Duplicate from ryfye181/distilbert_endian_classifier
Browse filesCo-authored-by: Ryan Feinberg <ryfye181@users.noreply.huggingface.co>
- .gitattributes +35 -0
- README.md +3 -0
- config.json +33 -0
- model.safetensors +3 -0
- optimizer.pt +3 -0
- rng_state.pth +3 -0
- scheduler.pt +3 -0
- special_tokens_map.json +7 -0
- tokenizer.json +0 -0
- tokenizer_config.json +55 -0
- trainer_state.json +1272 -0
- training_args.bin +3 -0
- vocab.txt +0 -0
.gitattributes
ADDED
|
@@ -0,0 +1,35 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
*.7z filter=lfs diff=lfs merge=lfs -text
|
| 2 |
+
*.arrow filter=lfs diff=lfs merge=lfs -text
|
| 3 |
+
*.bin filter=lfs diff=lfs merge=lfs -text
|
| 4 |
+
*.bz2 filter=lfs diff=lfs merge=lfs -text
|
| 5 |
+
*.ckpt filter=lfs diff=lfs merge=lfs -text
|
| 6 |
+
*.ftz filter=lfs diff=lfs merge=lfs -text
|
| 7 |
+
*.gz filter=lfs diff=lfs merge=lfs -text
|
| 8 |
+
*.h5 filter=lfs diff=lfs merge=lfs -text
|
| 9 |
+
*.joblib filter=lfs diff=lfs merge=lfs -text
|
| 10 |
+
*.lfs.* filter=lfs diff=lfs merge=lfs -text
|
| 11 |
+
*.mlmodel filter=lfs diff=lfs merge=lfs -text
|
| 12 |
+
*.model filter=lfs diff=lfs merge=lfs -text
|
| 13 |
+
*.msgpack filter=lfs diff=lfs merge=lfs -text
|
| 14 |
+
*.npy filter=lfs diff=lfs merge=lfs -text
|
| 15 |
+
*.npz filter=lfs diff=lfs merge=lfs -text
|
| 16 |
+
*.onnx filter=lfs diff=lfs merge=lfs -text
|
| 17 |
+
*.ot filter=lfs diff=lfs merge=lfs -text
|
| 18 |
+
*.parquet filter=lfs diff=lfs merge=lfs -text
|
| 19 |
+
*.pb filter=lfs diff=lfs merge=lfs -text
|
| 20 |
+
*.pickle filter=lfs diff=lfs merge=lfs -text
|
| 21 |
+
*.pkl filter=lfs diff=lfs merge=lfs -text
|
| 22 |
+
*.pt filter=lfs diff=lfs merge=lfs -text
|
| 23 |
+
*.pth filter=lfs diff=lfs merge=lfs -text
|
| 24 |
+
*.rar filter=lfs diff=lfs merge=lfs -text
|
| 25 |
+
*.safetensors filter=lfs diff=lfs merge=lfs -text
|
| 26 |
+
saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
| 27 |
+
*.tar.* filter=lfs diff=lfs merge=lfs -text
|
| 28 |
+
*.tar filter=lfs diff=lfs merge=lfs -text
|
| 29 |
+
*.tflite filter=lfs diff=lfs merge=lfs -text
|
| 30 |
+
*.tgz filter=lfs diff=lfs merge=lfs -text
|
| 31 |
+
*.wasm filter=lfs diff=lfs merge=lfs -text
|
| 32 |
+
*.xz filter=lfs diff=lfs merge=lfs -text
|
| 33 |
+
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
+
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
+
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
README.md
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
---
|
config.json
ADDED
|
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "distilbert/distilbert-base-uncased",
|
| 3 |
+
"activation": "gelu",
|
| 4 |
+
"architectures": [
|
| 5 |
+
"DistilBertForSequenceClassification"
|
| 6 |
+
],
|
| 7 |
+
"attention_dropout": 0.1,
|
| 8 |
+
"dim": 768,
|
| 9 |
+
"dropout": 0.1,
|
| 10 |
+
"hidden_dim": 3072,
|
| 11 |
+
"id2label": {
|
| 12 |
+
"0": "little",
|
| 13 |
+
"1": "big"
|
| 14 |
+
},
|
| 15 |
+
"initializer_range": 0.02,
|
| 16 |
+
"label2id": {
|
| 17 |
+
"big": 1,
|
| 18 |
+
"little": 0
|
| 19 |
+
},
|
| 20 |
+
"max_position_embeddings": 512,
|
| 21 |
+
"model_type": "distilbert",
|
| 22 |
+
"n_heads": 12,
|
| 23 |
+
"n_layers": 6,
|
| 24 |
+
"pad_token_id": 0,
|
| 25 |
+
"problem_type": "single_label_classification",
|
| 26 |
+
"qa_dropout": 0.1,
|
| 27 |
+
"seq_classif_dropout": 0.2,
|
| 28 |
+
"sinusoidal_pos_embds": false,
|
| 29 |
+
"tie_weights_": true,
|
| 30 |
+
"torch_dtype": "float32",
|
| 31 |
+
"transformers_version": "4.44.2",
|
| 32 |
+
"vocab_size": 30522
|
| 33 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9e7d3d29886f038d178fabd0e4eb9f13081acd7f6f28eb054de7b7c3bfd89d0a
|
| 3 |
+
size 267832560
|
optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:695397bde9ef2331a119ef01173219ab6ca5784714fcd1524cf4c3e7000f8f6d
|
| 3 |
+
size 535727290
|
rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b8ac14987a980c2804eb34cbf8345e363f4efb17b37b41ea4c78dee05d33cd23
|
| 3 |
+
size 14244
|
scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9184cbd8cd347b5270f85f1e3c20b290e8df12b3fb4bd94a3f49537906e512c5
|
| 3 |
+
size 1064
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cls_token": "[CLS]",
|
| 3 |
+
"mask_token": "[MASK]",
|
| 4 |
+
"pad_token": "[PAD]",
|
| 5 |
+
"sep_token": "[SEP]",
|
| 6 |
+
"unk_token": "[UNK]"
|
| 7 |
+
}
|
tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "[PAD]",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"100": {
|
| 12 |
+
"content": "[UNK]",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"101": {
|
| 20 |
