Gaduh Hartawan
commited on
Commit
•
8751838
1
Parent(s):
b95dcf0
initial 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
- special_tokens_map.json +51 -0
- tokenizer.json +0 -0
- trainer_state.json +1330 -0
- training_args.bin +3 -0
- vocab.txt +0 -0
config.json
ADDED
|
@@ -0,0 +1,68 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "facebook/bart-large-cnn",
|
| 3 |
+
"_num_labels": 3,
|
| 4 |
+
"activation_dropout": 0.0,
|
| 5 |
+
"activation_function": "gelu",
|
| 6 |
+
"add_final_layer_norm": false,
|
| 7 |
+
"architectures": ["BartForConditionalGeneration"],
|
| 8 |
+
"attention_dropout": 0.0,
|
| 9 |
+
"bos_token_id": 0,
|
| 10 |
+
"classif_dropout": 0.0,
|
| 11 |
+
"classifier_dropout": 0.0,
|
| 12 |
+
"d_model": 1024,
|
| 13 |
+
"decoder_attention_heads": 16,
|
| 14 |
+
"decoder_ffn_dim": 4096,
|
| 15 |
+
"decoder_layerdrop": 0.0,
|
| 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": 0.0,
|
| 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": 2.0,
|
| 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": 2.0,
|
| 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": 2.0,
|
| 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:cb98f4ef30260cf3dae3226ddcb4f72d2b8275b7a7cca81cbfac92e6ac0e599d
|
| 3 |
+
size 1625422896
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0a79d260ad25ff60576a3b81ee9173c9d396400c2b5894b5a7ff7745c883095b
|
| 3 |
+
size 3250751759
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:375e35d892a6fa01592c9545b1e6b60098a97aec2909cd43f820c0cd58a54ae5
|
| 3 |
+
size 14244
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:505e140348cf428a705b42cb8aafd791b8d0b0ac55c70e82b985098cfa08a46a
|
| 3 |
+
size 1064
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "[CLS]",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "[CLS]",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "[SEP]",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "[MASK]",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "[PAD]",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "[SEP]",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "[UNK]",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1330 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 5.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 18750,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.03,
|
| 13 |
+
"grad_norm": 13.29955005645752,
|
| 14 |
+
"learning_rate": 1e-5,
|
| 15 |
+
"loss": 3.9042,
|
| 16 |
+
"step": 100
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.05,
|
| 20 |
+
"grad_norm": 1.7849366664886475,
|
| 21 |
+
"learning_rate": 2e-5,
|
| 22 |
+
"loss": 0.6647,
|
| 23 |
+
"step": 200
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.08,
|
| 27 |
+
"grad_norm": 1.459277868270874,
|
| 28 |
+
"learning_rate": 3e-5,
|
| 29 |
+
"loss": 0.5227,
|
| 30 |
+
"step": 300
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.11,
|
| 34 |
+
"grad_norm": 1.3880829811096191,
|
| 35 |
+
"learning_rate": 4e-5,
|
| 36 |
+
"loss": 0.5149,
|
| 37 |
+
"step": 400
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.13,
|
| 41 |
+
"grad_norm": 1.3338109254837036,
|
| 42 |
+
"learning_rate": 5e-5,
|
| 43 |
+
"loss": 0.4984,
|
| 44 |
+
"step": 500
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.16,
|
| 48 |
+
"grad_norm": 1.065587043762207,
|
| 49 |
+
"learning_rate": 4.972602739726028e-5,
|
| 50 |
+
"loss": 0.4847,
|
| 51 |
+
"step": 600
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.19,
|
| 55 |
+
"grad_norm": 1.0945565700531006,
|
| 56 |
+
"learning_rate": 4.945205479452055e-5,
|
| 57 |
+
"loss": 0.4615,
|
| 58 |
+
"step": 700
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.21,
|
| 62 |
+
"grad_norm": 1.0960415601730347,
|
| 63 |
+
"learning_rate": 4.917808219178082e-5,
|
| 64 |
+
"loss": 0.4504,
|
| 65 |
+
"step": 800
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.24,
|
| 69 |
+
"grad_norm": 1.0467826128005981,
|
| 70 |
+
"learning_rate": 4.89041095890411e-5,
|
| 71 |
+
"loss": 0.4415,
|
| 72 |
+
"step": 900
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.27,
|
| 76 |
+
"grad_norm": 1.0690410137176514,
|
| 77 |
+
"learning_rate": 4.863013698630137e-5,
|
| 78 |
+
"loss": 0.4497,
|
| 79 |
+
"step": 1000
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.29,
|
| 83 |
+
"grad_norm": 0.9305112957954407,
|
| 84 |
+
"learning_rate": 4.835616438356165e-5,
|
| 85 |
+
"loss": 0.4363,
|
| 86 |
+
"step": 1100
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.32,
|
| 90 |
+
"grad_norm": 0.8500822186470032,
|
| 91 |
+
"learning_rate": 4.808219178082192e-5,
|
| 92 |
+
"loss": 0.4306,
|
| 93 |
+
"step": 1200
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.35,
|
| 97 |
+
"grad_norm": 1.2137833833694458,
|
| 98 |
+
"learning_rate": 4.780821917808219e-5,
|
| 99 |
+
"loss": 0.423,
|
| 100 |
+
"step": 1300
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.37,
|
| 104 |
+
"grad_norm": 0.9352328777313232,
|
| 105 |
+
"learning_rate": 4.753424657534247e-5,
|
| 106 |
+
"loss": 0.4161,
|
| 107 |
+
"step": 1400
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.4,
|
| 111 |
+
"grad_norm": 1.1027398109436035,
|
| 112 |
+
"learning_rate": 4.726027397260274e-5,
|
| 113 |
+
"loss": 0.4083,
|
| 114 |
+
"step": 1500
