
|
| 28 |
+
"gradient_checkpointing": false,
|
| 29 |
+
"id2label": {
|
| 30 |
+
"0": "LABEL_0",
|
| 31 |
+
"1": "LABEL_1",
|
| 32 |
+
"2": "LABEL_2"
|
| 33 |
+
},
|
| 34 |
+
"init_std": 0.02,
|
| 35 |
+
"is_encoder_decoder": true,
|
| 36 |
+
"label2id": {
|
| 37 |
+
"LABEL_0": 0,
|
| 38 |
+
"LABEL_1": 1,
|
| 39 |
+
"LABEL_2": 2
|
| 40 |
+
},
|
| 41 |
+
"length_penalty": 2,
|
| 42 |
+
"max_length": 142,
|
| 43 |
+
"max_position_embeddings": 1024,
|
| 44 |
+
"min_length": 56,
|
| 45 |
+
"model_type": "bart",
|
| 46 |
+
"no_repeat_ngram_size": 3,
|
| 47 |
+
"normalize_before": false,
|
| 48 |
+
"num_beams": 4,
|
| 49 |
+
"num_hidden_layers": 12,
|
| 50 |
+
"output_past": true,
|
| 51 |
+
"pad_token_id": 1,
|
| 52 |
+
"prefix": " ",
|
| 53 |
+
"scale_embedding": false,
|
| 54 |
+
"task_specific_params": {
|
| 55 |
+
"summarization": {
|
| 56 |
+
"early_stopping": true,
|
| 57 |
+
"length_penalty": 2,
|
| 58 |
+
"max_length": 142,
|
| 59 |
+
"min_length": 56,
|
| 60 |
+
"no_repeat_ngram_size": 3,
|
| 61 |
+
"num_beams": 4
|
| 62 |
+
}
|
| 63 |
+
},
|
| 64 |
+
"torch_dtype": "float32",
|
| 65 |
+
"transformers_version": "4.39.3",
|
| 66 |
+
"use_cache": true,
|
| 67 |
+
"vocab_size": 50264
|
| 68 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 0,
|
| 4 |
+
"decoder_start_token_id": 2,
|
| 5 |
+
"early_stopping": true,
|
| 6 |
+
"eos_token_id": 2,
|
| 7 |
+
"forced_bos_token_id": 0,
|
| 8 |
+
"forced_eos_token_id": 2,
|
| 9 |
+
"length_penalty": 2,
|
| 10 |
+
"max_length": 142,
|
| 11 |
+
"min_length": 56,
|
| 12 |
+
"no_repeat_ngram_size": 3,
|
| 13 |
+
"num_beams": 4,
|
| 14 |
+
"pad_token_id": 1,
|
| 15 |
+
"transformers_version": "4.39.3"
|
| 16 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8e8c441683222cd6beb10c91ea2770548a94904a0bf52a9b009b5f4cbf7f4e8d
|
| 3 |
+
size 1625422896
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6a6acdbb3821e0d7a42e709d12c007e0f0c8540a3ec90392679150e72fab8e3d
|
| 3 |
+
size 3250751759
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8971b12a176fe7e5b5fb742e30c037a42e0e6846be001ecfa37949a7b05835cb
|
| 3 |
+
size 14244
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c3bec67b96ed587061623fcdd062c65bf54535d7990f2df1520a1ce081ee727f
|
| 3 |
+
size 1064
|
trainer_state.json
ADDED
|
@@ -0,0 +1,3381 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.9902828495388996,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 48000,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0,
|
| 13 |
+
"grad_norm": 13.376683235168457,
|
| 14 |
+
"learning_rate": 0.00001,
|
| 15 |
+
"loss": 4.2362,
|
| 16 |
+
"step": 100
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0,
|
| 20 |
+
"grad_norm": 2.2011091709136963,
|
| 21 |
+
"learning_rate": 0.00002,
|
| 22 |
+
"loss": 0.6427,
|
| 23 |
+
"step": 200
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.01,
|
| 27 |
+
"grad_norm": 2.1116416454315186,
|
| 28 |
+
"learning_rate": 0.00003,
|
| 29 |
+
"loss": 0.5047,
|
| 30 |
+
"step": 300
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.01,
|
| 34 |
+
"grad_norm": 2.325719118118286,
|
| 35 |
+
"learning_rate": 0.00004,
|
| 36 |
+
"loss": 0.4955,
|
| 37 |
+
"step": 400
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.01,
|
| 41 |
+
"grad_norm": 1.2606059312820435,
|
| 42 |
+
"learning_rate": 0.00005,
|
| 43 |
+
"loss": 0.5104,
|
| 44 |
+
"step": 500
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.01,
|
| 48 |
+
"grad_norm": 1.9306057691574097,
|
| 49 |
+
"learning_rate": 0.000049895770361259934,
|
| 50 |
+
"loss": 0.5024,
|
| 51 |
+
"step": 600
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.01,
|
| 55 |
+
"grad_norm": 1.3757009506225586,
|
| 56 |
+
"learning_rate": 0.00004979154072251986,
|
| 57 |
+
"loss": 0.4814,
|
| 58 |
+
"step": 700
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.02,
|
| 62 |
+
"grad_norm": 1.6575660705566406,
|
| 63 |
+
"learning_rate": 0.00004968731108377979,
|
| 64 |
+
"loss": 0.4675,
|
| 65 |
+
"step": 800
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.02,
|
| 69 |
+
"grad_norm": 1.6257671117782593,
|
| 70 |
+
"learning_rate": 0.000049583081445039714,
|
| 71 |
+
"loss": 0.4673,
|
| 72 |
+
"step": 900
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.02,
|
| 76 |
+
"grad_norm": 1.1650177240371704,
|
| 77 |
+
"learning_rate": 0.000049478851806299645,
|
| 78 |
+
"loss": 0.4635,
|
| 79 |
+
"step": 1000
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.02,
|
| 83 |
+
"grad_norm": 1.2291123867034912,
|
| 84 |
+
"learning_rate": 0.00004937462216755957,
|
| 85 |
+
"loss": 0.4645,
|
| 86 |
+
"step": 1100
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.02,
|
| 90 |
+
"grad_norm": 1.094923973083496,
|
| 91 |
+
"learning_rate": 0.0000492703925288195,
|
| 92 |
+
"loss": 0.463,
|
| 93 |
+
"step": 1200
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.03,
|
| 97 |
+
"grad_norm": 1.4243600368499756,
|
| 98 |
+
"learning_rate": 0.000049166162890079426,
|
| 99 |
+
"loss": 0.4305,
|
| 100 |
+
"step": 1300
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.03,
|
| 104 |
+
"grad_norm": 1.5323872566223145,
|
| 105 |
+
"learning_rate": 0.00004906193325133936,
|
| 106 |
+
"loss": 0.4431,
|
| 107 |
+
"step": 1400
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.03,
|
| 111 |
+
"grad_norm": 1.4820982217788696,
|
| 112 |
+
"learning_rate": 0.00004895770361259928,
|
| 113 |
+
"loss": 0.4301,
|
| 114 |
+
"step": 1500
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.03,
|
| 118 |
+
"grad_norm": 1.6124228239059448,
|
| 119 |
+
"learning_rate": 0.00004885347397385921,
|
| 120 |
+
"loss": 0.4537,
|
| 121 |
+
"step": 1600
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.04,
|
| 125 |
+
"grad_norm": 1.5138214826583862,
|
| 126 |
+
"learning_rate": 0.00004874924433511914,
|
| 127 |
+
"loss": 0.4101,
|
| 128 |
+
"step": 1700
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.04,
|
| 132 |
+
"grad_norm": 2.3704726696014404,
|
| 133 |
+
"learning_rate": 0.00004864501469637907,
|
| 134 |
+
"loss": 0.4113,
|
| 135 |
+
"step": 1800
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.04,
|
| 139 |
+
"grad_norm": 1.4032459259033203,
|
| 140 |
+
"learning_rate": 0.00004854078505763899,
|
| 141 |
+
"loss": 0.4297,
|
| 142 |
+
"step": 1900
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.04,
|
| 146 |
+
"grad_norm": 1.3240689039230347,
|
| 147 |
+
"learning_rate": 0.000048436555418898924,
|
| 148 |
+
"loss": 0.4193,
|
| 149 |
+
"step": 2000
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.04,
|
| 153 |
+
"grad_norm": 1.411370038986206,
|
| 154 |
+
"learning_rate": 0.00004833232578015885,
|
| 155 |
+
"loss": 0.4118,
|
| 156 |
+
"step": 2100
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.05,
|
| 160 |
+
"grad_norm": 1.4880633354187012,
|
| 161 |
+
"learning_rate": 0.00004822809614141878,
|
| 162 |
+
"loss": 0.4284,
|
| 163 |
+
"step": 2200
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.05,
|
| 167 |
+
"grad_norm": 1.3598105907440186,
|
| 168 |
+
"learning_rate": 0.000048123866502678705,
|
| 169 |
+
"loss": 0.4119,
|
| 170 |
+
"step": 2300
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.05,
|
| 174 |
+
"grad_norm": 1.2027459144592285,
|
| 175 |
+
"learning_rate": 0.000048019636863938636,
|
| 176 |
+
"loss": 0.4073,
|
| 177 |
+
"step": 2400
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.05,
|
| 181 |
+
"grad_norm": 1.35155189037323,
|
| 182 |
+
"learning_rate": 0.00004791540722519856,
|
| 183 |
+
"loss": 0.4134,
|
| 184 |
+
"step": 2500
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.05,
|
| 188 |
+
"grad_norm": 1.8489034175872803,
|
| 189 |
+
"learning_rate": 0.00004781117758645849,
|
| 190 |
+
"loss": 0.4069,
|
| 191 |
+
"step": 2600
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.06,
|
| 195 |
+
"grad_norm": 1.2636560201644897,
|
| 196 |
+
"learning_rate": 0.000047706947947718416,
|
| 197 |
+
"loss": 0.4224,
|
| 198 |
+
"step": 2700
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.06,
|
| 202 |
+
"grad_norm": 1.4396322965621948,
|
| 203 |
+
"learning_rate": 0.00004760271830897835,
|
| 204 |
+
"loss": 0.3951,
|
| 205 |
+
"step": 2800
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.06,
|
| 209 |
+
"grad_norm": 1.3303143978118896,
|
| 210 |
+
"learning_rate": 0.00004749848867023827,
|
| 211 |
+
"loss": 0.4111,
|
| 212 |
+
"step": 2900
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.06,
|
| 216 |
+
"grad_norm": 1.2054022550582886,
|
| 217 |
+
"learning_rate": 0.0000473942590314982,
|
| 218 |
+
"loss": 0.4155,
|
| 219 |
+
"step": 3000
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.06,
|
| 223 |
+
"grad_norm": 1.1086294651031494,
|
| 224 |
+
"learning_rate": 0.00004729002939275813,
|
| 225 |
+
"loss": 0.3912,
|
| 226 |
+
"step": 3100
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.07,
|
| 230 |
+
"grad_norm": 1.0423736572265625,
|
| 231 |
+
"learning_rate": 0.00004718579975401806,
|
| 232 |
+
"loss": 0.4174,
|
| 233 |
+
"step": 3200
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.07,
|
| 237 |
+
"grad_norm": 1.1115955114364624,
|
| 238 |
+
"learning_rate": 0.000047081570115277984,
|
| 239 |
+
"loss": 0.4293,
|
| 240 |
+
"step": 3300
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.07,
|
| 244 |
+
"grad_norm": 1.1089532375335693,
|
| 245 |
+
"learning_rate": 0.000046977340476537915,
|
| 246 |
+
"loss": 0.3957,
|
| 247 |
+
"step": 3400
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.07,
|
| 251 |
+
"grad_norm": 1.1033278703689575,
|
| 252 |
+
"learning_rate": 0.00004687311083779784,
|
| 253 |
+
"loss": 0.3913,
|
| 254 |
+
"step": 3500
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.07,
|
| 258 |
+
"grad_norm": 1.0068538188934326,
|
| 259 |
+
"learning_rate": 0.00004676888119905777,
|
| 260 |
+
"loss": 0.3931,
|
| 261 |
+
"step": 3600
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.08,
|
| 265 |
+
"grad_norm": 1.0825397968292236,
|
| 266 |
+
"learning_rate": 0.000046664651560317695,
|
| 267 |
+
"loss": 0.4043,
|
| 268 |
+
"step": 3700
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.08,
|
| 272 |
+
"grad_norm": 1.1124091148376465,
|
| 273 |
+
"learning_rate": 0.000046560421921577626,
|
| 274 |
+
"loss": 0.3892,
|
| 275 |
+
"step": 3800
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.08,
|
| 279 |
+
"grad_norm": 0.8287498354911804,
|
| 280 |
+
"learning_rate": 0.00004645619228283755,
|
| 281 |
+
"loss": 0.3902,
|
| 282 |
+
"step": 3900
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.08,
|
| 286 |
+
"grad_norm": 1.3230834007263184,
|
| 287 |
+
"learning_rate": 0.00004635196264409748,
|
| 288 |
+
"loss": 0.3946,
|
| 289 |
+
"step": 4000
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.08,
|
| 293 |
+
"grad_norm": 1.1308443546295166,
|
| 294 |
+
"learning_rate": 0.00004624773300535741,
|
| 295 |
+
"loss": 0.388,
|
| 296 |
+
"step": 4100
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.09,
|
| 300 |
+
"grad_norm": 1.016345739364624,
|
| 301 |
+
"learning_rate": 0.00004614350336661734,
|
| 302 |
+
"loss": 0.3885,
|
| 303 |
+
"step": 4200
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.09,
|
| 307 |
+
"grad_norm": 0.7567281723022461,
|
| 308 |
+
"learning_rate": 0.00004603927372787726,
|
| 309 |
+
"loss": 0.3733,
|
| 310 |
+
"step": 4300
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.09,
|
| 314 |
+
"grad_norm": 1.0191072225570679,
|
| 315 |
+
"learning_rate": 0.000045935044089137194,
|
| 316 |
+
"loss": 0.3755,
|
| 317 |
+
"step": 4400
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.09,
|
| 321 |
+
"grad_norm": 0.9752495288848877,
|
| 322 |
+
"learning_rate": 0.00004583081445039712,
|
| 323 |
+
"loss": 0.3891,
|
| 324 |
+
"step": 4500
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.09,
|
| 328 |
+
"grad_norm": 1.063784122467041,
|
| 329 |
+
"learning_rate": 0.00004572658481165705,
|
| 330 |
+
"loss": 0.3502,
|
| 331 |
+
"step": 4600
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.1,
|
| 335 |
+
"grad_norm": 1.4277328252792358,
|
| 336 |
+
"learning_rate": 0.000045622355172916974,
|
| 337 |
+
"loss": 0.3832,
|
| 338 |
+
"step": 4700
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.1,
|
| 342 |
+
"grad_norm": 1.051148533821106,
|
| 343 |
+
"learning_rate": 0.000045518125534176906,
|
| 344 |
+
"loss": 0.3844,
|
| 345 |
+
"step": 4800
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.1,
|
| 349 |
+
"grad_norm": 1.0719451904296875,
|
| 350 |
+
"learning_rate": 0.00004541389589543683,
|
| 351 |
+
"loss": 0.3591,
|
| 352 |
+
"step": 4900
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.1,
|
| 356 |
+
"grad_norm": 1.0701349973678589,
|
| 357 |
+
"learning_rate": 0.00004530966625669676,
|
| 358 |
+
"loss": 0.3549,
|
| 359 |
+
"step": 5000
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.11,
|
| 363 |
+
"grad_norm": 0.9545413255691528,
|
| 364 |
+
"learning_rate": 0.000045205436617956686,
|
| 365 |
+
"loss": 0.3791,
|
| 366 |
+
"step": 5100
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.11,
|
| 370 |
+
"grad_norm": 1.2077405452728271,
|
| 371 |
+
"learning_rate": 0.00004510120697921662,
|
| 372 |
+
"loss": 0.3794,
|
| 373 |
+
"step": 5200