+
"content": "[CLS]",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"102": {
|
| 28 |
+
"content": "[SEP]",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"103": {
|
| 36 |
+
"content": "[MASK]",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"clean_up_tokenization_spaces": true,
|
| 45 |
+
"cls_token": "[CLS]",
|
| 46 |
+
"do_lower_case": true,
|
| 47 |
+
"mask_token": "[MASK]",
|
| 48 |
+
"model_max_length": 512,
|
| 49 |
+
"pad_token": "[PAD]",
|
| 50 |
+
"sep_token": "[SEP]",
|
| 51 |
+
"strip_accents": null,
|
| 52 |
+
"tokenize_chinese_chars": true,
|
| 53 |
+
"tokenizer_class": "DistilBertTokenizer",
|
| 54 |
+
"unk_token": "[UNK]"
|
| 55 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,1272 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.10040388455255037,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 88500,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0005672535850426574,
|
| 13 |
+
"grad_norm": 3.387953996658325,
|
| 14 |
+
"learning_rate": 1.9994327464149576e-05,
|
| 15 |
+
"loss": 0.4217,
|
| 16 |
+
"step": 500
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.0011345071700853148,
|
| 20 |
+
"grad_norm": 7.831711769104004,
|
| 21 |
+
"learning_rate": 1.998865492829915e-05,
|
| 22 |
+
"loss": 0.3872,
|
| 23 |
+
"step": 1000
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.0017017607551279724,
|
| 27 |
+
"grad_norm": 0.5793220400810242,
|
| 28 |
+
"learning_rate": 1.998298239244872e-05,
|
| 29 |
+
"loss": 0.3817,
|
| 30 |
+
"step": 1500
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.0022690143401706297,
|
| 34 |
+
"grad_norm": 1.234447956085205,
|
| 35 |
+
"learning_rate": 1.9977309856598294e-05,
|
| 36 |
+
"loss": 0.3843,
|
| 37 |
+
"step": 2000
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.0028362679252132874,
|
| 41 |
+
"grad_norm": 1.1251826286315918,
|
| 42 |
+
"learning_rate": 1.997163732074787e-05,
|
| 43 |
+
"loss": 0.3823,
|
| 44 |
+
"step": 2500
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.0034035215102559448,
|
| 48 |
+
"grad_norm": 5.509366512298584,
|
| 49 |
+
"learning_rate": 1.9965964784897443e-05,
|
| 50 |
+
"loss": 0.3689,
|
| 51 |
+
"step": 3000
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.0039707750952986025,
|
| 55 |
+
"grad_norm": 1.2918850183486938,
|
| 56 |
+
"learning_rate": 1.9960292249047017e-05,
|
| 57 |
+
"loss": 0.368,
|
| 58 |
+
"step": 3500
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.004538028680341259,
|
| 62 |
+
"grad_norm": 0.7386884689331055,
|
| 63 |
+
"learning_rate": 1.9954619713196587e-05,
|
| 64 |
+
"loss": 0.3588,
|
| 65 |
+
"step": 4000
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.005105282265383917,
|
| 69 |
+
"grad_norm": 0.40995702147483826,
|
| 70 |
+
"learning_rate": 1.994894717734616e-05,
|
| 71 |
+
"loss": 0.3648,
|
| 72 |
+
"step": 4500
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.005672535850426575,
|
| 76 |
+
"grad_norm": 1.1950393915176392,
|
| 77 |
+
"learning_rate": 1.9943274641495735e-05,
|
| 78 |
+
"loss": 0.3636,
|
| 79 |
+
"step": 5000
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.006239789435469232,
|
| 83 |
+
"grad_norm": 0.33484068512916565,
|
| 84 |
+
"learning_rate": 1.993760210564531e-05,
|
| 85 |
+
"loss": 0.3727,
|
| 86 |
+
"step": 5500
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.0068070430205118895,
|
| 90 |
+
"grad_norm": 0.6621952056884766,
|
| 91 |
+
"learning_rate": 1.9931929569794883e-05,
|
| 92 |
+
"loss": 0.3645,
|
| 93 |
+
"step": 6000
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.007374296605554547,
|
| 97 |
+
"grad_norm": 0.3775445222854614,
|
| 98 |
+
"learning_rate": 1.9926257033944457e-05,
|
| 99 |
+
"loss": 0.3463,
|
| 100 |
+
"step": 6500
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.007941550190597205,
|
| 104 |
+
"grad_norm": 3.5549304485321045,
|
| 105 |
+
"learning_rate": 1.9920584498094028e-05,
|
| 106 |
+
"loss": 0.3486,
|
| 107 |
+
"step": 7000
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.008508803775639862,
|
| 111 |
+
"grad_norm": 0.8449244499206543,
|
| 112 |
+
"learning_rate": 1.9914911962243602e-05,
|
| 113 |
+
"loss": 0.3516,
|
| 114 |
+
"step": 7500
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.009076057360682519,
|
| 118 |
+
"grad_norm": 0.4936882257461548,
|
| 119 |
+
"learning_rate": 1.9909239426393176e-05,
|
| 120 |
+
"loss": 0.3726,
|
| 121 |
+
"step": 8000
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.009643310945725177,
|
| 125 |
+
"grad_norm": 0.3846686780452728,
|
| 126 |
+
"learning_rate": 1.990356689054275e-05,