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.43,
|
| 118 |
+
"grad_norm": 0.924199104309082,
|
| 119 |
+
"learning_rate": 4.698630136986302e-5,
|
| 120 |
+
"loss": 0.4254,
|
| 121 |
+
"step": 1600
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.45,
|
| 125 |
+
"grad_norm": 0.8912659883499146,
|
| 126 |
+
"learning_rate": 4.671232876712329e-5,
|
| 127 |
+
"loss": 0.3918,
|
| 128 |
+
"step": 1700
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.48,
|
| 132 |
+
"grad_norm": 1.0025393962860107,
|
| 133 |
+
"learning_rate": 4.643835616438356e-5,
|
| 134 |
+
"loss": 0.4046,
|
| 135 |
+
"step": 1800
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.51,
|
| 139 |
+
"grad_norm": 0.9362453818321228,
|
| 140 |
+
"learning_rate": 4.616438356164384e-5,
|
| 141 |
+
"loss": 0.3979,
|
| 142 |
+
"step": 1900
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.53,
|
| 146 |
+
"grad_norm": 0.8841680884361267,
|
| 147 |
+
"learning_rate": 4.589041095890411e-5,
|
| 148 |
+
"loss": 0.3938,
|
| 149 |
+
"step": 2000
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.56,
|
| 153 |
+
"grad_norm": 0.898572564125061,
|
| 154 |
+
"learning_rate": 4.561643835616439e-5,
|
| 155 |
+
"loss": 0.3974,
|
| 156 |
+
"step": 2100
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.59,
|
| 160 |
+
"grad_norm": 0.9011989831924438,
|
| 161 |
+
"learning_rate": 4.534246575342466e-5,
|
| 162 |
+
"loss": 0.3955,
|
| 163 |
+
"step": 2200
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.61,
|
| 167 |
+
"grad_norm": 0.913512110710144,
|
| 168 |
+
"learning_rate": 4.506849315068493e-5,
|
| 169 |
+
"loss": 0.3842,
|
| 170 |
+
"step": 2300
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.64,
|
| 174 |
+
"grad_norm": 0.9077229499816895,
|
| 175 |
+
"learning_rate": 4.479452054794521e-5,
|
| 176 |
+
"loss": 0.3756,
|
| 177 |
+
"step": 2400
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.67,
|
| 181 |
+
"grad_norm": 0.8107369542121887,
|
| 182 |
+
"learning_rate": 4.452054794520548e-5,
|
| 183 |
+
"loss": 0.3781,
|
| 184 |
+
"step": 2500
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.69,
|
| 188 |
+
"grad_norm": 0.962982714176178,
|
| 189 |
+
"learning_rate": 4.424657534246576e-5,
|
| 190 |
+
"loss": 0.3728,
|
| 191 |
+
"step": 2600
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.72,
|
| 195 |
+
"grad_norm": 0.9333651065826416,
|
| 196 |
+
"learning_rate": 4.3972602739726035e-5,
|
| 197 |
+
"loss": 0.3731,
|
| 198 |
+
"step": 2700
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.75,
|
| 202 |
+
"grad_norm": 0.9969388246536255,
|
| 203 |
+
"learning_rate": 4.36986301369863e-5,
|
| 204 |
+
"loss": 0.3793,
|
| 205 |
+
"step": 2800
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.77,
|
| 209 |
+
"grad_norm": 0.8959200978279114,
|
| 210 |
+
"learning_rate": 4.342465753424658e-5,
|
| 211 |
+
"loss": 0.3787,
|
| 212 |
+
"step": 2900
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.8,
|
| 216 |
+
"grad_norm": 0.8185614943504333,
|
| 217 |
+
"learning_rate": 4.3150684931506855e-5,
|
| 218 |
+
"loss": 0.3626,
|
| 219 |
+
"step": 3000
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.83,
|
| 223 |
+
"grad_norm": 1.1243007183074951,
|
| 224 |
+
"learning_rate": 4.2876712328767126e-5,
|
| 225 |
+
"loss": 0.3681,
|
| 226 |
+
"step": 3100
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.85,
|
| 230 |
+
"grad_norm": 1.0251150131225586,
|
| 231 |
+
"learning_rate": 4.2602739726027404e-5,
|
| 232 |
+
"loss": 0.3609,
|
| 233 |
+
"step": 3200
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.88,
|
| 237 |
+
"grad_norm": 0.8459119200706482,
|
| 238 |
+
"learning_rate": 4.232876712328767e-5,
|
| 239 |
+
"loss": 0.3608,
|
| 240 |
+
"step": 3300
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.91,
|
| 244 |
+
"grad_norm": 0.9018300175666809,
|
| 245 |
+
"learning_rate": 4.2054794520547946e-5,
|
| 246 |
+
"loss": 0.3779,
|
| 247 |
+
"step": 3400
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.93,
|
| 251 |
+
"grad_norm": 0.804707407951355,
|
| 252 |
+
"learning_rate": 4.1780821917808224e-5,
|
| 253 |
+
"loss": 0.3674,
|
| 254 |
+
"step": 3500
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.96,
|
| 258 |
+
"grad_norm": 0.8781819343566895,
|
| 259 |
+
"learning_rate": 4.1506849315068495e-5,
|
| 260 |
+
"loss": 0.3635,
|
| 261 |
+
"step": 3600
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.99,
|
| 265 |
+
"grad_norm": 0.9687257409095764,
|
| 266 |
+
"learning_rate": 4.123287671232877e-5,
|
| 267 |
+
"loss": 0.3617,
|
| 268 |
+
"step": 3700
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 1.01,
|
| 272 |
+
"grad_norm": 0.7628118991851807,
|
| 273 |
+
"learning_rate": 4.0958904109589044e-5,
|
| 274 |
+
"loss": 0.3352,
|
| 275 |
+
"step": 3800
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 1.04,
|
| 279 |
+
"grad_norm": 0.9802232980728149,
|
| 280 |
+
"learning_rate": 4.0684931506849315e-5,
|
| 281 |
+
"loss": 0.3244,
|
| 282 |
+
"step": 3900
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 1.07,
|
| 286 |
+
"grad_norm": 0.7366902828216553,
|
| 287 |
+
"learning_rate": 4.041095890410959e-5,
|
| 288 |
+
"loss": 0.3237,
|
| 289 |
+
"step": 4000
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 1.09,