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.11,
|
| 377 |
+
"grad_norm": 1.0325422286987305,
|
| 378 |
+
"learning_rate": 0.00004499697734047654,
|
| 379 |
+
"loss": 0.3709,
|
| 380 |
+
"step": 5300
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.11,
|
| 384 |
+
"grad_norm": 1.231856107711792,
|
| 385 |
+
"learning_rate": 0.00004489274770173647,
|
| 386 |
+
"loss": 0.375,
|
| 387 |
+
"step": 5400
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.11,
|
| 391 |
+
"grad_norm": 1.0496745109558105,
|
| 392 |
+
"learning_rate": 0.0000447885180629964,
|
| 393 |
+
"loss": 0.3633,
|
| 394 |
+
"step": 5500
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.12,
|
| 398 |
+
"grad_norm": 0.9745790362358093,
|
| 399 |
+
"learning_rate": 0.00004468428842425633,
|
| 400 |
+
"loss": 0.3541,
|
| 401 |
+
"step": 5600
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.12,
|
| 405 |
+
"grad_norm": 0.8768660426139832,
|
| 406 |
+
"learning_rate": 0.000044580058785516246,
|
| 407 |
+
"loss": 0.354,
|
| 408 |
+
"step": 5700
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.12,
|
| 412 |
+
"grad_norm": 1.2748613357543945,
|
| 413 |
+
"learning_rate": 0.00004447582914677618,
|
| 414 |
+
"loss": 0.3463,
|
| 415 |
+
"step": 5800
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.12,
|
| 419 |
+
"grad_norm": 1.6291444301605225,
|
| 420 |
+
"learning_rate": 0.0000443715995080361,
|
| 421 |
+
"loss": 0.3483,
|
| 422 |
+
"step": 5900
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.12,
|
| 426 |
+
"grad_norm": 1.1184431314468384,
|
| 427 |
+
"learning_rate": 0.000044267369869296034,
|
| 428 |
+
"loss": 0.3616,
|
| 429 |
+
"step": 6000
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.13,
|
| 433 |
+
"grad_norm": 1.042864441871643,
|
| 434 |
+
"learning_rate": 0.00004416314023055596,
|
| 435 |
+
"loss": 0.3659,
|
| 436 |
+
"step": 6100
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.13,
|
| 440 |
+
"grad_norm": 1.2588552236557007,
|
| 441 |
+
"learning_rate": 0.00004405891059181589,
|
| 442 |
+
"loss": 0.3685,
|
| 443 |
+
"step": 6200
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.13,
|
| 447 |
+
"grad_norm": 0.9592663645744324,
|
| 448 |
+
"learning_rate": 0.000043954680953075814,
|
| 449 |
+
"loss": 0.3346,
|
| 450 |
+
"step": 6300
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.13,
|
| 454 |
+
"grad_norm": 1.0394409894943237,
|
| 455 |
+
"learning_rate": 0.000043850451314335745,
|
| 456 |
+
"loss": 0.3543,
|
| 457 |
+
"step": 6400
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.13,
|
| 461 |
+
"grad_norm": 1.0655689239501953,
|
| 462 |
+
"learning_rate": 0.00004374622167559567,
|
| 463 |
+
"loss": 0.3608,
|
| 464 |
+
"step": 6500
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.14,
|
| 468 |
+
"grad_norm": 1.0429097414016724,
|
| 469 |
+
"learning_rate": 0.0000436419920368556,
|
| 470 |
+
"loss": 0.3752,
|
| 471 |
+
"step": 6600
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.14,
|
| 475 |
+
"grad_norm": 1.105862021446228,
|
| 476 |
+
"learning_rate": 0.000043537762398115525,
|
| 477 |
+
"loss": 0.3656,
|
| 478 |
+
"step": 6700
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.14,
|
| 482 |
+
"grad_norm": 1.396264672279358,
|
| 483 |
+
"learning_rate": 0.00004343353275937546,
|
| 484 |
+
"loss": 0.3649,
|
| 485 |
+
"step": 6800
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.14,
|
| 489 |
+
"grad_norm": 1.1206328868865967,
|
| 490 |
+
"learning_rate": 0.00004332930312063538,
|
| 491 |
+
"loss": 0.3658,
|
| 492 |
+
"step": 6900
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.14,
|
| 496 |
+
"grad_norm": 1.1256790161132812,
|
| 497 |
+
"learning_rate": 0.00004322507348189531,
|
| 498 |
+
"loss": 0.3604,
|
| 499 |
+
"step": 7000
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.15,
|
| 503 |
+
"grad_norm": 1.0484185218811035,
|
| 504 |
+
"learning_rate": 0.00004312084384315524,
|
| 505 |
+
"loss": 0.3538,
|
| 506 |
+
"step": 7100
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.15,
|
| 510 |
+
"grad_norm": 1.1310747861862183,
|
| 511 |
+
"learning_rate": 0.00004301661420441517,
|
| 512 |
+
"loss": 0.3454,
|
| 513 |
+
"step": 7200
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.15,
|
| 517 |
+
"grad_norm": 1.3282368183135986,
|
| 518 |
+
"learning_rate": 0.00004291238456567509,
|
| 519 |
+
"loss": 0.3531,
|
| 520 |
+
"step": 7300
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.15,
|
| 524 |
+
"grad_norm": 1.1081411838531494,
|
| 525 |
+
"learning_rate": 0.000042808154926935024,
|
| 526 |
+
"loss": 0.3572,
|
| 527 |
+
"step": 7400
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.15,
|
| 531 |
+
"grad_norm": 1.0838216543197632,
|
| 532 |
+
"learning_rate": 0.00004270392528819495,
|
| 533 |
+
"loss": 0.3609,
|
| 534 |
+
"step": 7500
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.16,
|
| 538 |
+
"grad_norm": 1.0632649660110474,
|
| 539 |
+
"learning_rate": 0.00004259969564945488,
|
| 540 |
+
"loss": 0.3555,
|
| 541 |
+
"step": 7600
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.16,
|
| 545 |
+
"grad_norm": 1.2421514987945557,
|
| 546 |
+
"learning_rate": 0.000042495466010714804,
|
| 547 |
+
"loss": 0.3504,
|
| 548 |
+
"step": 7700
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.16,
|
| 552 |
+
"grad_norm": 1.0846999883651733,
|
| 553 |
+
"learning_rate": 0.000042391236371974736,
|
| 554 |
+
"loss": 0.3697,
|
| 555 |
+
"step": 7800
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.16,
|
| 559 |
+
"grad_norm": 1.0196987390518188,
|
| 560 |
+
"learning_rate": 0.00004228700673323466,
|
| 561 |
+
"loss": 0.331,
|
| 562 |
+
"step": 7900
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.17,
|
| 566 |
+
"grad_norm": 0.896594226360321,
|
| 567 |
+
"learning_rate": 0.00004218277709449459,
|
| 568 |
+
"loss": 0.3501,
|
| 569 |
+
"step": 8000
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.17,
|
| 573 |
+
"grad_norm": 1.0059070587158203,
|
| 574 |
+
"learning_rate": 0.000042078547455754516,
|
| 575 |
+
"loss": 0.3414,
|
| 576 |
+
"step": 8100
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.17,
|
| 580 |
+
"grad_norm": 0.9597504734992981,
|
| 581 |
+
"learning_rate": 0.00004197431781701445,
|
| 582 |
+
"loss": 0.3405,
|
| 583 |
+
"step": 8200
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.17,
|
| 587 |
+
"grad_norm": 0.9826390743255615,
|
| 588 |
+
"learning_rate": 0.00004187008817827437,
|
| 589 |
+
"loss": 0.3444,
|
| 590 |
+
"step": 8300
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.17,
|
| 594 |
+
"grad_norm": 1.1618746519088745,
|
| 595 |
+
"learning_rate": 0.0000417658585395343,
|
| 596 |
+
"loss": 0.3428,
|
| 597 |
+
"step": 8400
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.18,
|
| 601 |
+
"grad_norm": 1.3454729318618774,
|
| 602 |
+
"learning_rate": 0.00004166162890079423,
|
| 603 |
+
"loss": 0.3546,
|
| 604 |
+
"step": 8500
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.18,
|
| 608 |
+
"grad_norm": 1.0757421255111694,
|
| 609 |
+
"learning_rate": 0.00004155739926205416,
|
| 610 |
+
"loss": 0.3409,
|
| 611 |
+
"step": 8600
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.18,
|
| 615 |
+
"grad_norm": 1.1405227184295654,
|
| 616 |
+
"learning_rate": 0.000041453169623314083,
|
| 617 |
+
"loss": 0.367,
|
| 618 |
+
"step": 8700
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.18,
|
| 622 |
+
"grad_norm": 1.336688756942749,
|
| 623 |
+
"learning_rate": 0.000041348939984574015,
|
| 624 |
+
"loss": 0.353,
|
| 625 |
+
"step": 8800
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.18,
|
| 629 |
+
"grad_norm": 1.0525643825531006,
|
| 630 |
+
"learning_rate": 0.00004124471034583394,
|
| 631 |
+
"loss": 0.3394,
|
| 632 |
+
"step": 8900
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.19,
|
| 636 |
+
"grad_norm": 0.9103542566299438,
|
| 637 |
+
"learning_rate": 0.00004114048070709387,
|
| 638 |
+
"loss": 0.3542,
|
| 639 |
+
"step": 9000
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.19,
|
| 643 |
+
"grad_norm": 1.1108696460723877,
|
| 644 |
+
"learning_rate": 0.000041036251068353795,
|
| 645 |
+
"loss": 0.3468,
|
| 646 |
+
"step": 9100
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.19,
|
| 650 |
+
"grad_norm": 0.9561033248901367,
|
| 651 |
+
"learning_rate": 0.000040932021429613726,
|
| 652 |
+
"loss": 0.3234,
|
| 653 |
+
"step": 9200
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.19,
|
| 657 |
+
"grad_norm": 1.0222822427749634,
|
| 658 |
+
"learning_rate": 0.00004082779179087365,
|
| 659 |
+
"loss": 0.3557,
|
| 660 |
+
"step": 9300
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.19,
|
| 664 |
+
"grad_norm": 0.7020508050918579,
|
| 665 |
+
"learning_rate": 0.00004072356215213358,
|
| 666 |
+
"loss": 0.3387,
|
| 667 |
+
"step": 9400
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.2,
|
| 671 |
+
"grad_norm": 1.1984186172485352,
|
| 672 |
+
"learning_rate": 0.00004061933251339351,
|
| 673 |
+
"loss": 0.3396,
|
| 674 |
+
"step": 9500
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.2,
|
| 678 |
+
"grad_norm": 1.0309916734695435,
|
| 679 |
+
"learning_rate": 0.00004051510287465344,
|
| 680 |
+
"loss": 0.3361,
|
| 681 |
+
"step": 9600
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.2,
|
| 685 |
+
"grad_norm": 0.9544962048530579,
|
| 686 |
+
"learning_rate": 0.00004041087323591336,
|
| 687 |
+
"loss": 0.3198,
|
| 688 |
+
"step": 9700
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.2,
|
| 692 |
+
"grad_norm": 0.924387514591217,
|
| 693 |
+
"learning_rate": 0.000040306643597173294,
|
| 694 |
+
"loss": 0.346,
|
| 695 |
+
"step": 9800
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.2,
|
| 699 |
+
"grad_norm": 0.9332679510116577,
|
| 700 |
+
"learning_rate": 0.00004020241395843322,
|
| 701 |
+
"loss": 0.3583,
|
| 702 |
+
"step": 9900
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.21,
|
| 706 |
+
"grad_norm": 0.8733212351799011,
|
| 707 |
+
"learning_rate": 0.00004009818431969315,
|
| 708 |
+
"loss": 0.3258,
|
| 709 |
+
"step": 10000
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.21,
|
| 713 |
+
"grad_norm": 1.1735975742340088,
|
| 714 |
+
"learning_rate": 0.000039993954680953074,
|
| 715 |
+
"loss": 0.3319,
|
| 716 |
+
"step": 10100
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.21,
|
| 720 |
+
"grad_norm": 1.1871587038040161,
|
| 721 |
+
"learning_rate": 0.000039889725042213005,
|
| 722 |
+
"loss": 0.3323,
|
| 723 |
+
"step": 10200
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.21,
|
| 727 |
+
"grad_norm": 1.0501062870025635,
|
| 728 |
+
"learning_rate": 0.00003978549540347294,
|
| 729 |
+
"loss": 0.3501,
|
| 730 |
+
"step": 10300
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.21,
|
| 734 |
+
"grad_norm": 1.6133193969726562,
|
| 735 |
+
"learning_rate": 0.00003968126576473286,
|
| 736 |
+
"loss": 0.3353,
|
| 737 |
+
"step": 10400
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.22,
|
| 741 |
+
"grad_norm": 1.0860315561294556,
|
| 742 |
+
"learning_rate": 0.00003957703612599279,
|
| 743 |
+
"loss": 0.3247,
|
| 744 |
+
"step": 10500
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.22,
|
| 748 |
+
"grad_norm": 0.9817527532577515,
|
| 749 |
+
"learning_rate": 0.00003947280648725272,
|
| 750 |
+
"loss": 0.3486,
|
| 751 |
+
"step": 10600
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.22,
|
| 755 |
+
"grad_norm": 1.185621738433838,
|
| 756 |
+
"learning_rate": 0.00003936857684851265,
|
| 757 |
+
"loss": 0.3344,
|
| 758 |
+
"step": 10700
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.22,
|
| 762 |
+
"grad_norm": 1.211653709411621,
|
| 763 |
+
"learning_rate": 0.00003926434720977257,
|
| 764 |
+
"loss": 0.3308,
|
| 765 |
+
"step": 10800
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.22,
|
| 769 |
+
"grad_norm": 1.1859573125839233,
|
| 770 |
+
"learning_rate": 0.000039160117571032504,
|
| 771 |
+
"loss": 0.3462,
|
| 772 |
+
"step": 10900
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.23,
|
| 776 |
+
"grad_norm": 0.9363047480583191,
|
| 777 |
+
"learning_rate": 0.00003905588793229243,
|
| 778 |
+
"loss": 0.3331,
|
| 779 |
+
"step": 11000
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.23,
|
| 783 |
+
"grad_norm": 0.8858780860900879,
|
| 784 |
+
"learning_rate": 0.00003895165829355236,
|
| 785 |
+
"loss": 0.3501,
|
| 786 |
+
"step": 11100
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.23,
|
| 790 |
+
"grad_norm": 1.187888264656067,
|
| 791 |
+
"learning_rate": 0.000038847428654812284,
|
| 792 |
+
"loss": 0.3202,
|
| 793 |
+
"step": 11200
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.23,
|
| 797 |
+
"grad_norm": 1.33687424659729,
|
| 798 |
+
"learning_rate": 0.000038743199016072216,
|
| 799 |
+
"loss": 0.3488,
|
| 800 |
+
"step": 11300
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.24,
|
| 804 |
+
"grad_norm": 1.230508804321289,
|
| 805 |
+
"learning_rate": 0.00003863896937733214,
|
| 806 |
+
"loss": 0.322,
|
| 807 |
+
"step": 11400
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.24,
|
| 811 |
+
"grad_norm": 1.2062541246414185,
|
| 812 |
+
"learning_rate": 0.00003853473973859207,
|
| 813 |
+
"loss": 0.3446,
|
| 814 |
+
"step": 11500