|
| 127 |
+
"loss": 0.3559,
|
| 128 |
+
"step": 8500
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.010210564530767834,
|
| 132 |
+
"grad_norm": 1.0206539630889893,
|
| 133 |
+
"learning_rate": 1.9897894354692324e-05,
|
| 134 |
+
"loss": 0.3465,
|
| 135 |
+
"step": 9000
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.010777818115810491,
|
| 139 |
+
"grad_norm": 0.32130077481269836,
|
| 140 |
+
"learning_rate": 1.9892221818841898e-05,
|
| 141 |
+
"loss": 0.356,
|
| 142 |
+
"step": 9500
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.01134507170085315,
|
| 146 |
+
"grad_norm": 1.2418744564056396,
|
| 147 |
+
"learning_rate": 1.988654928299147e-05,
|
| 148 |
+
"loss": 0.3536,
|
| 149 |
+
"step": 10000
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.011912325285895807,
|
| 153 |
+
"grad_norm": 3.4250869750976562,
|
| 154 |
+
"learning_rate": 1.9880876747141043e-05,
|
| 155 |
+
"loss": 0.34,
|
| 156 |
+
"step": 10500
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.012479578870938464,
|
| 160 |
+
"grad_norm": 0.8089951872825623,
|
| 161 |
+
"learning_rate": 1.9875204211290617e-05,
|
| 162 |
+
"loss": 0.3542,
|
| 163 |
+
"step": 11000
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.013046832455981122,
|
| 167 |
+
"grad_norm": 2.089003562927246,
|
| 168 |
+
"learning_rate": 1.986953167544019e-05,
|
| 169 |
+
"loss": 0.3639,
|
| 170 |
+
"step": 11500
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.013614086041023779,
|
| 174 |
+
"grad_norm": 0.32313957810401917,
|
| 175 |
+
"learning_rate": 1.9863859139589762e-05,
|
| 176 |
+
"loss": 0.3437,
|
| 177 |
+
"step": 12000
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.014181339626066436,
|
| 181 |
+
"grad_norm": 0.3242214620113373,
|
| 182 |
+
"learning_rate": 1.985818660373934e-05,
|
| 183 |
+
"loss": 0.3498,
|
| 184 |
+
"step": 12500
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.014748593211109095,
|
| 188 |
+
"grad_norm": 0.5132392644882202,
|
| 189 |
+
"learning_rate": 1.985251406788891e-05,
|
| 190 |
+
"loss": 0.3476,
|
| 191 |
+
"step": 13000
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.015315846796151751,
|
| 195 |
+
"grad_norm": 1.0964300632476807,
|
| 196 |
+
"learning_rate": 1.9846841532038484e-05,
|
| 197 |
+
"loss": 0.3345,
|
| 198 |
+
"step": 13500
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.01588310038119441,
|
| 202 |
+
"grad_norm": 0.9808000326156616,
|
| 203 |
+
"learning_rate": 1.9841168996188058e-05,
|
| 204 |
+
"loss": 0.3337,
|
| 205 |
+
"step": 14000
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.016450353966237067,
|
| 209 |
+
"grad_norm": 21.7490291595459,
|
| 210 |
+
"learning_rate": 1.9835496460337632e-05,
|
| 211 |
+
"loss": 0.3562,
|
| 212 |
+
"step": 14500
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.017017607551279724,
|
| 216 |
+
"grad_norm": 0.37766537070274353,
|
| 217 |
+
"learning_rate": 1.9829823924487203e-05,
|
| 218 |
+
"loss": 0.3343,
|
| 219 |
+
"step": 15000
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.01758486113632238,
|
| 223 |
+
"grad_norm": 1.8476216793060303,
|
| 224 |
+
"learning_rate": 1.982415138863678e-05,
|
| 225 |
+
"loss": 0.3381,
|
| 226 |
+
"step": 15500
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.018152114721365038,
|
| 230 |
+
"grad_norm": 0.5553542971611023,
|
| 231 |
+
"learning_rate": 1.981847885278635e-05,
|
| 232 |
+
"loss": 0.3497,
|
| 233 |
+
"step": 16000
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.018719368306407698,
|
| 237 |
+
"grad_norm": 0.21085184812545776,
|
| 238 |
+
"learning_rate": 1.9812806316935925e-05,
|
| 239 |
+
"loss": 0.3472,
|
| 240 |
+
"step": 16500
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.019286621891450355,
|
| 244 |
+
"grad_norm": 9.589337348937988,
|
| 245 |
+
"learning_rate": 1.98071337810855e-05,
|
| 246 |
+
"loss": 0.3449,
|
| 247 |
+
"step": 17000
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.01985387547649301,
|
| 251 |
+
"grad_norm": 0.9183555245399475,
|
| 252 |
+
"learning_rate": 1.9801461245235073e-05,
|
| 253 |
+
"loss": 0.3339,
|
| 254 |
+
"step": 17500
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.02042112906153567,
|
| 258 |
+
"grad_norm": 23.96393585205078,
|
| 259 |
+
"learning_rate": 1.9795788709384643e-05,
|
| 260 |
+
"loss": 0.3385,
|
| 261 |
+
"step": 18000
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.020988382646578325,
|
| 265 |
+
"grad_norm": 1.359579086303711,
|
| 266 |
+
"learning_rate": 1.9790116173534217e-05,
|
| 267 |
+
"loss": 0.3428,
|
| 268 |
+
"step": 18500