|
| 293 |
+
"grad_norm": 0.8844860196113586,
|
| 294 |
+
"learning_rate": 4.0136986301369864e-5,
|
| 295 |
+
"loss": 0.3296,
|
| 296 |
+
"step": 4100
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 1.12,
|
| 300 |
+
"grad_norm": 0.692650556564331,
|
| 301 |
+
"learning_rate": 3.9863013698630135e-5,
|
| 302 |
+
"loss": 0.3165,
|
| 303 |
+
"step": 4200
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 1.15,
|
| 307 |
+
"grad_norm": 0.8171700239181519,
|
| 308 |
+
"learning_rate": 3.958904109589041e-5,
|
| 309 |
+
"loss": 0.323,
|
| 310 |
+
"step": 4300
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 1.17,
|
| 314 |
+
"grad_norm": 0.9350169897079468,
|
| 315 |
+
"learning_rate": 3.9315068493150684e-5,
|
| 316 |
+
"loss": 0.3259,
|
| 317 |
+
"step": 4400
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 1.2,
|
| 321 |
+
"grad_norm": 0.9551327228546143,
|
| 322 |
+
"learning_rate": 3.904109589041096e-5,
|
| 323 |
+
"loss": 0.3252,
|
| 324 |
+
"step": 4500
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 1.23,
|
| 328 |
+
"grad_norm": 0.8646096587181091,
|
| 329 |
+
"learning_rate": 3.8767123287671233e-5,
|
| 330 |
+
"loss": 0.3267,
|
| 331 |
+
"step": 4600
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 1.25,
|
| 335 |
+
"grad_norm": 0.8012389540672302,
|
| 336 |
+
"learning_rate": 3.8493150684931505e-5,
|
| 337 |
+
"loss": 0.3149,
|
| 338 |
+
"step": 4700
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 1.28,
|
| 342 |
+
"grad_norm": 0.833848774433136,
|
| 343 |
+
"learning_rate": 3.821917808219178e-5,
|
| 344 |
+
"loss": 0.3164,
|
| 345 |
+
"step": 4800
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 1.31,
|
| 349 |
+
"grad_norm": 0.7836089730262756,
|
| 350 |
+
"learning_rate": 3.7945205479452054e-5,
|
| 351 |
+
"loss": 0.3206,
|
| 352 |
+
"step": 4900
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 1.33,
|
| 356 |
+
"grad_norm": 0.8694811463356018,
|
| 357 |
+
"learning_rate": 3.767123287671233e-5,
|
| 358 |
+
"loss": 0.3187,
|
| 359 |
+
"step": 5000
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 1.36,
|
| 363 |
+
"grad_norm": 0.8749567866325378,
|
| 364 |
+
"learning_rate": 3.739726027397261e-5,
|
| 365 |
+
"loss": 0.3165,
|
| 366 |
+
"step": 5100
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 1.39,
|
| 370 |
+
"grad_norm": 0.8689484596252441,
|
| 371 |
+
"learning_rate": 3.7123287671232874e-5,
|
| 372 |
+
"loss": 0.3154,
|
| 373 |
+
"step": 5200
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 1.41,
|
| 377 |
+
"grad_norm": 0.8809706568717957,
|
| 378 |
+
"learning_rate": 3.684931506849315e-5,
|
| 379 |
+
"loss": 0.3301,
|
| 380 |
+
"step": 5300
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 1.44,
|
| 384 |
+
"grad_norm": 0.8677769899368286,
|
| 385 |
+
"learning_rate": 3.657534246575342e-5,
|
| 386 |
+
"loss": 0.3184,
|
| 387 |
+
"step": 5400
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 1.47,
|
| 391 |
+
"grad_norm": 0.8212382793426514,
|
| 392 |
+
"learning_rate": 3.63013698630137e-5,
|
| 393 |
+
"loss": 0.3181,
|
| 394 |
+
"step": 5500
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 1.49,
|
| 398 |
+
"grad_norm": 0.8636347651481628,
|
| 399 |
+
"learning_rate": 3.602739726027398e-5,
|
| 400 |
+
"loss": 0.3138,
|
| 401 |
+
"step": 5600
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 1.52,
|
| 405 |
+
"grad_norm": 0.8136293292045593,
|
| 406 |
+
"learning_rate": 3.575342465753424e-5,
|
| 407 |
+
"loss": 0.3156,
|
| 408 |
+
"step": 5700
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 1.55,
|
| 412 |
+
"grad_norm": 0.7700251936912537,
|
| 413 |
+
"learning_rate": 3.547945205479452e-5,
|
| 414 |
+
"loss": 0.3179,
|
| 415 |
+
"step": 5800
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 1.57,
|
| 419 |
+
"grad_norm": 0.7282480597496033,
|
| 420 |
+
"learning_rate": 3.52054794520548e-5,
|
| 421 |
+
"loss": 0.3188,
|
| 422 |
+
"step": 5900
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 1.6,
|
| 426 |
+
"grad_norm": 0.7657186388969421,
|
| 427 |
+
"learning_rate": 3.493150684931507e-5,
|
| 428 |
+
"loss": 0.3137,
|
| 429 |
+
"step": 6000
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 1.63,
|
| 433 |
+
"grad_norm": 0.8558144569396973,
|
| 434 |
+
"learning_rate": 3.465753424657535e-5,
|
| 435 |
+
"loss": 0.3192,
|
| 436 |
+
"step": 6100
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 1.65,
|
| 440 |
+
"grad_norm": 0.7496147751808167,
|
| 441 |
+
"learning_rate": 3.438356164383562e-5,
|
| 442 |
+
"loss": 0.3175,
|
| 443 |
+
"step": 6200
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 1.68,
|
| 447 |
+
"grad_norm": 0.9365683794021606,
|
| 448 |
+
"learning_rate": 3.410958904109589e-5,
|
| 449 |
+
"loss": 0.3124,
|
| 450 |
+
"step": 6300
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 1.71,
|
| 454 |
+
"grad_norm": 0.8127835392951965,
|
| 455 |
+
"learning_rate": 3.383561643835617e-5,
|
| 456 |
+
"loss": 0.3056,
|
| 457 |
+
"step": 6400
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 1.73,
|
| 461 |
+
"grad_norm": 0.819684624671936,
|
| 462 |
+
"learning_rate": 3.356164383561644e-5,
|
| 463 |
+
"loss": 0.3144,
|
| 464 |
+
"step": 6500
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 1.76,
|
| 468 |
+