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.24,
|
| 818 |
+
"grad_norm": 0.9526596069335938,
|
| 819 |
+
"learning_rate": 0.000038430510099851996,
|
| 820 |
+
"loss": 0.3167,
|
| 821 |
+
"step": 11600
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.24,
|
| 825 |
+
"grad_norm": 0.8020007014274597,
|
| 826 |
+
"learning_rate": 0.00003832628046111193,
|
| 827 |
+
"loss": 0.3289,
|
| 828 |
+
"step": 11700
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.24,
|
| 832 |
+
"grad_norm": 0.8741321563720703,
|
| 833 |
+
"learning_rate": 0.00003822205082237185,
|
| 834 |
+
"loss": 0.3307,
|
| 835 |
+
"step": 11800
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.25,
|
| 839 |
+
"grad_norm": 0.8475280404090881,
|
| 840 |
+
"learning_rate": 0.00003811782118363178,
|
| 841 |
+
"loss": 0.3116,
|
| 842 |
+
"step": 11900
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.25,
|
| 846 |
+
"grad_norm": 0.8570387959480286,
|
| 847 |
+
"learning_rate": 0.00003801359154489171,
|
| 848 |
+
"loss": 0.3345,
|
| 849 |
+
"step": 12000
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.25,
|
| 853 |
+
"grad_norm": 1.237730860710144,
|
| 854 |
+
"learning_rate": 0.00003790936190615164,
|
| 855 |
+
"loss": 0.3457,
|
| 856 |
+
"step": 12100
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.25,
|
| 860 |
+
"grad_norm": 1.1261372566223145,
|
| 861 |
+
"learning_rate": 0.00003780513226741156,
|
| 862 |
+
"loss": 0.3261,
|
| 863 |
+
"step": 12200
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.25,
|
| 867 |
+
"grad_norm": 0.7940207719802856,
|
| 868 |
+
"learning_rate": 0.000037700902628671495,
|
| 869 |
+
"loss": 0.3272,
|
| 870 |
+
"step": 12300
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.26,
|
| 874 |
+
"grad_norm": 1.0150914192199707,
|
| 875 |
+
"learning_rate": 0.00003759667298993142,
|
| 876 |
+
"loss": 0.3338,
|
| 877 |
+
"step": 12400
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.26,
|
| 881 |
+
"grad_norm": 0.8289426565170288,
|
| 882 |
+
"learning_rate": 0.00003749244335119135,
|
| 883 |
+
"loss": 0.3485,
|
| 884 |
+
"step": 12500
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.26,
|
| 888 |
+
"grad_norm": 1.11201012134552,
|
| 889 |
+
"learning_rate": 0.000037388213712451275,
|
| 890 |
+
"loss": 0.321,
|
| 891 |
+
"step": 12600
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.26,
|
| 895 |
+
"grad_norm": 0.7759777903556824,
|
| 896 |
+
"learning_rate": 0.000037283984073711206,
|
| 897 |
+
"loss": 0.336,
|
| 898 |
+
"step": 12700
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.26,
|
| 902 |
+
"grad_norm": 1.2184884548187256,
|
| 903 |
+
"learning_rate": 0.00003717975443497113,
|
| 904 |
+
"loss": 0.3029,
|
| 905 |
+
"step": 12800
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.27,
|
| 909 |
+
"grad_norm": 0.7660321593284607,
|
| 910 |
+
"learning_rate": 0.00003707552479623106,
|
| 911 |
+
"loss": 0.336,
|
| 912 |
+
"step": 12900
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.27,
|
| 916 |
+
"grad_norm": 0.8834217190742493,
|
| 917 |
+
"learning_rate": 0.000036971295157490987,
|
| 918 |
+
"loss": 0.3184,
|
| 919 |
+
"step": 13000
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.27,
|
| 923 |
+
"grad_norm": 1.1337840557098389,
|
| 924 |
+
"learning_rate": 0.00003686706551875092,
|
| 925 |
+
"loss": 0.3216,
|
| 926 |
+
"step": 13100
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.27,
|
| 930 |
+
"grad_norm": 1.0339350700378418,
|
| 931 |
+
"learning_rate": 0.00003676283588001084,
|
| 932 |
+
"loss": 0.3162,
|
| 933 |
+
"step": 13200
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.27,
|
| 937 |
+
"grad_norm": 1.222301721572876,
|
| 938 |
+
"learning_rate": 0.000036658606241270774,
|
| 939 |
+
"loss": 0.3344,
|
| 940 |
+
"step": 13300
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.28,
|
| 944 |
+
"grad_norm": 1.290209412574768,
|
| 945 |
+
"learning_rate": 0.0000365543766025307,
|
| 946 |
+
"loss": 0.3212,
|
| 947 |
+
"step": 13400
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.28,
|
| 951 |
+
"grad_norm": 1.372733235359192,
|
| 952 |
+
"learning_rate": 0.00003645014696379063,
|
| 953 |
+
"loss": 0.3375,
|
| 954 |
+
"step": 13500
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.28,
|
| 958 |
+
"grad_norm": 0.7824010252952576,
|
| 959 |
+
"learning_rate": 0.000036345917325050554,
|
| 960 |
+
"loss": 0.3393,
|
| 961 |
+
"step": 13600
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.28,
|
| 965 |
+
"grad_norm": 1.637102484703064,
|
| 966 |
+
"learning_rate": 0.000036241687686310485,
|
| 967 |
+
"loss": 0.3318,
|
| 968 |
+
"step": 13700
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.28,
|
| 972 |
+
"grad_norm": 0.9548413753509521,
|
| 973 |
+
"learning_rate": 0.00003613745804757041,
|
| 974 |
+
"loss": 0.3041,
|
| 975 |
+
"step": 13800
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.29,
|
| 979 |
+
"grad_norm": 1.3793456554412842,
|
| 980 |
+
"learning_rate": 0.00003603322840883034,
|
| 981 |
+
"loss": 0.3208,
|
| 982 |
+
"step": 13900
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.29,
|
| 986 |
+
"grad_norm": 1.323983073234558,
|
| 987 |
+
"learning_rate": 0.000035928998770090266,
|
| 988 |
+
"loss": 0.3077,
|
| 989 |
+
"step": 14000
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.29,
|
| 993 |
+
"grad_norm": 0.8947316408157349,
|
| 994 |
+
"learning_rate": 0.0000358247691313502,
|
| 995 |
+
"loss": 0.3081,
|
| 996 |
+
"step": 14100
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.29,
|
| 1000 |
+
"grad_norm": 1.2341340780258179,
|
| 1001 |
+
"learning_rate": 0.00003572053949261012,
|
| 1002 |
+
"loss": 0.3243,
|
| 1003 |
+
"step": 14200
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.3,
|
| 1007 |
+
"grad_norm": 1.1560657024383545,
|
| 1008 |
+
"learning_rate": 0.00003561630985387005,
|
| 1009 |
+
"loss": 0.3045,
|
| 1010 |
+
"step": 14300
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.3,
|
| 1014 |
+
"grad_norm": 1.0063599348068237,
|
| 1015 |
+
"learning_rate": 0.00003551208021512998,
|
| 1016 |
+
"loss": 0.2984,
|
| 1017 |
+
"step": 14400
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.3,
|
| 1021 |
+
"grad_norm": 0.9587939977645874,
|
| 1022 |
+
"learning_rate": 0.00003540785057638991,
|
| 1023 |
+
"loss": 0.3215,
|
| 1024 |
+
"step": 14500
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.3,
|
| 1028 |
+
"grad_norm": 0.9009751081466675,
|
| 1029 |
+
"learning_rate": 0.00003530362093764983,
|
| 1030 |
+
"loss": 0.3257,
|
| 1031 |
+
"step": 14600
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.3,
|
| 1035 |
+
"grad_norm": 1.1289176940917969,
|
| 1036 |
+
"learning_rate": 0.000035199391298909764,
|
| 1037 |
+
"loss": 0.3151,
|
| 1038 |
+
"step": 14700
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.31,
|
| 1042 |
+
"grad_norm": 1.0544753074645996,
|
| 1043 |
+
"learning_rate": 0.00003509516166016969,
|
| 1044 |
+
"loss": 0.3123,
|
| 1045 |
+
"step": 14800
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.31,
|
| 1049 |
+
"grad_norm": 1.6340482234954834,
|
| 1050 |
+
"learning_rate": 0.00003499093202142962,
|
| 1051 |
+
"loss": 0.3238,
|
| 1052 |
+
"step": 14900
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.31,
|
| 1056 |
+
"grad_norm": 0.8671633005142212,
|
| 1057 |
+
"learning_rate": 0.000034886702382689545,
|
| 1058 |
+
"loss": 0.3174,
|
| 1059 |
+
"step": 15000
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.31,
|
| 1063 |
+
"grad_norm": 0.8871846199035645,
|
| 1064 |
+
"learning_rate": 0.000034782472743949476,
|
| 1065 |
+
"loss": 0.3045,
|
| 1066 |
+
"step": 15100
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.31,
|
| 1070 |
+
"grad_norm": 0.9014525413513184,
|
| 1071 |
+
"learning_rate": 0.0000346782431052094,
|
| 1072 |
+
"loss": 0.3335,
|
| 1073 |
+
"step": 15200
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.32,
|
| 1077 |
+
"grad_norm": 0.819088876247406,
|
| 1078 |
+
"learning_rate": 0.00003457401346646933,
|
| 1079 |
+
"loss": 0.3222,
|
| 1080 |
+
"step": 15300
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 0.32,
|
| 1084 |
+
"grad_norm": 1.1166553497314453,
|
| 1085 |
+
"learning_rate": 0.000034469783827729256,
|
| 1086 |
+
"loss": 0.3303,
|
| 1087 |
+
"step": 15400
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.32,
|
| 1091 |
+
"grad_norm": 1.0952930450439453,
|
| 1092 |
+
"learning_rate": 0.00003436555418898919,
|
| 1093 |
+
"loss": 0.3095,
|
| 1094 |
+
"step": 15500
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.32,
|
| 1098 |
+
"grad_norm": 1.186812400817871,
|
| 1099 |
+
"learning_rate": 0.00003426132455024911,
|
| 1100 |
+
"loss": 0.3146,
|
| 1101 |
+
"step": 15600
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.32,
|
| 1105 |
+
"grad_norm": 0.9790433049201965,
|
| 1106 |
+
"learning_rate": 0.00003415709491150904,
|
| 1107 |
+
"loss": 0.321,
|
| 1108 |
+
"step": 15700
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 0.33,
|
| 1112 |
+
"grad_norm": 0.9488071203231812,
|
| 1113 |
+
"learning_rate": 0.00003405286527276897,
|
| 1114 |
+
"loss": 0.321,
|
| 1115 |
+
"step": 15800
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.33,
|
| 1119 |
+
"grad_norm": 0.9488551616668701,
|
| 1120 |
+
"learning_rate": 0.00003394863563402889,
|
| 1121 |
+
"loss": 0.2923,
|
| 1122 |
+
"step": 15900
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 0.33,
|
| 1126 |
+
"grad_norm": 0.9801424145698547,
|
| 1127 |
+
"learning_rate": 0.00003384440599528882,
|
| 1128 |
+
"loss": 0.3131,
|
| 1129 |
+
"step": 16000
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.33,
|
| 1133 |
+
"grad_norm": 1.5585353374481201,
|
| 1134 |
+
"learning_rate": 0.00003374017635654875,
|
| 1135 |
+
"loss": 0.3274,
|
| 1136 |
+
"step": 16100
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.33,
|
| 1140 |
+
"grad_norm": 1.0194026231765747,
|
| 1141 |
+
"learning_rate": 0.00003363594671780867,
|
| 1142 |
+
"loss": 0.3294,
|
| 1143 |
+
"step": 16200
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.34,
|
| 1147 |
+
"grad_norm": 0.89570552110672,
|
| 1148 |
+
"learning_rate": 0.000033531717079068604,
|
| 1149 |
+
"loss": 0.3041,
|
| 1150 |
+
"step": 16300
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 0.34,
|
| 1154 |
+
"grad_norm": 1.361720085144043,
|
| 1155 |
+
"learning_rate": 0.00003342748744032853,
|
| 1156 |
+
"loss": 0.3001,
|
| 1157 |
+
"step": 16400
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.34,
|
| 1161 |
+
"grad_norm": 0.6049641370773315,
|
| 1162 |
+
"learning_rate": 0.00003332325780158846,
|
| 1163 |
+
"loss": 0.2993,
|
| 1164 |
+
"step": 16500
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 0.34,
|
| 1168 |
+
"grad_norm": 1.1714413166046143,
|
| 1169 |
+
"learning_rate": 0.000033219028162848384,
|
| 1170 |
+
"loss": 0.3252,
|
| 1171 |
+
"step": 16600
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.34,
|
| 1175 |
+
"grad_norm": 0.6133936643600464,
|
| 1176 |
+
"learning_rate": 0.000033114798524108315,
|
| 1177 |
+
"loss": 0.3162,
|
| 1178 |
+
"step": 16700
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.35,
|
| 1182 |
+
"grad_norm": 0.9461352229118347,
|
| 1183 |
+
"learning_rate": 0.00003301056888536824,
|
| 1184 |
+
"loss": 0.3,
|
| 1185 |
+
"step": 16800
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.35,
|
| 1189 |
+
"grad_norm": 1.1674548387527466,
|
| 1190 |
+
"learning_rate": 0.00003290633924662817,
|
| 1191 |
+
"loss": 0.3053,
|
| 1192 |
+
"step": 16900
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 0.35,
|
| 1196 |
+
"grad_norm": 0.9445935487747192,
|
| 1197 |
+
"learning_rate": 0.000032802109607888096,
|
| 1198 |
+
"loss": 0.319,
|
| 1199 |
+
"step": 17000
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.35,
|
| 1203 |
+
"grad_norm": 0.7129838466644287,
|
| 1204 |
+
"learning_rate": 0.00003269787996914803,
|
| 1205 |
+
"loss": 0.3095,
|
| 1206 |
+
"step": 17100
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 0.35,
|
| 1210 |
+
"grad_norm": 0.9326237440109253,
|
| 1211 |
+
"learning_rate": 0.00003259365033040795,
|
| 1212 |
+
"loss": 0.3174,
|
| 1213 |
+
"step": 17200
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.36,
|
| 1217 |
+
"grad_norm": 0.8677362203598022,
|
| 1218 |
+
"learning_rate": 0.00003248942069166788,
|
| 1219 |
+
"loss": 0.3056,
|
| 1220 |
+
"step": 17300
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 0.36,
|
| 1224 |
+
"grad_norm": 0.7561855316162109,
|
| 1225 |
+
"learning_rate": 0.00003238519105292781,
|
| 1226 |
+
"loss": 0.2939,
|
| 1227 |
+
"step": 17400
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.36,
|
| 1231 |
+
"grad_norm": 1.0813933610916138,
|
| 1232 |
+
"learning_rate": 0.00003228096141418774,
|
| 1233 |
+
"loss": 0.296,
|
| 1234 |
+
"step": 17500
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 0.36,
|
| 1238 |
+
"grad_norm": 1.1002167463302612,
|
| 1239 |
+
"learning_rate": 0.00003217673177544766,
|
| 1240 |
+
"loss": 0.3142,
|
| 1241 |
+
"step": 17600
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.37,
|
| 1245 |
+
"grad_norm": 0.709709882736206,
|
| 1246 |
+
"learning_rate": 0.000032072502136707594,
|
| 1247 |
+
"loss": 0.2931,
|
| 1248 |
+
"step": 17700