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.021555636231620982,
|
| 272 |
+
"grad_norm": 0.5926823019981384,
|
| 273 |
+
"learning_rate": 1.978444363768379e-05,
|
| 274 |
+
"loss": 0.3444,
|
| 275 |
+
"step": 19000
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.022122889816663643,
|
| 279 |
+
"grad_norm": 0.7892150282859802,
|
| 280 |
+
"learning_rate": 1.9778771101833366e-05,
|
| 281 |
+
"loss": 0.3321,
|
| 282 |
+
"step": 19500
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.0226901434017063,
|
| 286 |
+
"grad_norm": 0.3750669062137604,
|
| 287 |
+
"learning_rate": 1.9773098565982936e-05,
|
| 288 |
+
"loss": 0.3548,
|
| 289 |
+
"step": 20000
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.023257396986748956,
|
| 293 |
+
"grad_norm": 4.070109844207764,
|
| 294 |
+
"learning_rate": 1.9767426030132514e-05,
|
| 295 |
+
"loss": 0.3522,
|
| 296 |
+
"step": 20500
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.023824650571791613,
|
| 300 |
+
"grad_norm": 0.38032421469688416,
|
| 301 |
+
"learning_rate": 1.9761753494282084e-05,
|
| 302 |
+
"loss": 0.3474,
|
| 303 |
+
"step": 21000
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.02439190415683427,
|
| 307 |
+
"grad_norm": 21.677780151367188,
|
| 308 |
+
"learning_rate": 1.975608095843166e-05,
|
| 309 |
+
"loss": 0.3453,
|
| 310 |
+
"step": 21500
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.024959157741876927,
|
| 314 |
+
"grad_norm": 5.641119480133057,
|
| 315 |
+
"learning_rate": 1.9750408422581232e-05,
|
| 316 |
+
"loss": 0.3449,
|
| 317 |
+
"step": 22000
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.025526411326919587,
|
| 321 |
+
"grad_norm": 25.08061981201172,
|
| 322 |
+
"learning_rate": 1.9744735886730806e-05,
|
| 323 |
+
"loss": 0.345,
|
| 324 |
+
"step": 22500
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.026093664911962244,
|
| 328 |
+
"grad_norm": 0.43347591161727905,
|
| 329 |
+
"learning_rate": 1.9739063350880377e-05,
|
| 330 |
+
"loss": 0.3366,
|
| 331 |
+
"step": 23000
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.0266609184970049,
|
| 335 |
+
"grad_norm": 0.31947267055511475,
|
| 336 |
+
"learning_rate": 1.9733390815029955e-05,
|
| 337 |
+
"loss": 0.3358,
|
| 338 |
+
"step": 23500
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.027228172082047558,
|
| 342 |
+
"grad_norm": 1.0318926572799683,
|
| 343 |
+
"learning_rate": 1.9727718279179525e-05,
|
| 344 |
+
"loss": 0.3322,
|
| 345 |
+
"step": 24000
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.027795425667090215,
|
| 349 |
+
"grad_norm": 1.8301944732666016,
|
| 350 |
+
"learning_rate": 1.97220457433291e-05,
|
| 351 |
+
"loss": 0.3396,
|
| 352 |
+
"step": 24500
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.028362679252132872,
|
| 356 |
+
"grad_norm": 0.3236497640609741,
|
| 357 |
+
"learning_rate": 1.9716373207478673e-05,
|
| 358 |
+
"loss": 0.3516,
|
| 359 |
+
"step": 25000
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.028929932837175532,
|
| 363 |
+
"grad_norm": 2.550765037536621,
|
| 364 |
+
"learning_rate": 1.9710700671628247e-05,
|
| 365 |
+
"loss": 0.3478,
|
| 366 |
+
"step": 25500
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.02949718642221819,
|
| 370 |
+
"grad_norm": 0.537734866142273,
|
| 371 |
+
"learning_rate": 1.9705028135777818e-05,
|
| 372 |
+
"loss": 0.3263,
|
| 373 |
+
"step": 26000
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.030064440007260846,
|
| 377 |
+
"grad_norm": 0.5279741883277893,
|
| 378 |
+
"learning_rate": 1.9699355599927392e-05,
|
| 379 |
+
"loss": 0.3268,
|
| 380 |
+
"step": 26500
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.030631693592303503,
|
| 384 |
+
"grad_norm": 4.043364524841309,
|
| 385 |
+
"learning_rate": 1.9693683064076966e-05,
|
| 386 |
+
"loss": 0.3476,
|
| 387 |
+
"step": 27000
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.03119894717734616,
|
| 391 |
+
"grad_norm": 0.4588943421840668,
|
| 392 |
+
"learning_rate": 1.968801052822654e-05,
|
| 393 |
+
"loss": 0.335,
|
| 394 |
+
"step": 27500
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.03176620076238882,
|
| 398 |
+
"grad_norm": 0.31232479214668274,
|
| 399 |
+
"learning_rate": 1.9682337992376114e-05,
|
| 400 |
+
"loss": 0.327,
|
| 401 |
+
"step": 28000
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.03233345434743148,
|
| 405 |
+
"grad_norm": 0.3791950047016144,
|
| 406 |
+
"learning_rate": 1.9676665456525688e-05,
|
| 407 |
+
"loss": 0.3347,
|
| 408 |
+
"step": 28500
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.032900707932474134,