"grad_norm": 0.7603724598884583,
|
| 469 |
+
"learning_rate": 3.328767123287672e-5,
|
| 470 |
+
"loss": 0.315,
|
| 471 |
+
"step": 6600
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 1.79,
|
| 475 |
+
"grad_norm": 0.8054817318916321,
|
| 476 |
+
"learning_rate": 3.301369863013699e-5,
|
| 477 |
+
"loss": 0.3073,
|
| 478 |
+
"step": 6700
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 1.81,
|
| 482 |
+
"grad_norm": 0.758423924446106,
|
| 483 |
+
"learning_rate": 3.273972602739726e-5,
|
| 484 |
+
"loss": 0.312,
|
| 485 |
+
"step": 6800
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 1.84,
|
| 489 |
+
"grad_norm": 0.8245046138763428,
|
| 490 |
+
"learning_rate": 3.246575342465754e-5,
|
| 491 |
+
"loss": 0.3125,
|
| 492 |
+
"step": 6900
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 1.87,
|
| 496 |
+
"grad_norm": 0.7906696796417236,
|
| 497 |
+
"learning_rate": 3.219178082191781e-5,
|
| 498 |
+
"loss": 0.3009,
|
| 499 |
+
"step": 7000
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 1.89,
|
| 503 |
+
"grad_norm": 0.8566040992736816,
|
| 504 |
+
"learning_rate": 3.1917808219178086e-5,
|
| 505 |
+
"loss": 0.3043,
|
| 506 |
+
"step": 7100
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 1.92,
|
| 510 |
+
"grad_norm": 0.7341597080230713,
|
| 511 |
+
"learning_rate": 3.164383561643836e-5,
|
| 512 |
+
"loss": 0.309,
|
| 513 |
+
"step": 7200
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 1.95,
|
| 517 |
+
"grad_norm": 0.7561280131340027,
|
| 518 |
+
"learning_rate": 3.136986301369863e-5,
|
| 519 |
+
"loss": 0.3051,
|
| 520 |
+
"step": 7300
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 1.97,
|
| 524 |
+
"grad_norm": 0.7900431156158447,
|
| 525 |
+
"learning_rate": 3.1095890410958906e-5,
|
| 526 |
+
"loss": 0.3093,
|
| 527 |
+
"step": 7400
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 2.0,
|
| 531 |
+
"grad_norm": 0.880424976348877,
|
| 532 |
+
"learning_rate": 3.082191780821918e-5,
|
| 533 |
+
"loss": 0.3058,
|
| 534 |
+
"step": 7500
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 2.03,
|
| 538 |
+
"grad_norm": 0.8830358982086182,
|
| 539 |
+
"learning_rate": 3.0547945205479455e-5,
|
| 540 |
+
"loss": 0.2673,
|
| 541 |
+
"step": 7600
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 2.05,
|
| 545 |
+
"grad_norm": 0.6983394026756287,
|
| 546 |
+
"learning_rate": 3.0273972602739726e-5,
|
| 547 |
+
"loss": 0.2739,
|
| 548 |
+
"step": 7700
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 2.08,
|
| 552 |
+
"grad_norm": 0.8467246890068054,
|
| 553 |
+
"learning_rate": 3e-5,
|
| 554 |
+
"loss": 0.2694,
|
| 555 |
+
"step": 7800
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 2.11,
|
| 559 |
+
"grad_norm": 0.8425388932228088,
|
| 560 |
+
"learning_rate": 2.9726027397260275e-5,
|
| 561 |
+
"loss": 0.2698,
|
| 562 |
+
"step": 7900
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 2.13,
|
| 566 |
+
"grad_norm": 0.6956115365028381,
|
| 567 |
+
"learning_rate": 2.945205479452055e-5,
|
| 568 |
+
"loss": 0.2616,
|
| 569 |
+
"step": 8000
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 2.16,
|
| 573 |
+
"grad_norm": 0.9649244546890259,
|
| 574 |
+
"learning_rate": 2.9178082191780824e-5,
|
| 575 |
+
"loss": 0.2763,
|
| 576 |
+
"step": 8100
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 2.19,
|
| 580 |
+
"grad_norm": 0.7081593871116638,
|
| 581 |
+
"learning_rate": 2.8904109589041095e-5,
|
| 582 |
+
"loss": 0.2683,
|
| 583 |
+
"step": 8200
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 2.21,
|
| 587 |
+
"grad_norm": 0.9411781430244446,
|
| 588 |
+
"learning_rate": 2.863013698630137e-5,
|
| 589 |
+
"loss": 0.2621,
|
| 590 |
+
"step": 8300
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 2.24,
|
| 594 |
+
"grad_norm": 0.8201924562454224,
|
| 595 |
+
"learning_rate": 2.8356164383561644e-5,
|
| 596 |
+
"loss": 0.2701,
|
| 597 |
+
"step": 8400
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 2.27,
|
| 601 |
+
"grad_norm": 0.8518856167793274,
|
| 602 |
+
"learning_rate": 2.808219178082192e-5,
|
| 603 |
+
"loss": 0.272,
|
| 604 |
+
"step": 8500
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 2.29,
|
| 608 |
+
"grad_norm": 0.8004194498062134,
|
| 609 |
+
"learning_rate": 2.7808219178082197e-5,
|
| 610 |
+
"loss": 0.267,
|
| 611 |
+
"step": 8600
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 2.32,
|
| 615 |
+
"grad_norm": 0.9312605857849121,
|
| 616 |
+
"learning_rate": 2.7534246575342465e-5,
|
| 617 |
+
"loss": 0.2632,
|
| 618 |
+
"step": 8700
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 2.35,
|
| 622 |
+
"grad_norm": 0.8414776921272278,
|
| 623 |
+
"learning_rate": 2.726027397260274e-5,
|
| 624 |
+
"loss": 0.2681,
|
| 625 |
+
"step": 8800
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 2.37,
|
| 629 |
+
"grad_norm": 0.6925989985466003,
|
| 630 |
+
"learning_rate": 2.6986301369863014e-5,
|
| 631 |
+
"loss": 0.2668,
|
| 632 |
+
"step": 8900
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 2.4,
|
| 636 |
+
"grad_norm": 0.9184579849243164,
|
| 637 |
+
"learning_rate": 2.671232876712329e-5,
|
| 638 |
+
"loss": 0.2673,
|
| 639 |
+
"step": 9000
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 2.43,
|
| 643 |
+
"grad_norm": 1.1033433675765991,
|
| 644 |
+