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 0.37,
|
| 1252 |
+
"grad_norm": 0.699057936668396,
|
| 1253 |
+
"learning_rate": 0.00003196827249796752,
|
| 1254 |
+
"loss": 0.3009,
|
| 1255 |
+
"step": 17800
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 0.37,
|
| 1259 |
+
"grad_norm": 1.4942712783813477,
|
| 1260 |
+
"learning_rate": 0.00003186404285922745,
|
| 1261 |
+
"loss": 0.3203,
|
| 1262 |
+
"step": 17900
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 0.37,
|
| 1266 |
+
"grad_norm": 1.069948434829712,
|
| 1267 |
+
"learning_rate": 0.000031759813220487375,
|
| 1268 |
+
"loss": 0.3118,
|
| 1269 |
+
"step": 18000
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 0.37,
|
| 1273 |
+
"grad_norm": 0.9980233311653137,
|
| 1274 |
+
"learning_rate": 0.000031655583581747306,
|
| 1275 |
+
"loss": 0.2937,
|
| 1276 |
+
"step": 18100
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 0.38,
|
| 1280 |
+
"grad_norm": 0.743611752986908,
|
| 1281 |
+
"learning_rate": 0.00003155135394300723,
|
| 1282 |
+
"loss": 0.3064,
|
| 1283 |
+
"step": 18200
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 0.38,
|
| 1287 |
+
"grad_norm": 0.8888816237449646,
|
| 1288 |
+
"learning_rate": 0.00003144712430426716,
|
| 1289 |
+
"loss": 0.309,
|
| 1290 |
+
"step": 18300
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 0.38,
|
| 1294 |
+
"grad_norm": 0.9147290587425232,
|
| 1295 |
+
"learning_rate": 0.000031342894665527086,
|
| 1296 |
+
"loss": 0.3104,
|
| 1297 |
+
"step": 18400
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 0.38,
|
| 1301 |
+
"grad_norm": 1.191470742225647,
|
| 1302 |
+
"learning_rate": 0.00003123866502678702,
|
| 1303 |
+
"loss": 0.307,
|
| 1304 |
+
"step": 18500
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 0.38,
|
| 1308 |
+
"grad_norm": 0.8025168776512146,
|
| 1309 |
+
"learning_rate": 0.00003113443538804694,
|
| 1310 |
+
"loss": 0.2981,
|
| 1311 |
+
"step": 18600
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 0.39,
|
| 1315 |
+
"grad_norm": 1.0088508129119873,
|
| 1316 |
+
"learning_rate": 0.000031030205749306873,
|
| 1317 |
+
"loss": 0.2946,
|
| 1318 |
+
"step": 18700
|
| 1319 |
+
},
|
| 1320 |
+
{
|
| 1321 |
+
"epoch": 0.39,
|
| 1322 |
+
"grad_norm": 0.9163653254508972,
|
| 1323 |
+
"learning_rate": 0.0000309259761105668,
|
| 1324 |
+
"loss": 0.314,
|
| 1325 |
+
"step": 18800
|
| 1326 |
+
},
|
| 1327 |
+
{
|
| 1328 |
+
"epoch": 0.39,
|
| 1329 |
+
"grad_norm": 0.7615305185317993,
|
| 1330 |
+
"learning_rate": 0.00003082174647182673,
|
| 1331 |
+
"loss": 0.2987,
|
| 1332 |
+
"step": 18900
|
| 1333 |
+
},
|
| 1334 |
+
{
|
| 1335 |
+
"epoch": 0.39,
|
| 1336 |
+
"grad_norm": 1.1720337867736816,
|
| 1337 |
+
"learning_rate": 0.000030717516833086654,
|
| 1338 |
+
"loss": 0.284,
|
| 1339 |
+
"step": 19000
|
| 1340 |
+
},
|
| 1341 |
+
{
|
| 1342 |
+
"epoch": 0.39,
|
| 1343 |
+
"grad_norm": 0.9668058156967163,
|
| 1344 |
+
"learning_rate": 0.000030613287194346585,
|
| 1345 |
+
"loss": 0.2961,
|
| 1346 |
+
"step": 19100
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 0.4,
|
| 1350 |
+
"grad_norm": 1.0331960916519165,
|
| 1351 |
+
"learning_rate": 0.000030509057555606513,
|
| 1352 |
+
"loss": 0.3126,
|
| 1353 |
+
"step": 19200
|
| 1354 |
+
},
|
| 1355 |
+
{
|
| 1356 |
+
"epoch": 0.4,
|
| 1357 |
+
"grad_norm": 1.011672854423523,
|
| 1358 |
+
"learning_rate": 0.00003040482791686644,
|
| 1359 |
+
"loss": 0.2909,
|
| 1360 |
+
"step": 19300
|
| 1361 |
+
},
|
| 1362 |
+
{
|
| 1363 |
+
"epoch": 0.4,
|
| 1364 |
+
"grad_norm": 0.7572203874588013,
|
| 1365 |
+
"learning_rate": 0.00003030059827812637,
|
| 1366 |
+
"loss": 0.3003,
|
| 1367 |
+
"step": 19400
|
| 1368 |
+
},
|
| 1369 |
+
{
|
| 1370 |
+
"epoch": 0.4,
|
| 1371 |
+
"grad_norm": 1.2314035892486572,
|
| 1372 |
+
"learning_rate": 0.000030196368639386297,
|
| 1373 |
+
"loss": 0.2967,
|
| 1374 |
+
"step": 19500
|
| 1375 |
+
},
|
| 1376 |
+
{
|
| 1377 |
+
"epoch": 0.4,
|
| 1378 |
+
"grad_norm": 0.7441611886024475,
|
| 1379 |
+
"learning_rate": 0.000030092139000646225,
|
| 1380 |
+
"loss": 0.3081,
|
| 1381 |
+
"step": 19600
|
| 1382 |
+
},
|
| 1383 |
+
{
|
| 1384 |
+
"epoch": 0.41,
|
| 1385 |
+
"grad_norm": 1.1014299392700195,
|
| 1386 |
+
"learning_rate": 0.000029987909361906152,
|
| 1387 |
+
"loss": 0.3076,
|
| 1388 |
+
"step": 19700
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 0.41,
|
| 1392 |
+
"grad_norm": 0.8321337699890137,
|
| 1393 |
+
"learning_rate": 0.00002988367972316608,
|
| 1394 |
+
"loss": 0.3,
|
| 1395 |
+
"step": 19800
|
| 1396 |
+
},
|
| 1397 |
+
{
|
| 1398 |
+
"epoch": 0.41,
|
| 1399 |
+
"grad_norm": 1.0691934823989868,
|
| 1400 |
+
"learning_rate": 0.000029779450084426008,
|
| 1401 |
+
"loss": 0.3113,
|
| 1402 |
+
"step": 19900
|
| 1403 |
+
},
|
| 1404 |
+
{
|
| 1405 |
+
"epoch": 0.41,
|
| 1406 |
+
"grad_norm": 0.8844496607780457,
|
| 1407 |
+
"learning_rate": 0.000029675220445685936,
|
| 1408 |
+
"loss": 0.3135,
|
| 1409 |
+
"step": 20000
|
| 1410 |
+
},
|
| 1411 |
+
{
|
| 1412 |
+
"epoch": 0.41,
|
| 1413 |
+
"grad_norm": 0.9948704242706299,
|
| 1414 |
+
"learning_rate": 0.000029570990806945864,
|
| 1415 |
+
"loss": 0.2905,
|
| 1416 |
+
"step": 20100
|
| 1417 |
+
},
|
| 1418 |
+
{
|
| 1419 |
+
"epoch": 0.42,
|
| 1420 |
+
"grad_norm": 0.7587433457374573,
|
| 1421 |
+
"learning_rate": 0.000029466761168205792,
|
| 1422 |
+
"loss": 0.3187,
|
| 1423 |
+
"step": 20200
|
| 1424 |
+
},
|
| 1425 |
+
{
|
| 1426 |
+
"epoch": 0.42,
|
| 1427 |
+
"grad_norm": 0.8162456750869751,
|
| 1428 |
+
"learning_rate": 0.00002936253152946572,
|
| 1429 |
+
"loss": 0.3101,
|
| 1430 |
+
"step": 20300
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 0.42,
|
| 1434 |
+
"grad_norm": 0.5907912850379944,
|
| 1435 |
+
"learning_rate": 0.000029258301890725648,
|
| 1436 |
+
"loss": 0.2926,
|
| 1437 |
+
"step": 20400
|
| 1438 |
+
},
|
| 1439 |
+
{
|
| 1440 |
+
"epoch": 0.42,
|
| 1441 |
+
"grad_norm": 0.9535139799118042,
|
| 1442 |
+
"learning_rate": 0.000029154072251985576,
|
| 1443 |
+
"loss": 0.3237,
|
| 1444 |
+
"step": 20500
|
| 1445 |
+
},
|
| 1446 |
+
{
|
| 1447 |
+
"epoch": 0.42,
|
| 1448 |
+
"grad_norm": 0.9831104874610901,
|
| 1449 |
+
"learning_rate": 0.000029049842613245504,
|
| 1450 |
+
"loss": 0.2922,
|
| 1451 |
+
"step": 20600
|
| 1452 |
+
},
|
| 1453 |
+
{
|
| 1454 |
+
"epoch": 0.43,
|
| 1455 |
+
"grad_norm": 0.8171008229255676,
|
| 1456 |
+
"learning_rate": 0.00002894561297450543,
|
| 1457 |
+
"loss": 0.3127,
|
| 1458 |
+
"step": 20700
|
| 1459 |
+
},
|
| 1460 |
+
{
|
| 1461 |
+
"epoch": 0.43,
|
| 1462 |
+
"grad_norm": 0.9013519287109375,
|
| 1463 |
+
"learning_rate": 0.00002884138333576536,
|
| 1464 |
+
"loss": 0.2974,
|
| 1465 |
+
"step": 20800
|
| 1466 |
+
},
|
| 1467 |
+
{
|
| 1468 |
+
"epoch": 0.43,
|
| 1469 |
+
"grad_norm": 1.0881197452545166,
|
| 1470 |
+
"learning_rate": 0.000028737153697025287,
|
| 1471 |
+
"loss": 0.3033,
|
| 1472 |
+
"step": 20900
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 0.43,
|
| 1476 |
+
"grad_norm": 1.0165661573410034,
|
| 1477 |
+
"learning_rate": 0.000028632924058285215,
|
| 1478 |
+
"loss": 0.2969,
|
| 1479 |
+
"step": 21000
|
| 1480 |
+
},
|
| 1481 |
+
{
|
| 1482 |
+
"epoch": 0.44,
|
| 1483 |
+
"grad_norm": 0.9244837760925293,
|
| 1484 |
+
"learning_rate": 0.000028528694419545143,
|
| 1485 |
+
"loss": 0.3074,
|
| 1486 |
+
"step": 21100
|
| 1487 |
+
},
|
| 1488 |
+
{
|
| 1489 |
+
"epoch": 0.44,
|
| 1490 |
+
"grad_norm": 0.8180138468742371,
|
| 1491 |
+
"learning_rate": 0.00002842446478080507,
|
| 1492 |
+
"loss": 0.3022,
|
| 1493 |
+
"step": 21200
|
| 1494 |
+
},
|
| 1495 |
+
{
|
| 1496 |
+
"epoch": 0.44,
|
| 1497 |
+
"grad_norm": 0.8913204073905945,
|
| 1498 |
+
"learning_rate": 0.000028320235142065,
|
| 1499 |
+
"loss": 0.3041,
|
| 1500 |
+
"step": 21300
|
| 1501 |
+
},
|
| 1502 |
+
{
|
| 1503 |
+
"epoch": 0.44,
|
| 1504 |
+
"grad_norm": 0.944990336894989,
|
| 1505 |
+
"learning_rate": 0.000028216005503324927,
|
| 1506 |
+
"loss": 0.3034,
|
| 1507 |
+
"step": 21400
|
| 1508 |
+
},
|
| 1509 |
+
{
|
| 1510 |
+
"epoch": 0.44,
|
| 1511 |
+
"grad_norm": 0.6920465230941772,
|
| 1512 |
+
"learning_rate": 0.000028111775864584855,
|
| 1513 |
+
"loss": 0.287,
|
| 1514 |
+
"step": 21500
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 0.45,
|
| 1518 |
+
"grad_norm": 0.9530170559883118,
|
| 1519 |
+
"learning_rate": 0.000028007546225844783,
|
| 1520 |
+
"loss": 0.2924,
|
| 1521 |
+
"step": 21600
|
| 1522 |
+
},
|
| 1523 |
+
{
|
| 1524 |
+
"epoch": 0.45,
|
| 1525 |
+
"grad_norm": 1.0395070314407349,
|
| 1526 |
+
"learning_rate": 0.00002790331658710471,
|
| 1527 |
+
"loss": 0.2976,
|
| 1528 |
+
"step": 21700
|
| 1529 |
+
},
|
| 1530 |
+
{
|
| 1531 |
+
"epoch": 0.45,
|
| 1532 |
+
"grad_norm": 1.0013209581375122,
|
| 1533 |
+
"learning_rate": 0.00002779908694836464,
|
| 1534 |
+
"loss": 0.2765,
|
| 1535 |
+
"step": 21800
|
| 1536 |
+
},
|
| 1537 |
+
{
|
| 1538 |
+
"epoch": 0.45,
|
| 1539 |
+
"grad_norm": 0.6239937543869019,
|
| 1540 |
+
"learning_rate": 0.000027694857309624566,
|
| 1541 |
+
"loss": 0.3001,
|
| 1542 |
+
"step": 21900
|
| 1543 |
+
},
|
| 1544 |
+
{
|
| 1545 |
+
"epoch": 0.45,
|
| 1546 |
+
"grad_norm": 0.6583734154701233,
|
| 1547 |
+
"learning_rate": 0.000027590627670884494,
|
| 1548 |
+
"loss": 0.2984,
|
| 1549 |
+
"step": 22000
|
| 1550 |
+
},
|
| 1551 |
+
{
|
| 1552 |
+
"epoch": 0.46,
|
| 1553 |
+
"grad_norm": 0.9175625443458557,
|
| 1554 |
+
"learning_rate": 0.000027486398032144422,
|
| 1555 |
+
"loss": 0.305,
|
| 1556 |
+
"step": 22100
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 0.46,
|
| 1560 |
+
"grad_norm": 1.192954182624817,
|
| 1561 |
+
"learning_rate": 0.00002738216839340435,
|
| 1562 |
+
"loss": 0.2913,
|
| 1563 |
+
"step": 22200
|
| 1564 |
+
},
|
| 1565 |
+
{
|
| 1566 |
+
"epoch": 0.46,
|
| 1567 |
+
"grad_norm": 0.9170815944671631,
|
| 1568 |
+
"learning_rate": 0.000027277938754664278,
|
| 1569 |
+
"loss": 0.2886,
|
| 1570 |
+
"step": 22300
|
| 1571 |
+
},
|
| 1572 |
+
{
|
| 1573 |
+
"epoch": 0.46,
|
| 1574 |
+
"grad_norm": 0.7114281058311462,
|
| 1575 |
+
"learning_rate": 0.000027173709115924206,
|
| 1576 |
+
"loss": 0.2872,
|
| 1577 |
+
"step": 22400
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 0.46,
|
| 1581 |
+
"grad_norm": 1.1340280771255493,
|
| 1582 |
+
"learning_rate": 0.000027069479477184134,
|
| 1583 |
+
"loss": 0.2927,
|
| 1584 |
+
"step": 22500
|
| 1585 |
+
},
|
| 1586 |
+
{
|
| 1587 |
+
"epoch": 0.47,
|
| 1588 |
+
"grad_norm": 1.0684434175491333,
|
| 1589 |
+
"learning_rate": 0.00002696524983844406,
|
| 1590 |
+
"loss": 0.2762,
|
| 1591 |
+
"step": 22600
|
| 1592 |
+
},
|
| 1593 |
+
{
|
| 1594 |
+
"epoch": 0.47,
|
| 1595 |
+
"grad_norm": 0.8208048343658447,
|
| 1596 |
+
"learning_rate": 0.00002686102019970399,
|
| 1597 |
+
"loss": 0.2933,
|
| 1598 |
+
"step": 22700
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 0.47,
|
| 1602 |
+
"grad_norm": 0.9516797065734863,
|
| 1603 |
+
"learning_rate": 0.000026756790560963917,
|
| 1604 |
+
"loss": 0.2809,
|
| 1605 |
+
"step": 22800
|
| 1606 |
+
},
|
| 1607 |
+
{
|
| 1608 |
+
"epoch": 0.47,
|
| 1609 |
+
"grad_norm": 0.7305499315261841,
|
| 1610 |
+
"learning_rate": 0.000026652560922223845,
|
| 1611 |
+
"loss": 0.2968,
|
| 1612 |
+
"step": 22900
|
| 1613 |
+
},
|
| 1614 |
+
{
|
| 1615 |
+
"epoch": 0.47,
|
| 1616 |
+
"grad_norm": 0.7523550391197205,
|
| 1617 |
+
"learning_rate": 0.000026548331283483773,
|
| 1618 |
+
"loss": 0.289,
|
| 1619 |
+
"step": 23000
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 0.48,
|
| 1623 |
+
"grad_norm": 1.0975223779678345,
|
| 1624 |
+
"learning_rate": 0.0000264441016447437,
|
| 1625 |
+
"loss": 0.2961,
|
| 1626 |
+
"step": 23100
|
| 1627 |
+
},
|
| 1628 |
+
{
|
| 1629 |
+
"epoch": 0.48,
|
| 1630 |
+
"grad_norm": 0.906672477722168,
|
| 1631 |
+
"learning_rate": 0.00002633987200600363,
|
| 1632 |
+
"loss": 0.2904,
|
| 1633 |
+
"step": 23200
|
| 1634 |
+
},
|
| 1635 |
+
{
|
| 1636 |
+
"epoch": 0.48,
|
| 1637 |
+
"grad_norm": 0.8232882022857666,
|
| 1638 |
+
"learning_rate": 0.000026235642367263557,
|
| 1639 |
+
"loss": 0.2876,
|
| 1640 |
+
"step": 23300
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 0.48,
|
| 1644 |
+
"grad_norm": 1.0251737833023071,
|
| 1645 |
+
"learning_rate": 0.000026131412728523485,
|
| 1646 |
+
"loss": 0.2842,
|
| 1647 |
+
"step": 23400
|
| 1648 |
+
},
|
| 1649 |
+
{
|
| 1650 |
+
"epoch": 0.48,
|
| 1651 |
+
"grad_norm": 1.1119544506072998,
|
| 1652 |
+
"learning_rate": 0.000026027183089783413,
|
| 1653 |
+
"loss": 0.2778,
|
| 1654 |
+
"step": 23500
|
| 1655 |
+
},
|
| 1656 |
+
{
|
| 1657 |
+
"epoch": 0.49,
|
| 1658 |
+
"grad_norm": 1.1404234170913696,
|
| 1659 |
+
"learning_rate": 0.00002592295345104334,
|
| 1660 |
+
"loss": 0.2865,
|
| 1661 |
+
"step": 23600
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 0.49,
|
| 1665 |
+
"grad_norm": 1.0083096027374268,
|
| 1666 |
+
"learning_rate": 0.00002581872381230327,
|
| 1667 |
+
"loss": 0.2923,
|
| 1668 |
+
"step": 23700
|
| 1669 |
+
},
|
| 1670 |
+
{
|
| 1671 |
+
"epoch": 0.49,
|
| 1672 |
+
"grad_norm": 1.1562743186950684,
|
| 1673 |
+
"learning_rate": 0.000025714494173563196,
|
| 1674 |
+
"loss": 0.3009,
|
| 1675 |
+
"step": 23800
|
| 1676 |
+
},
|
| 1677 |