|
| 412 |
+
"grad_norm": 0.11353936791419983,
|
| 413 |
+
"learning_rate": 1.967099292067526e-05,
|
| 414 |
+
"loss": 0.332,
|
| 415 |
+
"step": 29000
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.03346796151751679,
|
| 419 |
+
"grad_norm": 0.5194895267486572,
|
| 420 |
+
"learning_rate": 1.9665320384824833e-05,
|
| 421 |
+
"loss": 0.3322,
|
| 422 |
+
"step": 29500
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.03403521510255945,
|
| 426 |
+
"grad_norm": 0.7947537302970886,
|
| 427 |
+
"learning_rate": 1.9659647848974407e-05,
|
| 428 |
+
"loss": 0.3307,
|
| 429 |
+
"step": 30000
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.034602468687602105,
|
| 433 |
+
"grad_norm": 0.3675363063812256,
|
| 434 |
+
"learning_rate": 1.965397531312398e-05,
|
| 435 |
+
"loss": 0.3482,
|
| 436 |
+
"step": 30500
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.03516972227264476,
|
| 440 |
+
"grad_norm": 0.27730804681777954,
|
| 441 |
+
"learning_rate": 1.9648302777273555e-05,
|
| 442 |
+
"loss": 0.3409,
|
| 443 |
+
"step": 31000
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.03573697585768742,
|
| 447 |
+
"grad_norm": 0.44526398181915283,
|
| 448 |
+
"learning_rate": 1.964263024142313e-05,
|
| 449 |
+
"loss": 0.3315,
|
| 450 |
+
"step": 31500
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.036304229442730075,
|
| 454 |
+
"grad_norm": 0.28707778453826904,
|
| 455 |
+
"learning_rate": 1.96369577055727e-05,
|
| 456 |
+
"loss": 0.3236,
|
| 457 |
+
"step": 32000
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.03687148302777273,
|
| 461 |
+
"grad_norm": 0.7148706912994385,
|
| 462 |
+
"learning_rate": 1.9631285169722274e-05,
|
| 463 |
+
"loss": 0.332,
|
| 464 |
+
"step": 32500
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.037438736612815396,
|
| 468 |
+
"grad_norm": 132.02745056152344,
|
| 469 |
+
"learning_rate": 1.9625612633871848e-05,
|
| 470 |
+
"loss": 0.3325,
|
| 471 |
+
"step": 33000
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.03800599019785805,
|
| 475 |
+
"grad_norm": 0.3043186366558075,
|
| 476 |
+
"learning_rate": 1.9619940098021422e-05,
|
| 477 |
+
"loss": 0.3292,
|
| 478 |
+
"step": 33500
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.03857324378290071,
|
| 482 |
+
"grad_norm": 13.414630889892578,
|
| 483 |
+
"learning_rate": 1.9614267562170996e-05,
|
| 484 |
+
"loss": 0.3338,
|
| 485 |
+
"step": 34000
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.03914049736794337,
|
| 489 |
+
"grad_norm": 0.20249101519584656,
|
| 490 |
+
"learning_rate": 1.9608595026320567e-05,
|
| 491 |
+
"loss": 0.3362,
|
| 492 |
+
"step": 34500
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.03970775095298602,
|
| 496 |
+
"grad_norm": 0.8548006415367126,
|
| 497 |
+
"learning_rate": 1.960292249047014e-05,
|
| 498 |
+
"loss": 0.3499,
|
| 499 |
+
"step": 35000
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.04027500453802868,
|
| 503 |
+
"grad_norm": 0.12220838665962219,
|
| 504 |
+
"learning_rate": 1.9597249954619715e-05,
|
| 505 |
+
"loss": 0.3285,
|
| 506 |
+
"step": 35500
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.04084225812307134,
|
| 510 |
+
"grad_norm": 0.29653048515319824,
|
| 511 |
+
"learning_rate": 1.959157741876929e-05,
|
| 512 |
+
"loss": 0.3454,
|
| 513 |
+
"step": 36000
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.041409511708113994,
|
| 517 |
+
"grad_norm": 0.09974247217178345,
|
| 518 |
+
"learning_rate": 1.9585904882918863e-05,
|
| 519 |
+
"loss": 0.3448,
|
| 520 |
+
"step": 36500
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.04197676529315665,
|
| 524 |
+
"grad_norm": 0.23804782330989838,
|
| 525 |
+
"learning_rate": 1.9580232347068437e-05,
|
| 526 |
+
"loss": 0.3224,
|
| 527 |
+
"step": 37000
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.04254401887819931,
|
| 531 |
+
"grad_norm": 0.26832830905914307,
|
| 532 |
+
"learning_rate": 1.9574559811218007e-05,
|
| 533 |
+
"loss": 0.3163,
|
| 534 |
+
"step": 37500
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.043111272463241965,
|
| 538 |
+
"grad_norm": 3.487936019897461,
|
| 539 |
+
"learning_rate": 1.956888727536758e-05,
|
| 540 |
+
"loss": 0.3343,
|
| 541 |
+
"step": 38000
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.04367852604828463,
|
| 545 |
+
"grad_norm": 0.18137934803962708,
|
| 546 |
+
"learning_rate": 1.9563214739517155e-05,
|
| 547 |
+
"loss": 0.3375,
|
| 548 |
+
"step": 38500
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.044245779633327285,
|
| 552 |
+
"grad_norm": 0.42142942547798157,
|
| 553 |
+