"learning_rate": 2.6438356164383566e-5,
|
| 645 |
+
"loss": 0.2684,
|
| 646 |
+
"step": 9100
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 2.45,
|
| 650 |
+
"grad_norm": 0.9113504886627197,
|
| 651 |
+
"learning_rate": 2.6164383561643834e-5,
|
| 652 |
+
"loss": 0.2644,
|
| 653 |
+
"step": 9200
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 2.48,
|
| 657 |
+
"grad_norm": 0.7905146479606628,
|
| 658 |
+
"learning_rate": 2.589041095890411e-5,
|
| 659 |
+
"loss": 0.2668,
|
| 660 |
+
"step": 9300
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 2.51,
|
| 664 |
+
"grad_norm": 0.6717493534088135,
|
| 665 |
+
"learning_rate": 2.5616438356164386e-5,
|
| 666 |
+
"loss": 0.271,
|
| 667 |
+
"step": 9400
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 2.53,
|
| 671 |
+
"grad_norm": 0.8438414335250854,
|
| 672 |
+
"learning_rate": 2.534246575342466e-5,
|
| 673 |
+
"loss": 0.2706,
|
| 674 |
+
"step": 9500
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 2.56,
|
| 678 |
+
"grad_norm": 0.8165556192398071,
|
| 679 |
+
"learning_rate": 2.5068493150684935e-5,
|
| 680 |
+
"loss": 0.2603,
|
| 681 |
+
"step": 9600
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 2.59,
|
| 685 |
+
"grad_norm": 0.8030436038970947,
|
| 686 |
+
"learning_rate": 2.4794520547945206e-5,
|
| 687 |
+
"loss": 0.2587,
|
| 688 |
+
"step": 9700
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 2.61,
|
| 692 |
+
"grad_norm": 0.8518214225769043,
|
| 693 |
+
"learning_rate": 2.452054794520548e-5,
|
| 694 |
+
"loss": 0.2533,
|
| 695 |
+
"step": 9800
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 2.64,
|
| 699 |
+
"grad_norm": 0.9882023930549622,
|
| 700 |
+
"learning_rate": 2.4246575342465755e-5,
|
| 701 |
+
"loss": 0.2561,
|
| 702 |
+
"step": 9900
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 2.67,
|
| 706 |
+
"grad_norm": 0.8175749182701111,
|
| 707 |
+
"learning_rate": 2.3972602739726026e-5,
|
| 708 |
+
"loss": 0.2572,
|
| 709 |
+
"step": 10000
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 2.69,
|
| 713 |
+
"grad_norm": 0.897048830986023,
|
| 714 |
+
"learning_rate": 2.36986301369863e-5,
|
| 715 |
+
"loss": 0.2587,
|
| 716 |
+
"step": 10100
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 2.72,
|
| 720 |
+
"grad_norm": 0.8218054175376892,
|
| 721 |
+
"learning_rate": 2.342465753424658e-5,
|
| 722 |
+
"loss": 0.2654,
|
| 723 |
+
"step": 10200
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 2.75,
|
| 727 |
+
"grad_norm": 0.7128798961639404,
|
| 728 |
+
"learning_rate": 2.315068493150685e-5,
|
| 729 |
+
"loss": 0.2642,
|
| 730 |
+
"step": 10300
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 2.77,
|
| 734 |
+
"grad_norm": 0.7982375621795654,
|
| 735 |
+
"learning_rate": 2.2876712328767124e-5,
|
| 736 |
+
"loss": 0.2537,
|
| 737 |
+
"step": 10400
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 2.8,
|
| 741 |
+
"grad_norm": 0.790105938911438,
|
| 742 |
+
"learning_rate": 2.2602739726027396e-5,
|
| 743 |
+
"loss": 0.2713,
|
| 744 |
+
"step": 10500
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 2.83,
|
| 748 |
+
"grad_norm": 0.7734562158584595,
|
| 749 |
+
"learning_rate": 2.2328767123287673e-5,
|
| 750 |
+
"loss": 0.2616,
|
| 751 |
+
"step": 10600
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 2.85,
|
| 755 |
+
"grad_norm": 0.8464659452438354,
|
| 756 |
+
"learning_rate": 2.2054794520547948e-5,
|
| 757 |
+
"loss": 0.2584,
|
| 758 |
+
"step": 10700
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 2.88,
|
| 762 |
+
"grad_norm": 0.7386855483055115,
|
| 763 |
+
"learning_rate": 2.178082191780822e-5,
|
| 764 |
+
"loss": 0.257,
|
| 765 |
+
"step": 10800
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 2.91,
|
| 769 |
+
"grad_norm": 0.7122279405593872,
|
| 770 |
+
"learning_rate": 2.1506849315068494e-5,
|
| 771 |
+
"loss": 0.2667,
|
| 772 |
+
"step": 10900
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 2.93,
|
| 776 |
+
"grad_norm": 0.8505749106407166,
|
| 777 |
+
"learning_rate": 2.1232876712328768e-5,
|
| 778 |
+
"loss": 0.2661,
|
| 779 |
+
"step": 11000
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 2.96,
|
| 783 |
+
"grad_norm": 0.8915577530860901,
|
| 784 |
+
"learning_rate": 2.0958904109589043e-5,
|
| 785 |
+
"loss": 0.2567,
|
| 786 |
+
"step": 11100
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 2.99,
|
| 790 |
+
"grad_norm": 0.9431042671203613,
|
| 791 |
+
"learning_rate": 2.0684931506849317e-5,
|
| 792 |
+
"loss": 0.2578,
|
| 793 |
+
"step": 11200
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 3.01,
|
| 797 |
+
"grad_norm": 0.7943726181983948,
|
| 798 |
+
"learning_rate": 2.0410958904109588e-5,
|
| 799 |
+
"loss": 0.2393,
|
| 800 |
+
"step": 11300
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 3.04,
|
| 804 |
+
"grad_norm": 0.8244442939758301,
|
| 805 |
+
"learning_rate": 2.0136986301369866e-5,
|
| 806 |
+
"loss": 0.2175,
|
| 807 |
+
"step": 11400
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 3.07,
|
| 811 |
+
"grad_norm": 0.7802647948265076,
|
| 812 |
+
"learning_rate": 1.9863013698630137e-5,
|
| 813 |
+
"loss": 0.2161,
|
| 814 |
+
"step": 11500
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 3.09,
|
| 818 |
+
"grad_norm": 1.1162070035934448,
|
| 819 |