+
{
|
| 1678 |
+
"epoch": 0.49,
|
| 1679 |
+
"grad_norm": 0.7771180272102356,
|
| 1680 |
+
"learning_rate": 0.000025610264534823124,
|
| 1681 |
+
"loss": 0.2938,
|
| 1682 |
+
"step": 23900
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 0.5,
|
| 1686 |
+
"grad_norm": 0.8902438282966614,
|
| 1687 |
+
"learning_rate": 0.000025506034896083052,
|
| 1688 |
+
"loss": 0.3102,
|
| 1689 |
+
"step": 24000
|
| 1690 |
+
},
|
| 1691 |
+
{
|
| 1692 |
+
"epoch": 0.5,
|
| 1693 |
+
"grad_norm": 0.9011629819869995,
|
| 1694 |
+
"learning_rate": 0.00002540180525734298,
|
| 1695 |
+
"loss": 0.2885,
|
| 1696 |
+
"step": 24100
|
| 1697 |
+
},
|
| 1698 |
+
{
|
| 1699 |
+
"epoch": 0.5,
|
| 1700 |
+
"grad_norm": 0.8738746643066406,
|
| 1701 |
+
"learning_rate": 0.000025297575618602908,
|
| 1702 |
+
"loss": 0.2739,
|
| 1703 |
+
"step": 24200
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 0.5,
|
| 1707 |
+
"grad_norm": 0.5818448662757874,
|
| 1708 |
+
"learning_rate": 0.000025193345979862836,
|
| 1709 |
+
"loss": 0.281,
|
| 1710 |
+
"step": 24300
|
| 1711 |
+
},
|
| 1712 |
+
{
|
| 1713 |
+
"epoch": 0.5,
|
| 1714 |
+
"grad_norm": 0.927828311920166,
|
| 1715 |
+
"learning_rate": 0.000025089116341122764,
|
| 1716 |
+
"loss": 0.3039,
|
| 1717 |
+
"step": 24400
|
| 1718 |
+
},
|
| 1719 |
+
{
|
| 1720 |
+
"epoch": 0.51,
|
| 1721 |
+
"grad_norm": 0.8303672671318054,
|
| 1722 |
+
"learning_rate": 0.00002498488670238269,
|
| 1723 |
+
"loss": 0.2921,
|
| 1724 |
+
"step": 24500
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 0.51,
|
| 1728 |
+
"grad_norm": 1.7229905128479004,
|
| 1729 |
+
"learning_rate": 0.000024880657063642616,
|
| 1730 |
+
"loss": 0.3003,
|
| 1731 |
+
"step": 24600
|
| 1732 |
+
},
|
| 1733 |
+
{
|
| 1734 |
+
"epoch": 0.51,
|
| 1735 |
+
"grad_norm": 0.898224949836731,
|
| 1736 |
+
"learning_rate": 0.000024776427424902544,
|
| 1737 |
+
"loss": 0.3034,
|
| 1738 |
+
"step": 24700
|
| 1739 |
+
},
|
| 1740 |
+
{
|
| 1741 |
+
"epoch": 0.51,
|
| 1742 |
+
"grad_norm": 0.7359880805015564,
|
| 1743 |
+
"learning_rate": 0.000024672197786162472,
|
| 1744 |
+
"loss": 0.2918,
|
| 1745 |
+
"step": 24800
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 0.51,
|
| 1749 |
+
"grad_norm": 0.9466255903244019,
|
| 1750 |
+
"learning_rate": 0.0000245679681474224,
|
| 1751 |
+
"loss": 0.2982,
|
| 1752 |
+
"step": 24900
|
| 1753 |
+
},
|
| 1754 |
+
{
|
| 1755 |
+
"epoch": 0.52,
|
| 1756 |
+
"grad_norm": 0.9300728440284729,
|
| 1757 |
+
"learning_rate": 0.00002446373850868233,
|
| 1758 |
+
"loss": 0.3049,
|
| 1759 |
+
"step": 25000
|
| 1760 |
+
},
|
| 1761 |
+
{
|
| 1762 |
+
"epoch": 0.52,
|
| 1763 |
+
"grad_norm": 1.2316802740097046,
|
| 1764 |
+
"learning_rate": 0.00002435950886994226,
|
| 1765 |
+
"loss": 0.3088,
|
| 1766 |
+
"step": 25100
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 0.52,
|
| 1770 |
+
"grad_norm": 1.1127874851226807,
|
| 1771 |
+
"learning_rate": 0.000024255279231202187,
|
| 1772 |
+
"loss": 0.2844,
|
| 1773 |
+
"step": 25200
|
| 1774 |
+
},
|
| 1775 |
+
{
|
| 1776 |
+
"epoch": 0.52,
|
| 1777 |
+
"grad_norm": 1.0291988849639893,
|
| 1778 |
+
"learning_rate": 0.000024151049592462115,
|
| 1779 |
+
"loss": 0.2844,
|
| 1780 |
+
"step": 25300
|
| 1781 |
+
},
|
| 1782 |
+
{
|
| 1783 |
+
"epoch": 0.52,
|
| 1784 |
+
"grad_norm": 1.0049411058425903,
|
| 1785 |
+
"learning_rate": 0.000024046819953722043,
|
| 1786 |
+
"loss": 0.2902,
|
| 1787 |
+
"step": 25400
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 0.53,
|
| 1791 |
+
"grad_norm": 7.540700912475586,
|
| 1792 |
+
"learning_rate": 0.00002394259031498197,
|
| 1793 |
+
"loss": 0.2966,
|
| 1794 |
+
"step": 25500
|
| 1795 |
+
},
|
| 1796 |
+
{
|
| 1797 |
+
"epoch": 0.53,
|
| 1798 |
+
"grad_norm": 0.7970338463783264,
|
| 1799 |
+
"learning_rate": 0.0000238383606762419,
|
| 1800 |
+
"loss": 0.2714,
|
| 1801 |
+
"step": 25600
|
| 1802 |
+
},
|
| 1803 |
+
{
|
| 1804 |
+
"epoch": 0.53,
|
| 1805 |
+
"grad_norm": 0.7852484583854675,
|
| 1806 |
+
"learning_rate": 0.000023734131037501827,
|
| 1807 |
+
"loss": 0.3103,
|
| 1808 |
+
"step": 25700
|
| 1809 |
+
},
|
| 1810 |
+
{
|
| 1811 |
+
"epoch": 0.53,
|
| 1812 |
+
"grad_norm": 1.1321452856063843,
|
| 1813 |
+
"learning_rate": 0.000023629901398761754,
|
| 1814 |
+
"loss": 0.2939,
|
| 1815 |
+
"step": 25800
|
| 1816 |
+
},
|
| 1817 |
+
{
|
| 1818 |
+
"epoch": 0.53,
|
| 1819 |
+
"grad_norm": 0.8240513801574707,
|
| 1820 |
+
"learning_rate": 0.000023525671760021682,
|
| 1821 |
+
"loss": 0.2933,
|
| 1822 |
+
"step": 25900
|
| 1823 |
+
},
|
| 1824 |
+
{
|
| 1825 |
+
"epoch": 0.54,
|
| 1826 |
+
"grad_norm": 0.8263924717903137,
|
| 1827 |
+
"learning_rate": 0.00002342144212128161,
|
| 1828 |
+
"loss": 0.2764,
|
| 1829 |
+
"step": 26000
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 0.54,
|
| 1833 |
+
"grad_norm": 0.48156994581222534,
|
| 1834 |
+
"learning_rate": 0.000023317212482541538,
|
| 1835 |
+
"loss": 0.2785,
|
| 1836 |
+
"step": 26100
|
| 1837 |
+
},
|
| 1838 |
+
{
|
| 1839 |
+
"epoch": 0.54,
|
| 1840 |
+
"grad_norm": 0.7034128904342651,
|
| 1841 |
+
"learning_rate": 0.000023212982843801466,
|
| 1842 |
+
"loss": 0.2884,
|
| 1843 |
+
"step": 26200
|
| 1844 |
+
},
|
| 1845 |
+
{
|
| 1846 |
+
"epoch": 0.54,
|
| 1847 |
+
"grad_norm": 0.7832284569740295,
|
| 1848 |
+
"learning_rate": 0.000023108753205061394,
|
| 1849 |
+
"loss": 0.2933,
|
| 1850 |
+
"step": 26300
|
| 1851 |
+
},
|
| 1852 |
+
{
|
| 1853 |
+
"epoch": 0.54,
|
| 1854 |
+
"grad_norm": 1.0077266693115234,
|
| 1855 |
+
"learning_rate": 0.000023004523566321322,
|
| 1856 |
+
"loss": 0.2929,
|
| 1857 |
+
"step": 26400
|
| 1858 |
+
},
|
| 1859 |
+
{
|
| 1860 |
+
"epoch": 0.55,
|
| 1861 |
+
"grad_norm": 0.7659132480621338,
|
| 1862 |
+
"learning_rate": 0.00002290029392758125,
|
| 1863 |
+
"loss": 0.277,
|
| 1864 |
+
"step": 26500
|
| 1865 |
+
},
|
| 1866 |
+
{
|
| 1867 |
+
"epoch": 0.55,
|
| 1868 |
+
"grad_norm": 0.7871528267860413,
|
| 1869 |
+
"learning_rate": 0.000022796064288841178,
|
| 1870 |
+
"loss": 0.2995,
|
| 1871 |
+
"step": 26600
|
| 1872 |
+
},
|
| 1873 |
+
{
|
| 1874 |
+
"epoch": 0.55,
|
| 1875 |
+
"grad_norm": 0.9168059825897217,
|
| 1876 |
+
"learning_rate": 0.000022691834650101106,
|
| 1877 |
+
"loss": 0.3038,
|
| 1878 |
+
"step": 26700
|
| 1879 |
+
},
|
| 1880 |
+
{
|
| 1881 |
+
"epoch": 0.55,
|
| 1882 |
+
"grad_norm": 0.6936691403388977,
|
| 1883 |
+
"learning_rate": 0.000022587605011361033,
|
| 1884 |
+
"loss": 0.2838,
|
| 1885 |
+
"step": 26800
|
| 1886 |
+
},
|
| 1887 |
+
{
|
| 1888 |
+
"epoch": 0.55,
|
| 1889 |
+
"grad_norm": 0.9125056266784668,
|
| 1890 |
+
"learning_rate": 0.00002248337537262096,
|
| 1891 |
+
"loss": 0.2832,
|
| 1892 |
+
"step": 26900
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 0.56,
|
| 1896 |
+
"grad_norm": 0.8433107733726501,
|
| 1897 |
+
"learning_rate": 0.00002237914573388089,
|
| 1898 |
+
"loss": 0.2963,
|
| 1899 |
+
"step": 27000
|
| 1900 |
+
},
|
| 1901 |
+
{
|
| 1902 |
+
"epoch": 0.56,
|
| 1903 |
+
"grad_norm": 0.6140872836112976,
|
| 1904 |
+
"learning_rate": 0.000022274916095140817,
|
| 1905 |
+
"loss": 0.256,
|
| 1906 |
+
"step": 27100
|
| 1907 |
+
},
|
| 1908 |
+
{
|
| 1909 |
+
"epoch": 0.56,
|
| 1910 |
+
"grad_norm": 0.9464013576507568,
|
| 1911 |
+
"learning_rate": 0.000022170686456400745,
|
| 1912 |
+
"loss": 0.279,
|
| 1913 |
+
"step": 27200
|
| 1914 |
+
},
|
| 1915 |
+
{
|
| 1916 |
+
"epoch": 0.56,
|
| 1917 |
+
"grad_norm": 0.6698806881904602,
|
| 1918 |
+
"learning_rate": 0.000022066456817660673,
|
| 1919 |
+
"loss": 0.2664,
|
| 1920 |
+
"step": 27300
|
| 1921 |
+
},
|
| 1922 |
+
{
|
| 1923 |
+
"epoch": 0.57,
|
| 1924 |
+
"grad_norm": 1.0536932945251465,
|
| 1925 |
+
"learning_rate": 0.0000219622271789206,
|
| 1926 |
+
"loss": 0.2768,
|
| 1927 |
+
"step": 27400
|
| 1928 |
+
},
|
| 1929 |
+
{
|
| 1930 |
+
"epoch": 0.57,
|
| 1931 |
+
"grad_norm": 0.9017496705055237,
|
| 1932 |
+
"learning_rate": 0.00002185799754018053,
|
| 1933 |
+
"loss": 0.2723,
|
| 1934 |
+
"step": 27500
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 0.57,
|
| 1938 |
+
"grad_norm": 0.9188879728317261,
|
| 1939 |
+
"learning_rate": 0.000021753767901440457,
|
| 1940 |
+
"loss": 0.2925,
|
| 1941 |
+
"step": 27600
|
| 1942 |
+
},
|
| 1943 |
+
{
|
| 1944 |
+
"epoch": 0.57,
|
| 1945 |
+
"grad_norm": 0.716852068901062,
|
| 1946 |
+
"learning_rate": 0.000021649538262700385,
|
| 1947 |
+
"loss": 0.2789,
|
| 1948 |
+
"step": 27700
|
| 1949 |
+
},
|
| 1950 |
+
{
|
| 1951 |
+
"epoch": 0.57,
|
| 1952 |
+
"grad_norm": 0.8829745650291443,
|
| 1953 |
+
"learning_rate": 0.000021545308623960312,
|
| 1954 |
+
"loss": 0.2613,
|
| 1955 |
+
"step": 27800
|
| 1956 |
+
},
|
| 1957 |
+
{
|
| 1958 |
+
"epoch": 0.58,
|
| 1959 |
+
"grad_norm": 0.8111447095870972,
|
| 1960 |
+
"learning_rate": 0.00002144107898522024,
|
| 1961 |
+
"loss": 0.2847,
|
| 1962 |
+
"step": 27900
|
| 1963 |
+
},
|
| 1964 |
+
{
|
| 1965 |
+
"epoch": 0.58,
|
| 1966 |
+
"grad_norm": 1.069851279258728,
|
| 1967 |
+
"learning_rate": 0.000021336849346480168,
|
| 1968 |
+
"loss": 0.2865,
|
| 1969 |
+
"step": 28000
|
| 1970 |
+
},
|
| 1971 |
+
{
|
| 1972 |
+
"epoch": 0.58,
|
| 1973 |
+
"grad_norm": 1.1249984502792358,
|
| 1974 |
+
"learning_rate": 0.000021232619707740096,
|
| 1975 |
+
"loss": 0.285,
|
| 1976 |
+
"step": 28100
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 0.58,
|
| 1980 |
+
"grad_norm": 0.8425958156585693,
|
| 1981 |
+
"learning_rate": 0.000021128390069000024,
|
| 1982 |
+
"loss": 0.2822,
|
| 1983 |
+
"step": 28200
|
| 1984 |
+
},
|
| 1985 |
+
{
|
| 1986 |
+
"epoch": 0.58,
|
| 1987 |
+
"grad_norm": 0.8334409594535828,
|
| 1988 |
+
"learning_rate": 0.000021024160430259952,
|
| 1989 |
+
"loss": 0.2915,
|
| 1990 |
+
"step": 28300
|
| 1991 |
+
},
|
| 1992 |
+
{
|
| 1993 |
+
"epoch": 0.59,
|
| 1994 |
+
"grad_norm": 0.8733804225921631,
|
| 1995 |
+
"learning_rate": 0.00002091993079151988,
|
| 1996 |
+
"loss": 0.2784,
|
| 1997 |
+
"step": 28400
|
| 1998 |
+
},
|
| 1999 |
+
{
|
| 2000 |
+
"epoch": 0.59,
|
| 2001 |
+
"grad_norm": 1.1405348777770996,
|
| 2002 |
+
"learning_rate": 0.000020815701152779808,
|
| 2003 |
+
"loss": 0.2632,
|
| 2004 |
+
"step": 28500
|
| 2005 |
+
},
|
| 2006 |
+
{
|
| 2007 |
+
"epoch": 0.59,
|
| 2008 |
+
"grad_norm": 0.930985689163208,
|
| 2009 |
+
"learning_rate": 0.000020711471514039732,
|
| 2010 |
+
"loss": 0.2804,
|
| 2011 |
+
"step": 28600
|
| 2012 |
+
},
|
| 2013 |
+
{
|
| 2014 |
+
"epoch": 0.59,
|
| 2015 |
+
"grad_norm": 0.7500423789024353,
|
| 2016 |
+
"learning_rate": 0.00002060724187529966,
|
| 2017 |
+
"loss": 0.271,
|
| 2018 |
+
"step": 28700
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 0.59,
|
| 2022 |
+
"grad_norm": 0.830214262008667,
|
| 2023 |
+
"learning_rate": 0.000020503012236559588,
|
| 2024 |
+
"loss": 0.2802,
|
| 2025 |
+
"step": 28800
|
| 2026 |
+
},
|
| 2027 |
+
{
|
| 2028 |
+
"epoch": 0.6,
|
| 2029 |
+
"grad_norm": 1.1190361976623535,
|
| 2030 |
+
"learning_rate": 0.000020398782597819516,
|
| 2031 |
+
"loss": 0.2978,
|
| 2032 |
+
"step": 28900
|
| 2033 |
+
},
|
| 2034 |
+
{
|
| 2035 |
+
"epoch": 0.6,
|
| 2036 |
+
"grad_norm": 0.938105583190918,
|
| 2037 |
+
"learning_rate": 0.000020294552959079444,
|
| 2038 |
+
"loss": 0.273,
|
| 2039 |
+
"step": 29000
|
| 2040 |
+
},
|
| 2041 |
+
{
|
| 2042 |
+
"epoch": 0.6,
|
| 2043 |
+
"grad_norm": 1.2504827976226807,
|
| 2044 |
+
"learning_rate": 0.000020190323320339372,
|
| 2045 |
+
"loss": 0.2802,
|
| 2046 |
+
"step": 29100
|
| 2047 |
+
},
|
| 2048 |
+
{
|
| 2049 |
+
"epoch": 0.6,
|
| 2050 |
+
"grad_norm": 0.6229212284088135,
|
| 2051 |
+
"learning_rate": 0.0000200860936815993,
|
| 2052 |
+
"loss": 0.2751,
|
| 2053 |
+
"step": 29200
|
| 2054 |
+
},
|
| 2055 |
+
{
|
| 2056 |
+
"epoch": 0.6,
|
| 2057 |
+
"grad_norm": 0.917674720287323,
|
| 2058 |
+
"learning_rate": 0.000019981864042859228,
|
| 2059 |
+
"loss": 0.2668,
|
| 2060 |
+
"step": 29300
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 0.61,
|
| 2064 |
+
"grad_norm": 0.7844977378845215,
|
| 2065 |
+
"learning_rate": 0.000019877634404119155,
|
| 2066 |
+
"loss": 0.2772,
|
| 2067 |
+
"step": 29400
|
| 2068 |
+
},
|
| 2069 |
+
{
|
| 2070 |
+
"epoch": 0.61,
|
| 2071 |
+
"grad_norm": 0.7926586866378784,
|
| 2072 |
+
"learning_rate": 0.000019773404765379083,
|
| 2073 |
+
"loss": 0.2851,
|
| 2074 |
+
"step": 29500
|
| 2075 |
+
},
|
| 2076 |
+
{
|
| 2077 |
+
"epoch": 0.61,
|
| 2078 |
+
"grad_norm": 0.9423878192901611,
|
| 2079 |
+
"learning_rate": 0.00001966917512663901,
|
| 2080 |
+
"loss": 0.2973,
|
| 2081 |
+
"step": 29600
|
| 2082 |
+
},
|
| 2083 |
+
{
|
| 2084 |
+
"epoch": 0.61,
|
| 2085 |
+
"grad_norm": 1.1130839586257935,
|
| 2086 |
+
"learning_rate": 0.00001956494548789894,
|
| 2087 |
+
"loss": 0.2858,
|
| 2088 |
+
"step": 29700
|
| 2089 |
+
},
|
| 2090 |
+
{
|
| 2091 |
+
"epoch": 0.61,
|
| 2092 |
+
"grad_norm": 0.8832184076309204,
|
| 2093 |
+
"learning_rate": 0.000019460715849158867,
|
| 2094 |
+
"loss": 0.2714,
|
| 2095 |
+
"step": 29800
|
| 2096 |
+
},
|
| 2097 |
+
{
|
| 2098 |
+
"epoch": 0.62,
|
| 2099 |
+
"grad_norm": 0.7974615693092346,
|
| 2100 |
+
"learning_rate": 0.000019356486210418795,
|
| 2101 |
+
"loss": 0.2672,
|
| 2102 |
+
"step": 29900
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 0.62,