"learning_rate": 1.955754220366673e-05,
|
| 554 |
+
"loss": 0.338,
|
| 555 |
+
"step": 39000
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.04481303321836994,
|
| 559 |
+
"grad_norm": 0.21193692088127136,
|
| 560 |
+
"learning_rate": 1.95518696678163e-05,
|
| 561 |
+
"loss": 0.3336,
|
| 562 |
+
"step": 39500
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.0453802868034126,
|
| 566 |
+
"grad_norm": 0.33044755458831787,
|
| 567 |
+
"learning_rate": 1.9546197131965878e-05,
|
| 568 |
+
"loss": 0.3238,
|
| 569 |
+
"step": 40000
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.045947540388455256,
|
| 573 |
+
"grad_norm": 0.25394052267074585,
|
| 574 |
+
"learning_rate": 1.9540524596115448e-05,
|
| 575 |
+
"loss": 0.3311,
|
| 576 |
+
"step": 40500
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.04651479397349791,
|
| 580 |
+
"grad_norm": 0.24994152784347534,
|
| 581 |
+
"learning_rate": 1.9534852060265022e-05,
|
| 582 |
+
"loss": 0.3173,
|
| 583 |
+
"step": 41000
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.04708204755854057,
|
| 587 |
+
"grad_norm": 4.048117637634277,
|
| 588 |
+
"learning_rate": 1.9529179524414596e-05,
|
| 589 |
+
"loss": 0.3449,
|
| 590 |
+
"step": 41500
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.04764930114358323,
|
| 594 |
+
"grad_norm": 0.7400296330451965,
|
| 595 |
+
"learning_rate": 1.952350698856417e-05,
|
| 596 |
+
"loss": 0.3474,
|
| 597 |
+
"step": 42000
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.048216554728625884,
|
| 601 |
+
"grad_norm": 0.26707279682159424,
|
| 602 |
+
"learning_rate": 1.951783445271374e-05,
|
| 603 |
+
"loss": 0.3386,
|
| 604 |
+
"step": 42500
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.04878380831366854,
|
| 608 |
+
"grad_norm": 13.8534517288208,
|
| 609 |
+
"learning_rate": 1.951216191686332e-05,
|
| 610 |
+
"loss": 0.3232,
|
| 611 |
+
"step": 43000
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.0493510618987112,
|
| 615 |
+
"grad_norm": 9.848762512207031,
|
| 616 |
+
"learning_rate": 1.950648938101289e-05,
|
| 617 |
+
"loss": 0.3431,
|
| 618 |
+
"step": 43500
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.049918315483753854,
|
| 622 |
+
"grad_norm": 0.6303707361221313,
|
| 623 |
+
"learning_rate": 1.9500816845162463e-05,
|
| 624 |
+
"loss": 0.3266,
|
| 625 |
+
"step": 44000
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.05048556906879652,
|
| 629 |
+
"grad_norm": 12.861529350280762,
|
| 630 |
+
"learning_rate": 1.9495144309312037e-05,
|
| 631 |
+
"loss": 0.3239,
|
| 632 |
+
"step": 44500
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.051052822653839175,
|
| 636 |
+
"grad_norm": 9.070085525512695,
|
| 637 |
+
"learning_rate": 1.948947177346161e-05,
|
| 638 |
+
"loss": 0.3319,
|
| 639 |
+
"step": 45000
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.05162007623888183,
|
| 643 |
+
"grad_norm": 14.648274421691895,
|
| 644 |
+
"learning_rate": 1.9483799237611182e-05,
|
| 645 |
+
"loss": 0.331,
|
| 646 |
+
"step": 45500
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.05218732982392449,
|
| 650 |
+
"grad_norm": 34.75390625,
|
| 651 |
+
"learning_rate": 1.947812670176076e-05,
|
| 652 |
+
"loss": 0.3295,
|
| 653 |
+
"step": 46000
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.052754583408967146,
|
| 657 |
+
"grad_norm": 1.1457617282867432,
|
| 658 |
+
"learning_rate": 1.947245416591033e-05,
|
| 659 |
+
"loss": 0.3236,
|
| 660 |
+
"step": 46500
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.0533218369940098,
|
| 664 |
+
"grad_norm": 1.6054667234420776,
|
| 665 |
+
"learning_rate": 1.9466781630059904e-05,
|
| 666 |
+
"loss": 0.3236,
|
| 667 |
+
"step": 47000
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.05388909057905246,
|
| 671 |
+
"grad_norm": 1.246711254119873,
|
| 672 |
+
"learning_rate": 1.9461109094209475e-05,
|
| 673 |
+
"loss": 0.332,
|
| 674 |
+
"step": 47500
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.054456344164095116,
|
| 678 |
+
"grad_norm": 0.5977410674095154,
|
| 679 |
+
"learning_rate": 1.9455436558359052e-05,
|
| 680 |
+
"loss": 0.3393,
|
| 681 |
+
"step": 48000
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.05502359774913777,
|
| 685 |
+
"grad_norm": 5.520767688751221,
|
| 686 |
+
"learning_rate": 1.9449764022508623e-05,
|
| 687 |
+
"loss": 0.3327,
|
| 688 |
+
"step": 48500
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 0.05559085133418043,
|
| 692 |
+
"grad_norm": 1.9742405414581299,
|
| 693 |
+
"learning_rate": 1.9444091486658197e-05,
|
| 694 |
+
"loss": 0.3444,
|
| 695 |
+