+
"learning_rate": 1.9589041095890412e-5,
|
| 820 |
+
"loss": 0.2211,
|
| 821 |
+
"step": 11600
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 3.12,
|
| 825 |
+
"grad_norm": 1.0273113250732422,
|
| 826 |
+
"learning_rate": 1.9315068493150686e-5,
|
| 827 |
+
"loss": 0.2253,
|
| 828 |
+
"step": 11700
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 3.15,
|
| 832 |
+
"grad_norm": 1.0477781295776367,
|
| 833 |
+
"learning_rate": 1.904109589041096e-5,
|
| 834 |
+
"loss": 0.2213,
|
| 835 |
+
"step": 11800
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 3.17,
|
| 839 |
+
"grad_norm": 0.9134103655815125,
|
| 840 |
+
"learning_rate": 1.8767123287671235e-5,
|
| 841 |
+
"loss": 0.2269,
|
| 842 |
+
"step": 11900
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 3.2,
|
| 846 |
+
"grad_norm": 0.8156262636184692,
|
| 847 |
+
"learning_rate": 1.8493150684931506e-5,
|
| 848 |
+
"loss": 0.2245,
|
| 849 |
+
"step": 12000
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 3.23,
|
| 853 |
+
"grad_norm": 0.9004743695259094,
|
| 854 |
+
"learning_rate": 1.821917808219178e-5,
|
| 855 |
+
"loss": 0.2254,
|
| 856 |
+
"step": 12100
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 3.25,
|
| 860 |
+
"grad_norm": 0.8386040925979614,
|
| 861 |
+
"learning_rate": 1.7945205479452055e-5,
|
| 862 |
+
"loss": 0.2292,
|
| 863 |
+
"step": 12200
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 3.28,
|
| 867 |
+
"grad_norm": 0.9777556657791138,
|
| 868 |
+
"learning_rate": 1.767123287671233e-5,
|
| 869 |
+
"loss": 0.2213,
|
| 870 |
+
"step": 12300
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 3.31,
|
| 874 |
+
"grad_norm": 0.7827901244163513,
|
| 875 |
+
"learning_rate": 1.7397260273972604e-5,
|
| 876 |
+
"loss": 0.2174,
|
| 877 |
+
"step": 12400
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 3.33,
|
| 881 |
+
"grad_norm": 0.7424948811531067,
|
| 882 |
+
"learning_rate": 1.7123287671232875e-5,
|
| 883 |
+
"loss": 0.2199,
|
| 884 |
+
"step": 12500
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 3.36,
|
| 888 |
+
"grad_norm": 0.8807641267776489,
|
| 889 |
+
"learning_rate": 1.684931506849315e-5,
|
| 890 |
+
"loss": 0.2204,
|
| 891 |
+
"step": 12600
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 3.39,
|
| 895 |
+
"grad_norm": 0.8479088544845581,
|
| 896 |
+
"learning_rate": 1.6575342465753428e-5,
|
| 897 |
+
"loss": 0.2241,
|
| 898 |
+
"step": 12700
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 3.41,
|
| 902 |
+
"grad_norm": 0.9211342334747314,
|
| 903 |
+
"learning_rate": 1.63013698630137e-5,
|
| 904 |
+
"loss": 0.2237,
|
| 905 |
+
"step": 12800
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 3.44,
|
| 909 |
+
"grad_norm": 0.8683446645736694,
|
| 910 |
+
"learning_rate": 1.6027397260273974e-5,
|
| 911 |
+
"loss": 0.2248,
|
| 912 |
+
"step": 12900
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 3.47,
|
| 916 |
+
"grad_norm": 0.8828756213188171,
|
| 917 |
+
"learning_rate": 1.5753424657534248e-5,
|
| 918 |
+
"loss": 0.233,
|
| 919 |
+
"step": 13000
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 3.49,
|
| 923 |
+
"grad_norm": 0.9421214461326599,
|
| 924 |
+
"learning_rate": 1.5479452054794523e-5,
|
| 925 |
+
"loss": 0.2294,
|
| 926 |
+
"step": 13100
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 3.52,
|
| 930 |
+
"grad_norm": 0.765132486820221,
|
| 931 |
+
"learning_rate": 1.5205479452054797e-5,
|
| 932 |
+
"loss": 0.2277,
|
| 933 |
+
"step": 13200
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 3.55,
|
| 937 |
+
"grad_norm": 0.9406650066375732,
|
| 938 |
+
"learning_rate": 1.4931506849315068e-5,
|
| 939 |
+
"loss": 0.217,
|
| 940 |
+
"step": 13300
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 3.57,
|
| 944 |
+
"grad_norm": 1.0174639225006104,
|
| 945 |
+
"learning_rate": 1.4657534246575344e-5,
|
| 946 |
+
"loss": 0.2265,
|
| 947 |
+
"step": 13400
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 3.6,
|
| 951 |
+
"grad_norm": 0.826392412185669,
|
| 952 |
+
"learning_rate": 1.4383561643835617e-5,
|
| 953 |
+
"loss": 0.222,
|
| 954 |
+
"step": 13500
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 3.63,
|
| 958 |
+
"grad_norm": 0.9821271300315857,
|
| 959 |
+
"learning_rate": 1.4109589041095892e-5,
|
| 960 |
+
"loss": 0.2186,
|
| 961 |
+
"step": 13600
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 3.65,
|
| 965 |
+
"grad_norm": 0.8172212839126587,
|
| 966 |
+
"learning_rate": 1.3835616438356164e-5,
|
| 967 |
+
"loss": 0.2238,
|
| 968 |
+
"step": 13700
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 3.68,
|
| 972 |
+
"grad_norm": 0.8128436207771301,
|
| 973 |
+
"learning_rate": 1.3561643835616439e-5,
|
| 974 |
+
"loss": 0.2168,
|
| 975 |
+
"step": 13800
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 3.71,
|
| 979 |
+
"grad_norm": 0.8061575293540955,
|
| 980 |
+
"learning_rate": 1.3287671232876714e-5,
|
| 981 |
+
"loss": 0.2244,
|
| 982 |
+
"step": 13900
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 3.73,
|
| 986 |
+
"grad_norm": 0.8976914882659912,
|
| 987 |
+
"learning_rate": 1.3013698630136986e-5,
|
| 988 |
+
"loss": 0.2212,
|
| 989 |
+
"step": 14000
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 3.76,
|
| 993 |
+
"grad_norm": 0.9973928332328796,