|
| 2106 |
+
"grad_norm": 1.0496609210968018,
|
| 2107 |
+
"learning_rate": 0.000019252256571678723,
|
| 2108 |
+
"loss": 0.2967,
|
| 2109 |
+
"step": 30000
|
| 2110 |
+
},
|
| 2111 |
+
{
|
| 2112 |
+
"epoch": 0.62,
|
| 2113 |
+
"grad_norm": 0.7052696943283081,
|
| 2114 |
+
"learning_rate": 0.00001914802693293865,
|
| 2115 |
+
"loss": 0.2682,
|
| 2116 |
+
"step": 30100
|
| 2117 |
+
},
|
| 2118 |
+
{
|
| 2119 |
+
"epoch": 0.62,
|
| 2120 |
+
"grad_norm": 0.8177437782287598,
|
| 2121 |
+
"learning_rate": 0.00001904379729419858,
|
| 2122 |
+
"loss": 0.2772,
|
| 2123 |
+
"step": 30200
|
| 2124 |
+
},
|
| 2125 |
+
{
|
| 2126 |
+
"epoch": 0.63,
|
| 2127 |
+
"grad_norm": 1.1367732286453247,
|
| 2128 |
+
"learning_rate": 0.000018939567655458507,
|
| 2129 |
+
"loss": 0.2751,
|
| 2130 |
+
"step": 30300
|
| 2131 |
+
},
|
| 2132 |
+
{
|
| 2133 |
+
"epoch": 0.63,
|
| 2134 |
+
"grad_norm": 1.4945389032363892,
|
| 2135 |
+
"learning_rate": 0.000018835338016718434,
|
| 2136 |
+
"loss": 0.2845,
|
| 2137 |
+
"step": 30400
|
| 2138 |
+
},
|
| 2139 |
+
{
|
| 2140 |
+
"epoch": 0.63,
|
| 2141 |
+
"grad_norm": 1.035019040107727,
|
| 2142 |
+
"learning_rate": 0.000018731108377978362,
|
| 2143 |
+
"loss": 0.2881,
|
| 2144 |
+
"step": 30500
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 0.63,
|
| 2148 |
+
"grad_norm": 0.4365576207637787,
|
| 2149 |
+
"learning_rate": 0.00001862687873923829,
|
| 2150 |
+
"loss": 0.2805,
|
| 2151 |
+
"step": 30600
|
| 2152 |
+
},
|
| 2153 |
+
{
|
| 2154 |
+
"epoch": 0.63,
|
| 2155 |
+
"grad_norm": 0.6886979341506958,
|
| 2156 |
+
"learning_rate": 0.000018522649100498218,
|
| 2157 |
+
"loss": 0.2755,
|
| 2158 |
+
"step": 30700
|
| 2159 |
+
},
|
| 2160 |
+
{
|
| 2161 |
+
"epoch": 0.64,
|
| 2162 |
+
"grad_norm": 0.7902783751487732,
|
| 2163 |
+
"learning_rate": 0.000018418419461758146,
|
| 2164 |
+
"loss": 0.2746,
|
| 2165 |
+
"step": 30800
|
| 2166 |
+
},
|
| 2167 |
+
{
|
| 2168 |
+
"epoch": 0.64,
|
| 2169 |
+
"grad_norm": 1.1486902236938477,
|
| 2170 |
+
"learning_rate": 0.000018314189823018074,
|
| 2171 |
+
"loss": 0.2734,
|
| 2172 |
+
"step": 30900
|
| 2173 |
+
},
|
| 2174 |
+
{
|
| 2175 |
+
"epoch": 0.64,
|
| 2176 |
+
"grad_norm": 0.8375297784805298,
|
| 2177 |
+
"learning_rate": 0.000018209960184278002,
|
| 2178 |
+
"loss": 0.2672,
|
| 2179 |
+
"step": 31000
|
| 2180 |
+
},
|
| 2181 |
+
{
|
| 2182 |
+
"epoch": 0.64,
|
| 2183 |
+
"grad_norm": 0.8426914811134338,
|
| 2184 |
+
"learning_rate": 0.00001810573054553793,
|
| 2185 |
+
"loss": 0.266,
|
| 2186 |
+
"step": 31100
|
| 2187 |
+
},
|
| 2188 |
+
{
|
| 2189 |
+
"epoch": 0.64,
|
| 2190 |
+
"grad_norm": 1.0207959413528442,
|
| 2191 |
+
"learning_rate": 0.000018001500906797858,
|
| 2192 |
+
"loss": 0.2879,
|
| 2193 |
+
"step": 31200
|
| 2194 |
+
},
|
| 2195 |
+
{
|
| 2196 |
+
"epoch": 0.65,
|
| 2197 |
+
"grad_norm": 0.8999843597412109,
|
| 2198 |
+
"learning_rate": 0.000017897271268057786,
|
| 2199 |
+
"loss": 0.2617,
|
| 2200 |
+
"step": 31300
|
| 2201 |
+
},
|
| 2202 |
+
{
|
| 2203 |
+
"epoch": 0.65,
|
| 2204 |
+
"grad_norm": 0.7014563083648682,
|
| 2205 |
+
"learning_rate": 0.000017793041629317713,
|
| 2206 |
+
"loss": 0.2757,
|
| 2207 |
+
"step": 31400
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 0.65,
|
| 2211 |
+
"grad_norm": 0.8926547765731812,
|
| 2212 |
+
"learning_rate": 0.00001768881199057764,
|
| 2213 |
+
"loss": 0.2805,
|
| 2214 |
+
"step": 31500
|
| 2215 |
+
},
|
| 2216 |
+
{
|
| 2217 |
+
"epoch": 0.65,
|
| 2218 |
+
"grad_norm": 0.9726804494857788,
|
| 2219 |
+
"learning_rate": 0.00001758458235183757,
|
| 2220 |
+
"loss": 0.2824,
|
| 2221 |
+
"step": 31600
|
| 2222 |
+
},
|
| 2223 |
+
{
|
| 2224 |
+
"epoch": 0.65,
|
| 2225 |
+
"grad_norm": 0.949054479598999,
|
| 2226 |
+
"learning_rate": 0.000017480352713097497,
|
| 2227 |
+
"loss": 0.2578,
|
| 2228 |
+
"step": 31700
|
| 2229 |
+
},
|
| 2230 |
+
{
|
| 2231 |
+
"epoch": 0.66,
|
| 2232 |
+
"grad_norm": 1.0712580680847168,
|
| 2233 |
+
"learning_rate": 0.000017376123074357425,
|
| 2234 |
+
"loss": 0.259,
|
| 2235 |
+
"step": 31800
|
| 2236 |
+
},
|
| 2237 |
+
{
|
| 2238 |
+
"epoch": 0.66,
|
| 2239 |
+
"grad_norm": 0.8288059234619141,
|
| 2240 |
+
"learning_rate": 0.000017271893435617353,
|
| 2241 |
+
"loss": 0.2788,
|
| 2242 |
+
"step": 31900
|
| 2243 |
+
},
|
| 2244 |
+
{
|
| 2245 |
+
"epoch": 0.66,
|
| 2246 |
+
"grad_norm": 0.9404420852661133,
|
| 2247 |
+
"learning_rate": 0.00001716766379687728,
|
| 2248 |
+
"loss": 0.2766,
|
| 2249 |
+
"step": 32000
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 0.66,
|
| 2253 |
+
"grad_norm": 0.8590795993804932,
|
| 2254 |
+
"learning_rate": 0.00001706343415813721,
|
| 2255 |
+
"loss": 0.2768,
|
| 2256 |
+
"step": 32100
|
| 2257 |
+
},
|
| 2258 |
+
{
|
| 2259 |
+
"epoch": 0.66,
|
| 2260 |
+
"grad_norm": 1.1562901735305786,
|
| 2261 |
+
"learning_rate": 0.000016959204519397137,
|
| 2262 |
+
"loss": 0.2876,
|
| 2263 |
+
"step": 32200
|
| 2264 |
+
},
|
| 2265 |
+
{
|
| 2266 |
+
"epoch": 0.67,
|
| 2267 |
+
"grad_norm": 0.8512832522392273,
|
| 2268 |
+
"learning_rate": 0.000016854974880657065,
|
| 2269 |
+
"loss": 0.2599,
|
| 2270 |
+
"step": 32300
|
| 2271 |
+
},
|
| 2272 |
+
{
|
| 2273 |
+
"epoch": 0.67,
|
| 2274 |
+
"grad_norm": 0.8616644740104675,
|
| 2275 |
+
"learning_rate": 0.000016750745241916992,
|
| 2276 |
+
"loss": 0.2725,
|
| 2277 |
+
"step": 32400
|
| 2278 |
+
},
|
| 2279 |
+
{
|
| 2280 |
+
"epoch": 0.67,
|
| 2281 |
+
"grad_norm": 1.4701212644577026,
|
| 2282 |
+
"learning_rate": 0.00001664651560317692,
|
| 2283 |
+
"loss": 0.2592,
|
| 2284 |
+
"step": 32500
|
| 2285 |
+
},
|
| 2286 |
+
{
|
| 2287 |
+
"epoch": 0.67,
|
| 2288 |
+
"grad_norm": 1.0344711542129517,
|
| 2289 |
+
"learning_rate": 0.000016542285964436848,
|
| 2290 |
+
"loss": 0.2782,
|
| 2291 |
+
"step": 32600
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 0.67,
|
| 2295 |
+
"grad_norm": 0.9090347290039062,
|
| 2296 |
+
"learning_rate": 0.000016438056325696776,
|
| 2297 |
+
"loss": 0.2683,
|
| 2298 |
+
"step": 32700
|
| 2299 |
+
},
|
| 2300 |
+
{
|
| 2301 |
+
"epoch": 0.68,
|
| 2302 |
+
"grad_norm": 0.9814984798431396,
|
| 2303 |
+
"learning_rate": 0.000016333826686956704,
|
| 2304 |
+
"loss": 0.2692,
|
| 2305 |
+
"step": 32800
|
| 2306 |
+
},
|
| 2307 |
+
{
|
| 2308 |
+
"epoch": 0.68,
|
| 2309 |
+
"grad_norm": 0.8351295590400696,
|
| 2310 |
+
"learning_rate": 0.000016229597048216632,
|
| 2311 |
+
"loss": 0.2707,
|
| 2312 |
+
"step": 32900
|
| 2313 |
+
},
|
| 2314 |
+
{
|
| 2315 |
+
"epoch": 0.68,
|
| 2316 |
+
"grad_norm": 0.8099172115325928,
|
| 2317 |
+
"learning_rate": 0.00001612536740947656,
|
| 2318 |
+
"loss": 0.2737,
|
| 2319 |
+
"step": 33000
|
| 2320 |
+
},
|
| 2321 |
+
{
|
| 2322 |
+
"epoch": 0.68,
|
| 2323 |
+
"grad_norm": 0.8135939240455627,
|
| 2324 |
+
"learning_rate": 0.000016021137770736488,
|
| 2325 |
+
"loss": 0.2641,
|
| 2326 |
+
"step": 33100
|
| 2327 |
+
},
|
| 2328 |
+
{
|
| 2329 |
+
"epoch": 0.68,
|
| 2330 |
+
"grad_norm": 1.0211602449417114,
|
| 2331 |
+
"learning_rate": 0.000015916908131996416,
|
| 2332 |
+
"loss": 0.2719,
|
| 2333 |
+
"step": 33200
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 0.69,
|
| 2337 |
+
"grad_norm": 1.499255657196045,
|
| 2338 |
+
"learning_rate": 0.000015812678493256344,
|
| 2339 |
+
"loss": 0.2774,
|
| 2340 |
+
"step": 33300
|
| 2341 |
+
},
|
| 2342 |
+
{
|
| 2343 |
+
"epoch": 0.69,
|
| 2344 |
+
"grad_norm": 1.062286376953125,
|
| 2345 |
+
"learning_rate": 0.00001570844885451627,
|
| 2346 |
+
"loss": 0.2631,
|
| 2347 |
+
"step": 33400
|
| 2348 |
+
},
|
| 2349 |
+
{
|
| 2350 |
+
"epoch": 0.69,
|
| 2351 |
+
"grad_norm": 0.7480090260505676,
|
| 2352 |
+
"learning_rate": 0.0000156042192157762,
|
| 2353 |
+
"loss": 0.2697,
|
| 2354 |
+
"step": 33500
|
| 2355 |
+
},
|
| 2356 |
+
{
|
| 2357 |
+
"epoch": 0.69,
|
| 2358 |
+
"grad_norm": 0.8319742679595947,
|
| 2359 |
+
"learning_rate": 0.000015499989577036127,
|
| 2360 |
+
"loss": 0.286,
|
| 2361 |
+
"step": 33600
|
| 2362 |
+
},
|
| 2363 |
+
{
|
| 2364 |
+
"epoch": 0.7,
|
| 2365 |
+
"grad_norm": 1.1275886297225952,
|
| 2366 |
+
"learning_rate": 0.000015395759938296052,
|
| 2367 |
+
"loss": 0.2855,
|
| 2368 |
+
"step": 33700
|
| 2369 |
+
},
|
| 2370 |
+
{
|
| 2371 |
+
"epoch": 0.7,
|
| 2372 |
+
"grad_norm": 0.9062207937240601,
|
| 2373 |
+
"learning_rate": 0.00001529153029955598,
|
| 2374 |
+
"loss": 0.268,
|
| 2375 |
+
"step": 33800
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 0.7,
|
| 2379 |
+
"grad_norm": 0.8947410583496094,
|
| 2380 |
+
"learning_rate": 0.00001518730066081591,
|
| 2381 |
+
"loss": 0.2679,
|
| 2382 |
+
"step": 33900
|
| 2383 |
+
},
|
| 2384 |
+
{
|
| 2385 |
+
"epoch": 0.7,
|
| 2386 |
+
"grad_norm": 1.0549769401550293,
|
| 2387 |
+
"learning_rate": 0.000015083071022075837,
|
| 2388 |
+
"loss": 0.2691,
|
| 2389 |
+
"step": 34000
|
| 2390 |
+
},
|
| 2391 |
+
{
|
| 2392 |
+
"epoch": 0.7,
|
| 2393 |
+
"grad_norm": 0.8998539447784424,
|
| 2394 |
+
"learning_rate": 0.000014978841383335765,
|
| 2395 |
+
"loss": 0.2688,
|
| 2396 |
+
"step": 34100
|
| 2397 |
+
},
|
| 2398 |
+
{
|
| 2399 |
+
"epoch": 0.71,
|
| 2400 |
+
"grad_norm": 0.6037629246711731,
|
| 2401 |
+
"learning_rate": 0.000014874611744595693,
|
| 2402 |
+
"loss": 0.267,
|
| 2403 |
+
"step": 34200
|
| 2404 |
+
},
|
| 2405 |
+
{
|
| 2406 |
+
"epoch": 0.71,
|
| 2407 |
+
"grad_norm": 0.92879718542099,
|
| 2408 |
+
"learning_rate": 0.000014770382105855621,
|
| 2409 |
+
"loss": 0.2667,
|
| 2410 |
+
"step": 34300
|
| 2411 |
+
},
|
| 2412 |
+
{
|
| 2413 |
+
"epoch": 0.71,
|
| 2414 |
+
"grad_norm": 1.562403678894043,
|
| 2415 |
+
"learning_rate": 0.000014666152467115549,
|
| 2416 |
+
"loss": 0.2812,
|
| 2417 |
+
"step": 34400
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 0.71,
|
| 2421 |
+
"grad_norm": 1.0794057846069336,
|
| 2422 |
+
"learning_rate": 0.000014561922828375477,
|
| 2423 |
+
"loss": 0.2691,
|
| 2424 |
+
"step": 34500
|
| 2425 |
+
},
|
| 2426 |
+
{
|
| 2427 |
+
"epoch": 0.71,
|
| 2428 |
+
"grad_norm": 1.1239275932312012,
|
| 2429 |
+
"learning_rate": 0.000014457693189635405,
|
| 2430 |
+
"loss": 0.2626,
|
| 2431 |
+
"step": 34600
|
| 2432 |
+
},
|
| 2433 |
+
{
|
| 2434 |
+
"epoch": 0.72,
|
| 2435 |
+
"grad_norm": 1.2695655822753906,
|
| 2436 |
+
"learning_rate": 0.000014353463550895332,
|
| 2437 |
+
"loss": 0.2679,
|
| 2438 |
+
"step": 34700
|
| 2439 |
+
},
|
| 2440 |
+
{
|
| 2441 |
+
"epoch": 0.72,
|
| 2442 |
+
"grad_norm": 0.9684453010559082,
|
| 2443 |
+
"learning_rate": 0.00001424923391215526,
|
| 2444 |
+
"loss": 0.2689,
|
| 2445 |
+
"step": 34800
|
| 2446 |
+
},
|
| 2447 |
+
{
|
| 2448 |
+
"epoch": 0.72,
|
| 2449 |
+
"grad_norm": 0.8536735773086548,
|
| 2450 |
+
"learning_rate": 0.000014145004273415188,
|
| 2451 |
+
"loss": 0.2505,
|
| 2452 |
+
"step": 34900
|
| 2453 |
+
},
|
| 2454 |
+
{
|
| 2455 |
+
"epoch": 0.72,
|
| 2456 |
+
"grad_norm": 0.8420373797416687,
|
| 2457 |
+
"learning_rate": 0.000014040774634675116,
|
| 2458 |
+
"loss": 0.2808,
|
| 2459 |
+
"step": 35000
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 0.72,
|
| 2463 |
+
"grad_norm": 1.7913522720336914,
|
| 2464 |
+
"learning_rate": 0.000013936544995935044,
|
| 2465 |
+
"loss": 0.2713,
|
| 2466 |
+
"step": 35100
|
| 2467 |
+
},
|
| 2468 |
+
{
|
| 2469 |
+
"epoch": 0.73,
|
| 2470 |
+
"grad_norm": 0.9944103360176086,
|
| 2471 |
+
"learning_rate": 0.000013832315357194972,
|
| 2472 |
+
"loss": 0.2767,
|
| 2473 |
+
"step": 35200
|
| 2474 |
+
},
|
| 2475 |
+
{
|
| 2476 |
+
"epoch": 0.73,
|
| 2477 |
+
"grad_norm": 0.647555947303772,
|
| 2478 |
+
"learning_rate": 0.0000137280857184549,
|
| 2479 |
+
"loss": 0.2637,
|
| 2480 |
+
"step": 35300
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"epoch": 0.73,
|
| 2484 |
+
"grad_norm": 0.6362162828445435,
|
| 2485 |
+
"learning_rate": 0.000013623856079714828,
|
| 2486 |
+
"loss": 0.2626,
|
| 2487 |
+
"step": 35400
|
| 2488 |
+
},
|
| 2489 |
+
{
|
| 2490 |
+
"epoch": 0.73,
|
| 2491 |
+
"grad_norm": 0.8359599113464355,
|
| 2492 |
+
"learning_rate": 0.000013519626440974756,
|
| 2493 |
+
"loss": 0.2819,
|
| 2494 |
+
"step": 35500
|
| 2495 |
+
},
|
| 2496 |
+
{
|
| 2497 |
+
"epoch": 0.73,
|
| 2498 |
+
"grad_norm": 0.8736098408699036,
|
| 2499 |
+
"learning_rate": 0.000013415396802234684,
|
| 2500 |
+
"loss": 0.2699,
|
| 2501 |
+
"step": 35600
|
| 2502 |
+
},
|
| 2503 |
+
{
|
| 2504 |
+
"epoch": 0.74,
|
| 2505 |
+
"grad_norm": 0.8288611173629761,
|
| 2506 |
+
"learning_rate": 0.000013311167163494611,
|
| 2507 |
+
"loss": 0.2621,
|
| 2508 |
+
"step": 35700
|
| 2509 |
+
},
|
| 2510 |
+
{
|
| 2511 |
+
"epoch": 0.74,
|
| 2512 |
+
"grad_norm": 0.7499838471412659,
|
| 2513 |
+
"learning_rate": 0.00001320693752475454,
|
| 2514 |
+
"loss": 0.2692,
|
| 2515 |
+
"step": 35800
|
| 2516 |
+
},
|
| 2517 |
+
{
|
| 2518 |
+
"epoch": 0.74,
|
| 2519 |
+
"grad_norm": 1.0514994859695435,
|
| 2520 |
+
"learning_rate": 0.000013102707886014467,
|
| 2521 |
+
"loss": 0.2712,
|
| 2522 |
+
"step": 35900
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 0.74,
|
| 2526 |
+
"grad_norm": 0.8135703206062317,
|
| 2527 |
+
"learning_rate": 0.000012998478247274395,
|
| 2528 |
+
"loss": 0.267,
|
| 2529 |
+
"step": 36000
|
| 2530 |
+
},
|
| 2531 |
+
{
|
| 2532 |
+
"epoch": 0.74,
|