"step": 49000
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 0.05615810491922309,
|
| 699 |
+
"grad_norm": 0.8282870650291443,
|
| 700 |
+
"learning_rate": 1.943841895080777e-05,
|
| 701 |
+
"loss": 0.3394,
|
| 702 |
+
"step": 49500
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 0.056725358504265744,
|
| 706 |
+
"grad_norm": 2.033897638320923,
|
| 707 |
+
"learning_rate": 1.9432746414957345e-05,
|
| 708 |
+
"loss": 0.3363,
|
| 709 |
+
"step": 50000
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 0.05729261208930841,
|
| 713 |
+
"grad_norm": 1.0143098831176758,
|
| 714 |
+
"learning_rate": 1.9427073879106916e-05,
|
| 715 |
+
"loss": 0.3196,
|
| 716 |
+
"step": 50500
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.057859865674351064,
|
| 720 |
+
"grad_norm": 1.6727173328399658,
|
| 721 |
+
"learning_rate": 1.9421401343256493e-05,
|
| 722 |
+
"loss": 0.3294,
|
| 723 |
+
"step": 51000
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 0.05842711925939372,
|
| 727 |
+
"grad_norm": 34.665252685546875,
|
| 728 |
+
"learning_rate": 1.9415728807406064e-05,
|
| 729 |
+
"loss": 0.3282,
|
| 730 |
+
"step": 51500
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 0.05899437284443638,
|
| 734 |
+
"grad_norm": 0.5767442584037781,
|
| 735 |
+
"learning_rate": 1.9410056271555638e-05,
|
| 736 |
+
"loss": 0.3327,
|
| 737 |
+
"step": 52000
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 0.059561626429479035,
|
| 741 |
+
"grad_norm": 0.6436382532119751,
|
| 742 |
+
"learning_rate": 1.9404383735705212e-05,
|
| 743 |
+
"loss": 0.3137,
|
| 744 |
+
"step": 52500
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 0.06012888001452169,
|
| 748 |
+
"grad_norm": 0.4447130858898163,
|
| 749 |
+
"learning_rate": 1.9398711199854786e-05,
|
| 750 |
+
"loss": 0.328,
|
| 751 |
+
"step": 53000
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 0.06069613359956435,
|
| 755 |
+
"grad_norm": 1.7067936658859253,
|
| 756 |
+
"learning_rate": 1.9393038664004356e-05,
|
| 757 |
+
"loss": 0.3268,
|
| 758 |
+
"step": 53500
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.061263387184607006,
|
| 762 |
+
"grad_norm": 97.85061645507812,
|
| 763 |
+
"learning_rate": 1.938736612815393e-05,
|
| 764 |
+
"loss": 0.3195,
|
| 765 |
+
"step": 54000
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 0.06183064076964966,
|
| 769 |
+
"grad_norm": 0.6765246391296387,
|
| 770 |
+
"learning_rate": 1.9381693592303505e-05,
|
| 771 |
+
"loss": 0.3233,
|
| 772 |
+
"step": 54500
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 0.06239789435469232,
|
| 776 |
+
"grad_norm": 71.85704803466797,
|
| 777 |
+
"learning_rate": 1.937602105645308e-05,
|
| 778 |
+
"loss": 0.3317,
|
| 779 |
+
"step": 55000
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 0.06296514793973498,
|
| 783 |
+
"grad_norm": 1.1403242349624634,
|
| 784 |
+
"learning_rate": 1.9370348520602653e-05,
|
| 785 |
+
"loss": 0.3349,
|
| 786 |
+
"step": 55500
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 0.06353240152477764,
|
| 790 |
+
"grad_norm": 22.765140533447266,
|
| 791 |
+
"learning_rate": 1.9364675984752227e-05,
|
| 792 |
+
"loss": 0.3192,
|
| 793 |
+
"step": 56000
|
| 794 |
+
},
|
| 795 |
+
{
|
| 796 |
+
"epoch": 0.0640996551098203,
|
| 797 |
+
"grad_norm": 0.7139065861701965,
|
| 798 |
+
"learning_rate": 1.9359003448901797e-05,
|
| 799 |
+
"loss": 0.3162,
|
| 800 |
+
"step": 56500
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.06466690869486295,
|
| 804 |
+
"grad_norm": 0.6095991134643555,
|
| 805 |
+
"learning_rate": 1.935333091305137e-05,
|
| 806 |
+
"loss": 0.3173,
|
| 807 |
+
"step": 57000
|
| 808 |
+
},
|
| 809 |
+
{
|
| 810 |
+
"epoch": 0.06523416227990561,
|
| 811 |
+
"grad_norm": 0.4620860815048218,
|
| 812 |
+
"learning_rate": 1.9347658377200945e-05,
|
| 813 |
+
"loss": 0.3405,
|
| 814 |
+
"step": 57500
|
| 815 |
+
},
|
| 816 |
+
{
|
| 817 |
+
"epoch": 0.06580141586494827,
|
| 818 |
+
"grad_norm": 2.4695820808410645,
|
| 819 |
+
"learning_rate": 1.934198584135052e-05,
|
| 820 |
+
"loss": 0.3263,
|
| 821 |
+
"step": 58000
|
| 822 |
+
},
|
| 823 |
+
{
|
| 824 |
+
"epoch": 0.06636866944999092,
|
| 825 |
+
"grad_norm": 0.49013492465019226,
|
| 826 |
+
"learning_rate": 1.933631330550009e-05,
|
| 827 |
+
"loss": 0.327,
|
| 828 |
+
"step": 58500
|
| 829 |
+
},
|
| 830 |
+
{
|
| 831 |
+
"epoch": 0.06693592303503358,
|
| 832 |
+
"grad_norm": 0.8440238237380981,
|
| 833 |
+
"learning_rate": 1.9330640769649668e-05,
|
| 834 |
+
"loss": 0.3326,
|
| 835 |
+
"step": 59000
|
| 836 |
+
},
|
| 837 |
+
{
|
| 838 |
+