|
| 994 |
+
"learning_rate": 1.273972602739726e-5,
|
| 995 |
+
"loss": 0.2248,
|
| 996 |
+
"step": 14100
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 3.79,
|
| 1000 |
+
"grad_norm": 0.8042004108428955,
|
| 1001 |
+
"learning_rate": 1.2465753424657535e-5,
|
| 1002 |
+
"loss": 0.2178,
|
| 1003 |
+
"step": 14200
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 3.81,
|
| 1007 |
+
"grad_norm": 0.8282990455627441,
|
| 1008 |
+
"learning_rate": 1.2191780821917808e-5,
|
| 1009 |
+
"loss": 0.2227,
|
| 1010 |
+
"step": 14300
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 3.84,
|
| 1014 |
+
"grad_norm": 0.6668768525123596,
|
| 1015 |
+
"learning_rate": 1.1917808219178083e-5,
|
| 1016 |
+
"loss": 0.2226,
|
| 1017 |
+
"step": 14400
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 3.87,
|
| 1021 |
+
"grad_norm": 0.7972692847251892,
|
| 1022 |
+
"learning_rate": 1.1643835616438355e-5,
|
| 1023 |
+
"loss": 0.2193,
|
| 1024 |
+
"step": 14500
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 3.89,
|
| 1028 |
+
"grad_norm": 0.7637550830841064,
|
| 1029 |
+
"learning_rate": 1.1369863013698632e-5,
|
| 1030 |
+
"loss": 0.2157,
|
| 1031 |
+
"step": 14600
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 3.92,
|
| 1035 |
+
"grad_norm": 0.8487162590026855,
|
| 1036 |
+
"learning_rate": 1.1095890410958904e-5,
|
| 1037 |
+
"loss": 0.2251,
|
| 1038 |
+
"step": 14700
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 3.95,
|
| 1042 |
+
"grad_norm": 0.8710606694221497,
|
| 1043 |
+
"learning_rate": 1.0821917808219179e-5,
|
| 1044 |
+
"loss": 0.2153,
|
| 1045 |
+
"step": 14800
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 3.97,
|
| 1049 |
+
"grad_norm": 0.8085966110229492,
|
| 1050 |
+
"learning_rate": 1.0547945205479452e-5,
|
| 1051 |
+
"loss": 0.2191,
|
| 1052 |
+
"step": 14900
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 4.0,
|
| 1056 |
+
"grad_norm": 0.94338059425354,
|
| 1057 |
+
"learning_rate": 1.0273972602739726e-5,
|
| 1058 |
+
"loss": 0.2184,
|
| 1059 |
+
"step": 15000
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 4.03,
|
| 1063 |
+
"grad_norm": 1.4945096969604492,
|
| 1064 |
+
"learning_rate": 1e-5,
|
| 1065 |
+
"loss": 0.1863,
|
| 1066 |
+
"step": 15100
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 4.05,
|
| 1070 |
+
"grad_norm": 0.9178032279014587,
|
| 1071 |
+
"learning_rate": 9.726027397260275e-6,
|
| 1072 |
+
"loss": 0.1854,
|
| 1073 |
+
"step": 15200
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 4.08,
|
| 1077 |
+
"grad_norm": 0.8616482615470886,
|
| 1078 |
+
"learning_rate": 9.452054794520548e-6,
|
| 1079 |
+
"loss": 0.1843,
|
| 1080 |
+
"step": 15300
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 4.11,
|
| 1084 |
+
"grad_norm": 0.9844592213630676,
|
| 1085 |
+
"learning_rate": 9.178082191780823e-6,
|
| 1086 |
+
"loss": 0.1909,
|
| 1087 |
+
"step": 15400
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 4.13,
|
| 1091 |
+
"grad_norm": 0.7312936186790466,
|
| 1092 |
+
"learning_rate": 8.904109589041095e-6,
|
| 1093 |
+
"loss": 0.1899,
|
| 1094 |
+
"step": 15500
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 4.16,
|
| 1098 |
+
"grad_norm": 0.9658412933349609,
|
| 1099 |
+
"learning_rate": 8.630136986301372e-6,
|
| 1100 |
+
"loss": 0.1878,
|
| 1101 |
+
"step": 15600
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 4.19,
|
| 1105 |
+
"grad_norm": 1.0498002767562866,
|
| 1106 |
+
"learning_rate": 8.356164383561644e-6,
|
| 1107 |
+
"loss": 0.1825,
|
| 1108 |
+
"step": 15700
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 4.21,
|
| 1112 |
+
"grad_norm": 0.7098029255867004,
|
| 1113 |
+
"learning_rate": 8.082191780821919e-6,
|
| 1114 |
+
"loss": 0.1864,
|
| 1115 |
+
"step": 15800
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 4.24,
|
| 1119 |
+
"grad_norm": 0.9946851134300232,
|
| 1120 |
+
"learning_rate": 7.808219178082192e-6,
|
| 1121 |
+
"loss": 0.1852,
|
| 1122 |
+
"step": 15900
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 4.27,
|
| 1126 |
+
"grad_norm": 0.9338549375534058,
|
| 1127 |
+
"learning_rate": 7.5342465753424655e-6,
|
| 1128 |
+
"loss": 0.1865,
|
| 1129 |
+
"step": 16000
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 4.29,
|
| 1133 |
+
"grad_norm": 0.8193784952163696,
|
| 1134 |
+
"learning_rate": 7.260273972602739e-6,
|
| 1135 |
+
"loss": 0.184,
|
| 1136 |
+
"step": 16100
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 4.32,
|
| 1140 |
+
"grad_norm": 0.9323195815086365,
|
| 1141 |
+
"learning_rate": 6.9863013698630145e-6,
|
| 1142 |
+
"loss": 0.1845,
|
| 1143 |
+
"step": 16200
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 4.35,
|
| 1147 |
+
"grad_norm": 0.9668224453926086,
|
| 1148 |
+
"learning_rate": 6.712328767123288e-6,
|
| 1149 |
+
"loss": 0.1911,
|
| 1150 |
+
"step": 16300
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 4.37,
|
| 1154 |
+
"grad_norm": 0.9941351413726807,
|
| 1155 |
+
"learning_rate": 6.438356164383562e-6,
|
| 1156 |
+
"loss": 0.1859,
|
| 1157 |
+
"step": 16400
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 4.4,
|
| 1161 |
+
"grad_norm": 0.9229924082756042,
|
| 1162 |
+
"learning_rate": 6.1643835616438354e-6,
|
| 1163 |
+
"loss": 0.1861,
|
| 1164 |
+
"step": 16500