| 2533 |
+
"grad_norm": 0.780437707901001,
|
| 2534 |
+
"learning_rate": 0.000012894248608534323,
|
| 2535 |
+
"loss": 0.2578,
|
| 2536 |
+
"step": 36100
|
| 2537 |
+
},
|
| 2538 |
+
{
|
| 2539 |
+
"epoch": 0.75,
|
| 2540 |
+
"grad_norm": 1.0182126760482788,
|
| 2541 |
+
"learning_rate": 0.000012790018969794251,
|
| 2542 |
+
"loss": 0.2848,
|
| 2543 |
+
"step": 36200
|
| 2544 |
+
},
|
| 2545 |
+
{
|
| 2546 |
+
"epoch": 0.75,
|
| 2547 |
+
"grad_norm": 0.9728681445121765,
|
| 2548 |
+
"learning_rate": 0.000012685789331054179,
|
| 2549 |
+
"loss": 0.2772,
|
| 2550 |
+
"step": 36300
|
| 2551 |
+
},
|
| 2552 |
+
{
|
| 2553 |
+
"epoch": 0.75,
|
| 2554 |
+
"grad_norm": 0.7952257394790649,
|
| 2555 |
+
"learning_rate": 0.000012581559692314107,
|
| 2556 |
+
"loss": 0.261,
|
| 2557 |
+
"step": 36400
|
| 2558 |
+
},
|
| 2559 |
+
{
|
| 2560 |
+
"epoch": 0.75,
|
| 2561 |
+
"grad_norm": 0.7164922952651978,
|
| 2562 |
+
"learning_rate": 0.000012477330053574035,
|
| 2563 |
+
"loss": 0.2403,
|
| 2564 |
+
"step": 36500
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 0.76,
|
| 2568 |
+
"grad_norm": 0.6388373970985413,
|
| 2569 |
+
"learning_rate": 0.000012373100414833963,
|
| 2570 |
+
"loss": 0.2705,
|
| 2571 |
+
"step": 36600
|
| 2572 |
+
},
|
| 2573 |
+
{
|
| 2574 |
+
"epoch": 0.76,
|
| 2575 |
+
"grad_norm": 1.4427376985549927,
|
| 2576 |
+
"learning_rate": 0.00001226887077609389,
|
| 2577 |
+
"loss": 0.277,
|
| 2578 |
+
"step": 36700
|
| 2579 |
+
},
|
| 2580 |
+
{
|
| 2581 |
+
"epoch": 0.76,
|
| 2582 |
+
"grad_norm": 0.9559366106987,
|
| 2583 |
+
"learning_rate": 0.000012164641137353818,
|
| 2584 |
+
"loss": 0.2784,
|
| 2585 |
+
"step": 36800
|
| 2586 |
+
},
|
| 2587 |
+
{
|
| 2588 |
+
"epoch": 0.76,
|
| 2589 |
+
"grad_norm": 0.8561313152313232,
|
| 2590 |
+
"learning_rate": 0.000012060411498613746,
|
| 2591 |
+
"loss": 0.272,
|
| 2592 |
+
"step": 36900
|
| 2593 |
+
},
|
| 2594 |
+
{
|
| 2595 |
+
"epoch": 0.76,
|
| 2596 |
+
"grad_norm": 0.9756980538368225,
|
| 2597 |
+
"learning_rate": 0.000011956181859873674,
|
| 2598 |
+
"loss": 0.2644,
|
| 2599 |
+
"step": 37000
|
| 2600 |
+
},
|
| 2601 |
+
{
|
| 2602 |
+
"epoch": 0.77,
|
| 2603 |
+
"grad_norm": 1.350045919418335,
|
| 2604 |
+
"learning_rate": 0.000011851952221133602,
|
| 2605 |
+
"loss": 0.2726,
|
| 2606 |
+
"step": 37100
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 0.77,
|
| 2610 |
+
"grad_norm": 1.2612251043319702,
|
| 2611 |
+
"learning_rate": 0.00001174772258239353,
|
| 2612 |
+
"loss": 0.2696,
|
| 2613 |
+
"step": 37200
|
| 2614 |
+
},
|
| 2615 |
+
{
|
| 2616 |
+
"epoch": 0.77,
|
| 2617 |
+
"grad_norm": 0.9203113913536072,
|
| 2618 |
+
"learning_rate": 0.000011643492943653458,
|
| 2619 |
+
"loss": 0.2721,
|
| 2620 |
+
"step": 37300
|
| 2621 |
+
},
|
| 2622 |
+
{
|
| 2623 |
+
"epoch": 0.77,
|
| 2624 |
+
"grad_norm": 1.0679327249526978,
|
| 2625 |
+
"learning_rate": 0.000011539263304913386,
|
| 2626 |
+
"loss": 0.2738,
|
| 2627 |
+
"step": 37400
|
| 2628 |
+
},
|
| 2629 |
+
{
|
| 2630 |
+
"epoch": 0.77,
|
| 2631 |
+
"grad_norm": 0.7020294070243835,
|
| 2632 |
+
"learning_rate": 0.000011435033666173314,
|
| 2633 |
+
"loss": 0.2806,
|
| 2634 |
+
"step": 37500
|
| 2635 |
+
},
|
| 2636 |
+
{
|
| 2637 |
+
"epoch": 0.78,
|
| 2638 |
+
"grad_norm": 0.9969758987426758,
|
| 2639 |
+
"learning_rate": 0.000011330804027433242,
|
| 2640 |
+
"loss": 0.2556,
|
| 2641 |
+
"step": 37600
|
| 2642 |
+
},
|
| 2643 |
+
{
|
| 2644 |
+
"epoch": 0.78,
|
| 2645 |
+
"grad_norm": 0.7139135003089905,
|
| 2646 |
+
"learning_rate": 0.00001122657438869317,
|
| 2647 |
+
"loss": 0.2739,
|
| 2648 |
+
"step": 37700
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 0.78,
|
| 2652 |
+
"grad_norm": 1.3120732307434082,
|
| 2653 |
+
"learning_rate": 0.000011122344749953097,
|
| 2654 |
+
"loss": 0.2774,
|
| 2655 |
+
"step": 37800
|
| 2656 |
+
},
|
| 2657 |
+
{
|
| 2658 |
+
"epoch": 0.78,
|
| 2659 |
+
"grad_norm": 0.9226138591766357,
|
| 2660 |
+
"learning_rate": 0.000011018115111213025,
|
| 2661 |
+
"loss": 0.2414,
|
| 2662 |
+
"step": 37900
|
| 2663 |
+
},
|
| 2664 |
+
{
|
| 2665 |
+
"epoch": 0.78,
|
| 2666 |
+
"grad_norm": 1.024018406867981,
|
| 2667 |
+
"learning_rate": 0.000010913885472472953,
|
| 2668 |
+
"loss": 0.2453,
|
| 2669 |
+
"step": 38000
|
| 2670 |
+
},
|
| 2671 |
+
{
|
| 2672 |
+
"epoch": 0.79,
|
| 2673 |
+
"grad_norm": 0.7024143934249878,
|
| 2674 |
+
"learning_rate": 0.000010809655833732881,
|
| 2675 |
+
"loss": 0.2681,
|
| 2676 |
+
"step": 38100
|
| 2677 |
+
},
|
| 2678 |
+
{
|
| 2679 |
+
"epoch": 0.79,
|
| 2680 |
+
"grad_norm": 1.008748173713684,
|
| 2681 |
+
"learning_rate": 0.000010705426194992809,
|
| 2682 |
+
"loss": 0.2656,
|
| 2683 |
+
"step": 38200
|
| 2684 |
+
},
|
| 2685 |
+
{
|
| 2686 |
+
"epoch": 0.79,
|
| 2687 |
+
"grad_norm": 0.8661462068557739,
|
| 2688 |
+
"learning_rate": 0.000010601196556252737,
|
| 2689 |
+
"loss": 0.2679,
|
| 2690 |
+
"step": 38300
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 0.79,
|
| 2694 |
+
"grad_norm": 0.9956786632537842,
|
| 2695 |
+
"learning_rate": 0.000010496966917512665,
|
| 2696 |
+
"loss": 0.2622,
|
| 2697 |
+
"step": 38400
|
| 2698 |
+
},
|
| 2699 |
+
{
|
| 2700 |
+
"epoch": 0.79,
|
| 2701 |
+
"grad_norm": 1.3166755437850952,
|
| 2702 |
+
"learning_rate": 0.000010392737278772593,
|
| 2703 |
+
"loss": 0.2595,
|
| 2704 |
+
"step": 38500
|
| 2705 |
+
},
|
| 2706 |
+
{
|
| 2707 |
+
"epoch": 0.8,
|
| 2708 |
+
"grad_norm": 1.0491749048233032,
|
| 2709 |
+
"learning_rate": 0.00001028850764003252,
|
| 2710 |
+
"loss": 0.2682,
|
| 2711 |
+
"step": 38600
|
| 2712 |
+
},
|
| 2713 |
+
{
|
| 2714 |
+
"epoch": 0.8,
|
| 2715 |
+
"grad_norm": 1.0107011795043945,
|
| 2716 |
+
"learning_rate": 0.000010184278001292448,
|
| 2717 |
+
"loss": 0.2544,
|
| 2718 |
+
"step": 38700
|
| 2719 |
+
},
|
| 2720 |
+
{
|
| 2721 |
+
"epoch": 0.8,
|
| 2722 |
+
"grad_norm": 1.0187588930130005,
|
| 2723 |
+
"learning_rate": 0.000010080048362552376,
|
| 2724 |
+
"loss": 0.2611,
|
| 2725 |
+
"step": 38800
|
| 2726 |
+
},
|
| 2727 |
+
{
|
| 2728 |
+
"epoch": 0.8,
|
| 2729 |
+
"grad_norm": 0.8718214631080627,
|
| 2730 |
+
"learning_rate": 0.000009975818723812304,
|
| 2731 |
+
"loss": 0.2567,
|
| 2732 |
+
"step": 38900
|
| 2733 |
+
},
|
| 2734 |
+
{
|
| 2735 |
+
"epoch": 0.8,
|
| 2736 |
+
"grad_norm": 0.7450688481330872,
|
| 2737 |
+
"learning_rate": 0.000009871589085072232,
|
| 2738 |
+
"loss": 0.2642,
|
| 2739 |
+
"step": 39000
|
| 2740 |
+
},
|
| 2741 |
+
{
|
| 2742 |
+
"epoch": 0.81,
|
| 2743 |
+
"grad_norm": 1.0107316970825195,
|
| 2744 |
+
"learning_rate": 0.00000976735944633216,
|
| 2745 |
+
"loss": 0.2613,
|
| 2746 |
+
"step": 39100
|
| 2747 |
+
},
|
| 2748 |
+
{
|
| 2749 |
+
"epoch": 0.81,
|
| 2750 |
+
"grad_norm": 0.8210121393203735,
|
| 2751 |
+
"learning_rate": 0.000009663129807592088,
|
| 2752 |
+
"loss": 0.2768,
|
| 2753 |
+
"step": 39200
|
| 2754 |
+
},
|
| 2755 |
+
{
|
| 2756 |
+
"epoch": 0.81,
|
| 2757 |
+
"grad_norm": 0.8526451587677002,
|
| 2758 |
+
"learning_rate": 0.000009558900168852016,
|
| 2759 |
+
"loss": 0.2473,
|
| 2760 |
+
"step": 39300
|
| 2761 |
+
},
|
| 2762 |
+
{
|
| 2763 |
+
"epoch": 0.81,
|
| 2764 |
+
"grad_norm": 0.7720713019371033,
|
| 2765 |
+
"learning_rate": 0.000009454670530111944,
|
| 2766 |
+
"loss": 0.2715,
|
| 2767 |
+
"step": 39400
|
| 2768 |
+
},
|
| 2769 |
+
{
|
| 2770 |
+
"epoch": 0.81,
|
| 2771 |
+
"grad_norm": 0.8657224774360657,
|
| 2772 |
+
"learning_rate": 0.000009350440891371872,
|
| 2773 |
+
"loss": 0.2612,
|
| 2774 |
+
"step": 39500
|
| 2775 |
+
},
|
| 2776 |
+
{
|
| 2777 |
+
"epoch": 0.82,
|
| 2778 |
+
"grad_norm": 0.9144985675811768,
|
| 2779 |
+
"learning_rate": 0.0000092462112526318,
|
| 2780 |
+
"loss": 0.2648,
|
| 2781 |
+
"step": 39600
|
| 2782 |
+
},
|
| 2783 |
+
{
|
| 2784 |
+
"epoch": 0.82,
|
| 2785 |
+
"grad_norm": 1.0658880472183228,
|
| 2786 |
+
"learning_rate": 0.000009141981613891727,
|
| 2787 |
+
"loss": 0.2661,
|
| 2788 |
+
"step": 39700
|
| 2789 |
+
},
|
| 2790 |
+
{
|
| 2791 |
+
"epoch": 0.82,
|
| 2792 |
+
"grad_norm": 0.9904933571815491,
|
| 2793 |
+
"learning_rate": 0.000009037751975151655,
|
| 2794 |
+
"loss": 0.2513,
|
| 2795 |
+
"step": 39800
|
| 2796 |
+
},
|
| 2797 |
+
{
|
| 2798 |
+
"epoch": 0.82,
|
| 2799 |
+
"grad_norm": 0.9688336253166199,
|
| 2800 |
+
"learning_rate": 0.000008933522336411583,
|
| 2801 |
+
"loss": 0.2622,
|
| 2802 |
+
"step": 39900
|
| 2803 |
+
},
|
| 2804 |
+
{
|
| 2805 |
+
"epoch": 0.83,
|
| 2806 |
+
"grad_norm": 0.6780712604522705,
|
| 2807 |
+
"learning_rate": 0.000008829292697671511,
|
| 2808 |
+
"loss": 0.2665,
|
| 2809 |
+
"step": 40000
|
| 2810 |
+
},
|
| 2811 |
+
{
|
| 2812 |
+
"epoch": 0.83,
|
| 2813 |
+
"grad_norm": 0.9881765246391296,
|
| 2814 |
+
"learning_rate": 0.000008725063058931437,
|
| 2815 |
+
"loss": 0.2486,
|
| 2816 |
+
"step": 40100
|
| 2817 |
+
},
|
| 2818 |
+
{
|
| 2819 |
+
"epoch": 0.83,
|
| 2820 |
+
"grad_norm": 0.9511051774024963,
|
| 2821 |
+
"learning_rate": 0.000008620833420191365,
|
| 2822 |
+
"loss": 0.2655,
|
| 2823 |
+
"step": 40200
|
| 2824 |
+
},
|
| 2825 |
+
{
|
| 2826 |
+
"epoch": 0.83,
|
| 2827 |
+
"grad_norm": 0.8047460317611694,
|
| 2828 |
+
"learning_rate": 0.000008516603781451293,
|
| 2829 |
+
"loss": 0.262,
|
| 2830 |
+
"step": 40300
|
| 2831 |
+
},
|
| 2832 |
+
{
|
| 2833 |
+
"epoch": 0.83,
|
| 2834 |
+
"grad_norm": 1.0902456045150757,
|
| 2835 |
+
"learning_rate": 0.000008412374142711221,
|
| 2836 |
+
"loss": 0.2681,
|
| 2837 |
+
"step": 40400
|
| 2838 |
+
},
|
| 2839 |
+
{
|
| 2840 |
+
"epoch": 0.84,
|
| 2841 |
+
"grad_norm": 0.9015824794769287,
|
| 2842 |
+
"learning_rate": 0.000008308144503971149,
|
| 2843 |
+
"loss": 0.2689,
|
| 2844 |
+
"step": 40500
|
| 2845 |
+
},
|
| 2846 |
+
{
|
| 2847 |
+
"epoch": 0.84,
|
| 2848 |
+
"grad_norm": 0.9453705549240112,
|
| 2849 |
+
"learning_rate": 0.000008203914865231077,
|
| 2850 |
+
"loss": 0.2693,
|
| 2851 |
+
"step": 40600
|
| 2852 |
+
},
|
| 2853 |
+
{
|
| 2854 |
+
"epoch": 0.84,
|
| 2855 |
+
"grad_norm": 0.8557701706886292,
|
| 2856 |
+
"learning_rate": 0.000008099685226491005,
|
| 2857 |
+
"loss": 0.2753,
|
| 2858 |
+
"step": 40700
|
| 2859 |
+
},
|
| 2860 |
+
{
|
| 2861 |
+
"epoch": 0.84,
|
| 2862 |
+
"grad_norm": 1.022003173828125,
|
| 2863 |
+
"learning_rate": 0.000007995455587750933,
|
| 2864 |
+
"loss": 0.2646,
|
| 2865 |
+
"step": 40800
|
| 2866 |
+
},
|
| 2867 |
+
{
|
| 2868 |
+
"epoch": 0.84,
|
| 2869 |
+
"grad_norm": 0.9399078488349915,
|
| 2870 |
+
"learning_rate": 0.00000789122594901086,
|
| 2871 |
+
"loss": 0.2671,
|
| 2872 |
+
"step": 40900
|
| 2873 |
+
},
|
| 2874 |
+
{
|
| 2875 |
+
"epoch": 0.85,
|
| 2876 |
+
"grad_norm": 1.1249728202819824,
|
| 2877 |
+
"learning_rate": 0.000007786996310270788,
|
| 2878 |
+
"loss": 0.2655,
|
| 2879 |
+
"step": 41000
|
| 2880 |
+
},
|
| 2881 |
+
{
|
| 2882 |
+
"epoch": 0.85,
|
| 2883 |
+
"grad_norm": 1.2603721618652344,
|
| 2884 |
+
"learning_rate": 0.000007682766671530716,
|
| 2885 |
+
"loss": 0.262,
|
| 2886 |
+
"step": 41100
|
| 2887 |
+
},
|
| 2888 |
+
{
|
| 2889 |
+
"epoch": 0.85,
|
| 2890 |
+
"grad_norm": 0.6507100462913513,
|
| 2891 |
+
"learning_rate": 0.000007578537032790644,
|
| 2892 |
+
"loss": 0.2648,
|
| 2893 |
+
"step": 41200
|
| 2894 |
+
},
|
| 2895 |
+
{
|
| 2896 |
+
"epoch": 0.85,
|
| 2897 |
+
"grad_norm": 1.0216801166534424,
|
| 2898 |
+
"learning_rate": 0.000007474307394050572,
|
| 2899 |
+
"loss": 0.2727,
|
| 2900 |
+
"step": 41300
|
| 2901 |
+
},
|
| 2902 |
+
{
|
| 2903 |
+
"epoch": 0.85,
|
| 2904 |
+
"grad_norm": 0.8479890823364258,
|
| 2905 |
+
"learning_rate": 0.0000073700777553105,
|
| 2906 |
+
"loss": 0.2644,
|
| 2907 |
+
"step": 41400
|
| 2908 |
+
},
|
| 2909 |
+
{
|
| 2910 |
+
"epoch": 0.86,
|
| 2911 |
+
"grad_norm": 0.953072726726532,
|
| 2912 |
+
"learning_rate": 0.000007265848116570428,
|
| 2913 |
+
"loss": 0.2601,
|
| 2914 |
+
"step": 41500
|
| 2915 |
+
},
|
| 2916 |
+
{
|
| 2917 |
+
"epoch": 0.86,
|
| 2918 |
+
"grad_norm": 0.7130771279335022,
|
| 2919 |
+
"learning_rate": 0.000007161618477830357,
|
| 2920 |
+
"loss": 0.2634,
|
| 2921 |
+
"step": 41600
|
| 2922 |
+
},
|
| 2923 |
+
{
|
| 2924 |
+
"epoch": 0.86,
|
| 2925 |
+
"grad_norm": 1.1879602670669556,
|
| 2926 |
+
"learning_rate": 0.000007057388839090285,
|
| 2927 |
+
"loss": 0.2588,
|
| 2928 |
+
"step": 41700
|
| 2929 |
+
},
|
| 2930 |
+
{
|
| 2931 |
+
"epoch": 0.86,
|
| 2932 |
+
"grad_norm": 0.9556657075881958,
|
| 2933 |
+
"learning_rate": 0.0000069531592003502125,
|
| 2934 |
+
"loss": 0.2673,
|
| 2935 |
+
"step": 41800
|
| 2936 |
+
},
|
| 2937 |
+
{
|
| 2938 |
+
"epoch": 0.86,
|
| 2939 |
+
"grad_norm": 1.1229112148284912,
|
| 2940 |
+
"learning_rate": 0.0000068489295616101404,
|
| 2941 |
+
"loss": 0.264,
|
| 2942 |
+
"step": 41900
|
| 2943 |
+
},
|
| 2944 |
+
{
|
| 2945 |
+
"epoch": 0.87,
|
| 2946 |
+
"grad_norm": 1.2128034830093384,
|
| 2947 |
+
"learning_rate": 0.000006744699922870068,
|
| 2948 |
+
"loss": 0.2626,
|
| 2949 |
+
"step": 42000
|
| 2950 |
+
},
|
| 2951 |
+
{
|
| 2952 |
+
"epoch": 0.87,
|
| 2953 |
+
"grad_norm": 0.9225343465805054,
|
| 2954 |
+
"learning_rate": 0.000006640470284129996,
|
| 2955 |
+
"loss": 0.2713,
|
| 2956 |
+
"step": 42100
|
| 2957 |
+
},
|
| 2958 |
+
{
|
| 2959 |
+
"epoch": 0.87,
|
| 2960 |
+