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 4.43,
|
| 1168 |
+
"grad_norm": 0.8792287111282349,
|
| 1169 |
+
"learning_rate": 5.89041095890411e-6,
|
| 1170 |
+
"loss": 0.1903,
|
| 1171 |
+
"step": 16600
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 4.45,
|
| 1175 |
+
"grad_norm": 0.682725191116333,
|
| 1176 |
+
"learning_rate": 5.616438356164384e-6,
|
| 1177 |
+
"loss": 0.1822,
|
| 1178 |
+
"step": 16700
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 4.48,
|
| 1182 |
+
"grad_norm": 0.8012785315513611,
|
| 1183 |
+
"learning_rate": 5.342465753424658e-6,
|
| 1184 |
+
"loss": 0.1888,
|
| 1185 |
+
"step": 16800
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 4.51,
|
| 1189 |
+
"grad_norm": 0.7928184270858765,
|
| 1190 |
+
"learning_rate": 5.068493150684932e-6,
|
| 1191 |
+
"loss": 0.1869,
|
| 1192 |
+
"step": 16900
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 4.53,
|
| 1196 |
+
"grad_norm": 1.2073571681976318,
|
| 1197 |
+
"learning_rate": 4.7945205479452054e-6,
|
| 1198 |
+
"loss": 0.184,
|
| 1199 |
+
"step": 17000
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 4.56,
|
| 1203 |
+
"grad_norm": 0.763810396194458,
|
| 1204 |
+
"learning_rate": 4.52054794520548e-6,
|
| 1205 |
+
"loss": 0.1824,
|
| 1206 |
+
"step": 17100
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 4.59,
|
| 1210 |
+
"grad_norm": 0.8932220935821533,
|
| 1211 |
+
"learning_rate": 4.246575342465754e-6,
|
| 1212 |
+
"loss": 0.1898,
|
| 1213 |
+
"step": 17200
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 4.61,
|
| 1217 |
+
"grad_norm": 0.7250128984451294,
|
| 1218 |
+
"learning_rate": 3.972602739726028e-6,
|
| 1219 |
+
"loss": 0.1886,
|
| 1220 |
+
"step": 17300
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 4.64,
|
| 1224 |
+
"grad_norm": 1.0617702007293701,
|
| 1225 |
+
"learning_rate": 3.6986301369863018e-6,
|
| 1226 |
+
"loss": 0.1889,
|
| 1227 |
+
"step": 17400
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 4.67,
|
| 1231 |
+
"grad_norm": 0.983672022819519,
|
| 1232 |
+
"learning_rate": 3.4246575342465754e-6,
|
| 1233 |
+
"loss": 0.1871,
|
| 1234 |
+
"step": 17500
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 4.69,
|
| 1238 |
+
"grad_norm": 0.9392043352127075,
|
| 1239 |
+
"learning_rate": 3.1506849315068495e-6,
|
| 1240 |
+
"loss": 0.1869,
|
| 1241 |
+
"step": 17600
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 4.72,
|
| 1245 |
+
"grad_norm": 0.8135913014411926,
|
| 1246 |
+
"learning_rate": 2.8767123287671236e-6,
|
| 1247 |
+
"loss": 0.1861,
|
| 1248 |
+
"step": 17700
|
| 1249 |
+
},
|
| 1250 |
+
{
|
| 1251 |
+
"epoch": 4.75,
|
| 1252 |
+
"grad_norm": 0.7956686615943909,
|
| 1253 |
+
"learning_rate": 2.6027397260273973e-6,
|
| 1254 |
+
"loss": 0.1864,
|
| 1255 |
+
"step": 17800
|
| 1256 |
+
},
|
| 1257 |
+
{
|
| 1258 |
+
"epoch": 4.77,
|
| 1259 |
+
"grad_norm": 0.8956461548805237,
|
| 1260 |
+
"learning_rate": 2.3287671232876713e-6,
|
| 1261 |
+
"loss": 0.1889,
|
| 1262 |
+
"step": 17900
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 4.8,
|
| 1266 |
+
"grad_norm": 0.9515472054481506,
|
| 1267 |
+
"learning_rate": 2.054794520547945e-6,
|
| 1268 |
+
"loss": 0.1871,
|
| 1269 |
+
"step": 18000
|
| 1270 |
+
},
|
| 1271 |
+
{
|
| 1272 |
+
"epoch": 4.83,
|
| 1273 |
+
"grad_norm": 0.8886680006980896,
|
| 1274 |
+
"learning_rate": 1.7808219178082193e-6,
|
| 1275 |
+
"loss": 0.187,
|
| 1276 |
+
"step": 18100
|
| 1277 |
+
},
|
| 1278 |
+
{
|
| 1279 |
+
"epoch": 4.85,
|
| 1280 |
+
"grad_norm": 0.8525242805480957,
|
| 1281 |
+
"learning_rate": 1.5068493150684932e-6,
|
| 1282 |
+
"loss": 0.1832,
|
| 1283 |
+
"step": 18200
|
| 1284 |
+
},
|
| 1285 |
+
{
|
| 1286 |
+
"epoch": 4.88,
|
| 1287 |
+
"grad_norm": 0.9522444009780884,
|
| 1288 |
+
"learning_rate": 1.232876712328767e-6,
|
| 1289 |
+
"loss": 0.186,
|
| 1290 |
+
"step": 18300
|
| 1291 |
+
},
|
| 1292 |
+
{
|
| 1293 |
+
"epoch": 4.91,
|
| 1294 |
+
"grad_norm": 0.8611086010932922,
|
| 1295 |
+
"learning_rate": 9.589041095890411e-7,
|
| 1296 |
+
"loss": 0.1855,
|
| 1297 |
+
"step": 18400
|
| 1298 |
+
},
|
| 1299 |
+
{
|
| 1300 |
+
"epoch": 4.93,
|
| 1301 |
+
"grad_norm": 0.9658819437026978,
|
| 1302 |
+
"learning_rate": 6.849315068493151e-7,
|
| 1303 |
+
"loss": 0.177,
|
| 1304 |
+
"step": 18500
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 4.96,
|
| 1308 |
+
"grad_norm": 0.9198510646820068,
|
| 1309 |
+
"learning_rate": 4.1095890410958903e-7,
|
| 1310 |
+
"loss": 0.178,
|
| 1311 |
+
"step": 18600
|
| 1312 |
+
},
|
| 1313 |
+
{
|
| 1314 |
+
"epoch": 4.99,
|
| 1315 |
+
"grad_norm": 0.8326091766357422,
|
| 1316 |
+
"learning_rate": 1.36986301369863e-7,
|
| 1317 |
+
"loss": 0.1865,
|
| 1318 |
+
"step": 18700
|
| 1319 |
+
}
|
| 1320 |
+
],
|
| 1321 |
+
"logging_steps": 100,
|
| 1322 |
+
"max_steps": 18750,
|
| 1323 |
+
"num_input_tokens_seen": 0,
|
| 1324 |
+
"num_train_epochs": 5,
|
| 1325 |
+
"save_steps": 6250,
|
| 1326 |
+
"total_flos": 8.12664225792e16,
|
| 1327 |
+
"train_batch_size": 8,
|
| 1328 |
+
"trial_name": null,
|
| 1329 |
+
"trial_params": null
|
| 1330 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e10560e54c27682fe1af61856c09bdac5dfcb77c62ecca0494c7e6f4ba3762bd
|
| 3 |
+
size 4920
|
vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|