"grad_norm": 1.0495810508728027,
|
| 2961 |
+
"learning_rate": 0.000006536240645389924,
|
| 2962 |
+
"loss": 0.2739,
|
| 2963 |
+
"step": 42200
|
| 2964 |
+
},
|
| 2965 |
+
{
|
| 2966 |
+
"epoch": 0.87,
|
| 2967 |
+
"grad_norm": 0.5960795879364014,
|
| 2968 |
+
"learning_rate": 0.000006432011006649852,
|
| 2969 |
+
"loss": 0.2629,
|
| 2970 |
+
"step": 42300
|
| 2971 |
+
},
|
| 2972 |
+
{
|
| 2973 |
+
"epoch": 0.87,
|
| 2974 |
+
"grad_norm": 0.653361439704895,
|
| 2975 |
+
"learning_rate": 0.00000632778136790978,
|
| 2976 |
+
"loss": 0.2585,
|
| 2977 |
+
"step": 42400
|
| 2978 |
+
},
|
| 2979 |
+
{
|
| 2980 |
+
"epoch": 0.88,
|
| 2981 |
+
"grad_norm": 0.8684765100479126,
|
| 2982 |
+
"learning_rate": 0.000006223551729169707,
|
| 2983 |
+
"loss": 0.2426,
|
| 2984 |
+
"step": 42500
|
| 2985 |
+
},
|
| 2986 |
+
{
|
| 2987 |
+
"epoch": 0.88,
|
| 2988 |
+
"grad_norm": 0.5730326771736145,
|
| 2989 |
+
"learning_rate": 0.000006119322090429635,
|
| 2990 |
+
"loss": 0.2537,
|
| 2991 |
+
"step": 42600
|
| 2992 |
+
},
|
| 2993 |
+
{
|
| 2994 |
+
"epoch": 0.88,
|
| 2995 |
+
"grad_norm": 0.8172651529312134,
|
| 2996 |
+
"learning_rate": 0.000006015092451689563,
|
| 2997 |
+
"loss": 0.2553,
|
| 2998 |
+
"step": 42700
|
| 2999 |
+
},
|
| 3000 |
+
{
|
| 3001 |
+
"epoch": 0.88,
|
| 3002 |
+
"grad_norm": 0.9099143147468567,
|
| 3003 |
+
"learning_rate": 0.000005910862812949491,
|
| 3004 |
+
"loss": 0.2569,
|
| 3005 |
+
"step": 42800
|
| 3006 |
+
},
|
| 3007 |
+
{
|
| 3008 |
+
"epoch": 0.89,
|
| 3009 |
+
"grad_norm": 0.9645257592201233,
|
| 3010 |
+
"learning_rate": 0.000005806633174209419,
|
| 3011 |
+
"loss": 0.2689,
|
| 3012 |
+
"step": 42900
|
| 3013 |
+
},
|
| 3014 |
+
{
|
| 3015 |
+
"epoch": 0.89,
|
| 3016 |
+
"grad_norm": 0.6569193005561829,
|
| 3017 |
+
"learning_rate": 0.0000057024035354693465,
|
| 3018 |
+
"loss": 0.2662,
|
| 3019 |
+
"step": 43000
|
| 3020 |
+
},
|
| 3021 |
+
{
|
| 3022 |
+
"epoch": 0.89,
|
| 3023 |
+
"grad_norm": 1.2637701034545898,
|
| 3024 |
+
"learning_rate": 0.000005598173896729274,
|
| 3025 |
+
"loss": 0.2874,
|
| 3026 |
+
"step": 43100
|
| 3027 |
+
},
|
| 3028 |
+
{
|
| 3029 |
+
"epoch": 0.89,
|
| 3030 |
+
"grad_norm": 1.0651448965072632,
|
| 3031 |
+
"learning_rate": 0.000005493944257989202,
|
| 3032 |
+
"loss": 0.255,
|
| 3033 |
+
"step": 43200
|
| 3034 |
+
},
|
| 3035 |
+
{
|
| 3036 |
+
"epoch": 0.89,
|
| 3037 |
+
"grad_norm": 1.1091278791427612,
|
| 3038 |
+
"learning_rate": 0.000005389714619249129,
|
| 3039 |
+
"loss": 0.283,
|
| 3040 |
+
"step": 43300
|
| 3041 |
+
},
|
| 3042 |
+
{
|
| 3043 |
+
"epoch": 0.9,
|
| 3044 |
+
"grad_norm": 0.9137535095214844,
|
| 3045 |
+
"learning_rate": 0.000005285484980509057,
|
| 3046 |
+
"loss": 0.2674,
|
| 3047 |
+
"step": 43400
|
| 3048 |
+
},
|
| 3049 |
+
{
|
| 3050 |
+
"epoch": 0.9,
|
| 3051 |
+
"grad_norm": 0.8751276731491089,
|
| 3052 |
+
"learning_rate": 0.000005181255341768985,
|
| 3053 |
+
"loss": 0.2638,
|
| 3054 |
+
"step": 43500
|
| 3055 |
+
},
|
| 3056 |
+
{
|
| 3057 |
+
"epoch": 0.9,
|
| 3058 |
+
"grad_norm": 0.9242710471153259,
|
| 3059 |
+
"learning_rate": 0.000005077025703028913,
|
| 3060 |
+
"loss": 0.2582,
|
| 3061 |
+
"step": 43600
|
| 3062 |
+
},
|
| 3063 |
+
{
|
| 3064 |
+
"epoch": 0.9,
|
| 3065 |
+
"grad_norm": 0.7860730886459351,
|
| 3066 |
+
"learning_rate": 0.000004972796064288841,
|
| 3067 |
+
"loss": 0.2557,
|
| 3068 |
+
"step": 43700
|
| 3069 |
+
},
|
| 3070 |
+
{
|
| 3071 |
+
"epoch": 0.9,
|
| 3072 |
+
"grad_norm": 0.9240506291389465,
|
| 3073 |
+
"learning_rate": 0.000004868566425548769,
|
| 3074 |
+
"loss": 0.2591,
|
| 3075 |
+
"step": 43800
|
| 3076 |
+
},
|
| 3077 |
+
{
|
| 3078 |
+
"epoch": 0.91,
|
| 3079 |
+
"grad_norm": 1.0865675210952759,
|
| 3080 |
+
"learning_rate": 0.000004764336786808697,
|
| 3081 |
+
"loss": 0.2511,
|
| 3082 |
+
"step": 43900
|
| 3083 |
+
},
|
| 3084 |
+
{
|
| 3085 |
+
"epoch": 0.91,
|
| 3086 |
+
"grad_norm": 1.0534336566925049,
|
| 3087 |
+
"learning_rate": 0.0000046601071480686255,
|
| 3088 |
+
"loss": 0.2651,
|
| 3089 |
+
"step": 44000
|
| 3090 |
+
},
|
| 3091 |
+
{
|
| 3092 |
+
"epoch": 0.91,
|
| 3093 |
+
"grad_norm": 0.9301556348800659,
|
| 3094 |
+
"learning_rate": 0.000004555877509328553,
|
| 3095 |
+
"loss": 0.2414,
|
| 3096 |
+
"step": 44100
|
| 3097 |
+
},
|
| 3098 |
+
{
|
| 3099 |
+
"epoch": 0.91,
|
| 3100 |
+
"grad_norm": 0.9242545366287231,
|
| 3101 |
+
"learning_rate": 0.000004451647870588481,
|
| 3102 |
+
"loss": 0.2636,
|
| 3103 |
+
"step": 44200
|
| 3104 |
+
},
|
| 3105 |
+
{
|
| 3106 |
+
"epoch": 0.91,
|
| 3107 |
+
"grad_norm": 0.7292113304138184,
|
| 3108 |
+
"learning_rate": 0.000004347418231848409,
|
| 3109 |
+
"loss": 0.2672,
|
| 3110 |
+
"step": 44300
|
| 3111 |
+
},
|
| 3112 |
+
{
|
| 3113 |
+
"epoch": 0.92,
|
| 3114 |
+
"grad_norm": 1.2003198862075806,
|
| 3115 |
+
"learning_rate": 0.000004243188593108337,
|
| 3116 |
+
"loss": 0.2661,
|
| 3117 |
+
"step": 44400
|
| 3118 |
+
},
|
| 3119 |
+
{
|
| 3120 |
+
"epoch": 0.92,
|
| 3121 |
+
"grad_norm": 0.9352708458900452,
|
| 3122 |
+
"learning_rate": 0.000004138958954368265,
|
| 3123 |
+
"loss": 0.275,
|
| 3124 |
+
"step": 44500
|
| 3125 |
+
},
|
| 3126 |
+
{
|
| 3127 |
+
"epoch": 0.92,
|
| 3128 |
+
"grad_norm": 0.7767049074172974,
|
| 3129 |
+
"learning_rate": 0.000004034729315628192,
|
| 3130 |
+
"loss": 0.2709,
|
| 3131 |
+
"step": 44600
|
| 3132 |
+
},
|
| 3133 |
+
{
|
| 3134 |
+
"epoch": 0.92,
|
| 3135 |
+
"grad_norm": 0.8384830355644226,
|
| 3136 |
+
"learning_rate": 0.00000393049967688812,
|
| 3137 |
+
"loss": 0.2655,
|
| 3138 |
+
"step": 44700
|
| 3139 |
+
},
|
| 3140 |
+
{
|
| 3141 |
+
"epoch": 0.92,
|
| 3142 |
+
"grad_norm": 0.9187309145927429,
|
| 3143 |
+
"learning_rate": 0.000003826270038148048,
|
| 3144 |
+
"loss": 0.2609,
|
| 3145 |
+
"step": 44800
|
| 3146 |
+
},
|
| 3147 |
+
{
|
| 3148 |
+
"epoch": 0.93,
|
| 3149 |
+
"grad_norm": 0.8865881562232971,
|
| 3150 |
+
"learning_rate": 0.0000037220403994079758,
|
| 3151 |
+
"loss": 0.2677,
|
| 3152 |
+
"step": 44900
|
| 3153 |
+
},
|
| 3154 |
+
{
|
| 3155 |
+
"epoch": 0.93,
|
| 3156 |
+
"grad_norm": 1.0104628801345825,
|
| 3157 |
+
"learning_rate": 0.0000036178107606679037,
|
| 3158 |
+
"loss": 0.2649,
|
| 3159 |
+
"step": 45000
|
| 3160 |
+
},
|
| 3161 |
+
{
|
| 3162 |
+
"epoch": 0.93,
|
| 3163 |
+
"grad_norm": 0.8638792634010315,
|
| 3164 |
+
"learning_rate": 0.0000035135811219278316,
|
| 3165 |
+
"loss": 0.2694,
|
| 3166 |
+
"step": 45100
|
| 3167 |
+
},
|
| 3168 |
+
{
|
| 3169 |
+
"epoch": 0.93,
|
| 3170 |
+
"grad_norm": 1.0379241704940796,
|
| 3171 |
+
"learning_rate": 0.0000034093514831877595,
|
| 3172 |
+
"loss": 0.2489,
|
| 3173 |
+
"step": 45200
|
| 3174 |
+
},
|
| 3175 |
+
{
|
| 3176 |
+
"epoch": 0.93,
|
| 3177 |
+
"grad_norm": 1.123955488204956,
|
| 3178 |
+
"learning_rate": 0.0000033051218444476874,
|
| 3179 |
+
"loss": 0.2576,
|
| 3180 |
+
"step": 45300
|
| 3181 |
+
},
|
| 3182 |
+
{
|
| 3183 |
+
"epoch": 0.94,
|
| 3184 |
+
"grad_norm": 1.1420966386795044,
|
| 3185 |
+
"learning_rate": 0.0000032008922057076153,
|
| 3186 |
+
"loss": 0.2706,
|
| 3187 |
+
"step": 45400
|
| 3188 |
+
},
|
| 3189 |
+
{
|
| 3190 |
+
"epoch": 0.94,
|
| 3191 |
+
"grad_norm": 1.0519020557403564,
|
| 3192 |
+
"learning_rate": 0.0000030966625669675427,
|
| 3193 |
+
"loss": 0.2616,
|
| 3194 |
+
"step": 45500
|
| 3195 |
+
},
|
| 3196 |
+
{
|
| 3197 |
+
"epoch": 0.94,
|
| 3198 |
+
"grad_norm": 1.040588140487671,
|
| 3199 |
+
"learning_rate": 0.0000029924329282274706,
|
| 3200 |
+
"loss": 0.2546,
|
| 3201 |
+
"step": 45600
|
| 3202 |
+
},
|
| 3203 |
+
{
|
| 3204 |
+
"epoch": 0.94,
|
| 3205 |
+
"grad_norm": 0.8306300640106201,
|
| 3206 |
+
"learning_rate": 0.0000028882032894873985,
|
| 3207 |
+
"loss": 0.2683,
|
| 3208 |
+
"step": 45700
|
| 3209 |
+
},
|
| 3210 |
+
{
|
| 3211 |
+
"epoch": 0.94,
|
| 3212 |
+
"grad_norm": 0.6718655228614807,
|
| 3213 |
+
"learning_rate": 0.0000027839736507473264,
|
| 3214 |
+
"loss": 0.2483,
|
| 3215 |
+
"step": 45800
|
| 3216 |
+
},
|
| 3217 |
+
{
|
| 3218 |
+
"epoch": 0.95,
|
| 3219 |
+
"grad_norm": 0.8920142650604248,
|
| 3220 |
+
"learning_rate": 0.0000026797440120072548,
|
| 3221 |
+
"loss": 0.2612,
|
| 3222 |
+
"step": 45900
|
| 3223 |
+
},
|
| 3224 |
+
{
|
| 3225 |
+
"epoch": 0.95,
|
| 3226 |
+
"grad_norm": 1.09241783618927,
|
| 3227 |
+
"learning_rate": 0.0000025755143732671827,
|
| 3228 |
+
"loss": 0.2483,
|
| 3229 |
+
"step": 46000
|
| 3230 |
+
},
|
| 3231 |
+
{
|
| 3232 |
+
"epoch": 0.95,
|
| 3233 |
+
"grad_norm": 0.9406448602676392,
|
| 3234 |
+
"learning_rate": 0.0000024712847345271106,
|
| 3235 |
+
"loss": 0.2639,
|
| 3236 |
+
"step": 46100
|
| 3237 |
+
},
|
| 3238 |
+
{
|
| 3239 |
+
"epoch": 0.95,
|
| 3240 |
+
"grad_norm": 0.9145969748497009,
|
| 3241 |
+
"learning_rate": 0.000002367055095787038,
|
| 3242 |
+
"loss": 0.2493,
|
| 3243 |
+
"step": 46200
|
| 3244 |
+
},
|
| 3245 |
+
{
|
| 3246 |
+
"epoch": 0.96,
|
| 3247 |
+
"grad_norm": 0.722114086151123,
|
| 3248 |
+
"learning_rate": 0.000002262825457046966,
|
| 3249 |
+
"loss": 0.266,
|
| 3250 |
+
"step": 46300
|
| 3251 |
+
},
|
| 3252 |
+
{
|
| 3253 |
+
"epoch": 0.96,
|
| 3254 |
+
"grad_norm": 0.7077659964561462,
|
| 3255 |
+
"learning_rate": 0.000002158595818306894,
|
| 3256 |
+
"loss": 0.2712,
|
| 3257 |
+
"step": 46400
|
| 3258 |
+
},
|
| 3259 |
+
{
|
| 3260 |
+
"epoch": 0.96,
|
| 3261 |
+
"grad_norm": 0.7417329549789429,
|
| 3262 |
+
"learning_rate": 0.0000020543661795668217,
|
| 3263 |
+
"loss": 0.2435,
|
| 3264 |
+
"step": 46500
|
| 3265 |
+
},
|
| 3266 |
+
{
|
| 3267 |
+
"epoch": 0.96,
|
| 3268 |
+
"grad_norm": 0.8229231238365173,
|
| 3269 |
+
"learning_rate": 0.0000019501365408267496,
|
| 3270 |
+
"loss": 0.2647,
|
| 3271 |
+
"step": 46600
|
| 3272 |
+
},
|
| 3273 |
+
{
|
| 3274 |
+
"epoch": 0.96,
|
| 3275 |
+
"grad_norm": 0.49395614862442017,
|
| 3276 |
+
"learning_rate": 0.0000018459069020866775,
|
| 3277 |
+
"loss": 0.2508,
|
| 3278 |
+
"step": 46700
|
| 3279 |
+
},
|
| 3280 |
+
{
|
| 3281 |
+
"epoch": 0.97,
|
| 3282 |
+
"grad_norm": 1.0020008087158203,
|
| 3283 |
+
"learning_rate": 0.0000017416772633466052,
|
| 3284 |
+
"loss": 0.2505,
|
| 3285 |
+
"step": 46800
|
| 3286 |
+
},
|
| 3287 |
+
{
|
| 3288 |
+
"epoch": 0.97,
|
| 3289 |
+
"grad_norm": 0.8525140881538391,
|
| 3290 |
+
"learning_rate": 0.0000016374476246065331,
|
| 3291 |
+
"loss": 0.2518,
|
| 3292 |
+
"step": 46900
|
| 3293 |
+
},
|
| 3294 |
+
{
|
| 3295 |
+
"epoch": 0.97,
|
| 3296 |
+
"grad_norm": 0.8383646011352539,
|
| 3297 |
+
"learning_rate": 0.000001533217985866461,
|
| 3298 |
+
"loss": 0.2768,
|
| 3299 |
+
"step": 47000
|
| 3300 |
+
},
|
| 3301 |
+
{
|
| 3302 |
+
"epoch": 0.97,
|
| 3303 |
+
"grad_norm": 0.586855947971344,
|
| 3304 |
+
"learning_rate": 0.000001428988347126389,
|
| 3305 |
+
"loss": 0.2379,
|
| 3306 |
+
"step": 47100
|
| 3307 |
+
},
|
| 3308 |
+
{
|
| 3309 |
+
"epoch": 0.97,
|
| 3310 |
+
"grad_norm": 0.7666946053504944,
|
| 3311 |
+
"learning_rate": 0.0000013247587083863168,
|
| 3312 |
+
"loss": 0.2501,
|
| 3313 |
+
"step": 47200
|
| 3314 |
+
},
|
| 3315 |
+
{
|
| 3316 |
+
"epoch": 0.98,
|
| 3317 |
+
"grad_norm": 1.0712182521820068,
|
| 3318 |
+
"learning_rate": 0.0000012205290696462445,
|
| 3319 |
+
"loss": 0.2539,
|
| 3320 |
+
"step": 47300
|
| 3321 |
+
},
|
| 3322 |
+
{
|
| 3323 |
+
"epoch": 0.98,
|
| 3324 |
+
"grad_norm": 0.9935363531112671,
|
| 3325 |
+
"learning_rate": 0.0000011162994309061726,
|
| 3326 |
+
"loss": 0.2603,
|
| 3327 |
+
"step": 47400
|
| 3328 |
+
},
|
| 3329 |
+
{
|
| 3330 |
+
"epoch": 0.98,
|
| 3331 |
+
"grad_norm": 0.8433282971382141,
|
| 3332 |
+
"learning_rate": 0.0000010120697921661005,
|
| 3333 |
+
"loss": 0.261,
|
| 3334 |
+
"step": 47500
|
| 3335 |
+
},
|
| 3336 |
+
{
|
| 3337 |
+
"epoch": 0.98,
|
| 3338 |
+
"grad_norm": 0.942387044429779,
|
| 3339 |
+
"learning_rate": 9.078401534260282e-7,
|
| 3340 |
+
"loss": 0.2702,
|
| 3341 |
+
"step": 47600
|
| 3342 |
+
},
|
| 3343 |
+
{
|
| 3344 |
+
"epoch": 0.98,
|
| 3345 |
+
"grad_norm": 1.199331521987915,
|
| 3346 |
+
"learning_rate": 8.036105146859561e-7,
|
| 3347 |
+
"loss": 0.2452,
|
| 3348 |
+
"step": 47700
|
| 3349 |
+
},
|
| 3350 |
+
{
|
| 3351 |
+
"epoch": 0.99,
|
| 3352 |
+
"grad_norm": 0.9567368626594543,
|
| 3353 |
+
"learning_rate": 6.993808759458839e-7,
|
| 3354 |
+
"loss": 0.2691,
|
| 3355 |
+
"step": 47800
|
| 3356 |
+
},
|
| 3357 |
+
{
|
| 3358 |
+
"epoch": 0.99,
|
| 3359 |
+
"grad_norm": 1.0403625965118408,
|
| 3360 |
+
"learning_rate": 5.951512372058119e-7,
|
| 3361 |
+
"loss": 0.2702,
|
| 3362 |
+
"step": 47900
|
| 3363 |
+
},
|
| 3364 |
+
{
|
| 3365 |
+
"epoch": 0.99,
|
| 3366 |
+
"grad_norm": 0.9573351740837097,
|
| 3367 |
+
"learning_rate": 4.909215984657397e-7,
|
| 3368 |
+
"loss": 0.2539,
|
| 3369 |
+
"step": 48000
|
| 3370 |
+
}
|
| 3371 |
+
],
|
| 3372 |
+
"logging_steps": 100,
|
| 3373 |
+
"max_steps": 48471,
|
| 3374 |
+
"num_input_tokens_seen": 0,
|
| 3375 |
+
"num_train_epochs": 1,
|
| 3376 |
+
"save_steps": 16000,
|
| 3377 |
+
"total_flos": 104021020901376000,
|
| 3378 |
+
"train_batch_size": 4,
|
| 3379 |
+
"trial_name": null,
|
| 3380 |
+
"trial_params": null
|
| 3381 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9d53b21a61a3b91d2383fb7034fe689c71b017de8d9c9fd63254e0b967fc9f55
|
| 3 |
+
size 4920
|