"epoch": 0.06750317662007624,
|
| 839 |
+
"grad_norm": 5.433935165405273,
|
| 840 |
+
"learning_rate": 1.9324968233799238e-05,
|
| 841 |
+
"loss": 0.3242,
|
| 842 |
+
"step": 59500
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.0680704302051189,
|
| 846 |
+
"grad_norm": 1.157573938369751,
|
| 847 |
+
"learning_rate": 1.9319295697948812e-05,
|
| 848 |
+
"loss": 0.3268,
|
| 849 |
+
"step": 60000
|
| 850 |
+
},
|
| 851 |
+
{
|
| 852 |
+
"epoch": 0.06863768379016155,
|
| 853 |
+
"grad_norm": 1.0837581157684326,
|
| 854 |
+
"learning_rate": 1.9313623162098386e-05,
|
| 855 |
+
"loss": 0.3228,
|
| 856 |
+
"step": 60500
|
| 857 |
+
},
|
| 858 |
+
{
|
| 859 |
+
"epoch": 0.06920493737520421,
|
| 860 |
+
"grad_norm": 1.0040887594223022,
|
| 861 |
+
"learning_rate": 1.930795062624796e-05,
|
| 862 |
+
"loss": 0.3115,
|
| 863 |
+
"step": 61000
|
| 864 |
+
},
|
| 865 |
+
{
|
| 866 |
+
"epoch": 0.06977219096024687,
|
| 867 |
+
"grad_norm": 1.13579523563385,
|
| 868 |
+
"learning_rate": 1.930227809039753e-05,
|
| 869 |
+
"loss": 0.3155,
|
| 870 |
+
"step": 61500
|
| 871 |
+
},
|
| 872 |
+
{
|
| 873 |
+
"epoch": 0.07033944454528952,
|
| 874 |
+
"grad_norm": 3.2068631649017334,
|
| 875 |
+
"learning_rate": 1.9296605554547105e-05,
|
| 876 |
+
"loss": 0.3132,
|
| 877 |
+
"step": 62000
|
| 878 |
+
},
|
| 879 |
+
{
|
| 880 |
+
"epoch": 0.07090669813033218,
|
| 881 |
+
"grad_norm": 0.586251974105835,
|
| 882 |
+
"learning_rate": 1.929093301869668e-05,
|
| 883 |
+
"loss": 0.3153,
|
| 884 |
+
"step": 62500
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.07147395171537484,
|
| 888 |
+
"grad_norm": 0.7550469636917114,
|
| 889 |
+
"learning_rate": 1.9285260482846253e-05,
|
| 890 |
+
"loss": 0.3283,
|
| 891 |
+
"step": 63000
|
| 892 |
+
},
|
| 893 |
+
{
|
| 894 |
+
"epoch": 0.0720412053004175,
|
| 895 |
+
"grad_norm": 1.0829483270645142,
|
| 896 |
+
"learning_rate": 1.9279587946995827e-05,
|
| 897 |
+
"loss": 0.3304,
|
| 898 |
+
"step": 63500
|
| 899 |
+
},
|
| 900 |
+
{
|
| 901 |
+
"epoch": 0.07260845888546015,
|
| 902 |
+
"grad_norm": 0.7655236124992371,
|
| 903 |
+
"learning_rate": 1.92739154111454e-05,
|
| 904 |
+
"loss": 0.326,
|
| 905 |
+
"step": 64000
|
| 906 |
+
},
|
| 907 |
+
{
|
| 908 |
+
"epoch": 0.07317571247050281,
|
| 909 |
+
"grad_norm": 0.857245922088623,
|
| 910 |
+
"learning_rate": 1.9268242875294972e-05,
|
| 911 |
+
"loss": 0.3212,
|
| 912 |
+
"step": 64500
|
| 913 |
+
},
|
| 914 |
+
{
|
| 915 |
+
"epoch": 0.07374296605554546,
|
| 916 |
+
"grad_norm": 0.4056306779384613,
|
| 917 |
+
"learning_rate": 1.9262570339444546e-05,
|
| 918 |
+
"loss": 0.32,
|
| 919 |
+
"step": 65000
|
| 920 |
+
},
|
| 921 |
+
{
|
| 922 |
+
"epoch": 0.07431021964058813,
|
| 923 |
+
"grad_norm": 10.08352279663086,
|
| 924 |
+
"learning_rate": 1.925689780359412e-05,
|
| 925 |
+
"loss": 0.3229,
|
| 926 |
+
"step": 65500
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.07487747322563079,
|
| 930 |
+
"grad_norm": 15.482842445373535,
|
| 931 |
+
"learning_rate": 1.9251225267743694e-05,
|
| 932 |
+
"loss": 0.3236,
|
| 933 |
+
"step": 66000
|
| 934 |
+
},
|
| 935 |
+
{
|
| 936 |
+
"epoch": 0.07544472681067345,
|
| 937 |
+
"grad_norm": 0.32501137256622314,
|
| 938 |
+
"learning_rate": 1.9245552731893268e-05,
|
| 939 |
+
"loss": 0.3212,
|
| 940 |
+
"step": 66500
|
| 941 |
+
},
|
| 942 |
+
{
|
| 943 |
+
"epoch": 0.0760119803957161,
|
| 944 |
+
"grad_norm": 0.5554530620574951,
|
| 945 |
+
"learning_rate": 1.9239880196042842e-05,
|
| 946 |
+
"loss": 0.3373,
|
| 947 |
+
"step": 67000
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 0.07657923398075876,
|
| 951 |
+
"grad_norm": 0.27948904037475586,
|
| 952 |
+
"learning_rate": 1.9234207660192413e-05,
|
| 953 |
+
"loss": 0.3305,
|
| 954 |
+
"step": 67500
|
| 955 |
+
},
|
| 956 |
+
{
|
| 957 |
+
"epoch": 0.07714648756580142,
|
| 958 |
+
"grad_norm": 1.6499723196029663,
|
| 959 |
+
"learning_rate": 1.9228535124341987e-05,
|
| 960 |
+
"loss": 0.3404,
|
| 961 |
+
"step": 68000
|
| 962 |
+
},
|
| 963 |
+
{
|
| 964 |
+
"epoch": 0.07771374115084408,
|
| 965 |
+
"grad_norm": 8.564773559570312,
|
| 966 |
+
"learning_rate": 1.922286258849156e-05,
|
| 967 |
+
"loss": 0.3298,
|
| 968 |
+
"step": 68500
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.07828099473588673,
|
| 972 |
+
"grad_norm": 7.0219573974609375,
|
| 973 |
+
"learning_rate": 1.9217190052641135e-05,
|
| 974 |
+
"loss": 0.3343,
|
| 975 |
+
"step": 69000
|
| 976 |
+
},
|
| 977 |
+
{
|
| 978 |
+
"epoch": 0.07884824832092939,
|
| 979 |
+
"grad_norm": 0.3555821180343628,
|
| 980 |
+
"learning_rate": 1.921151751679071e-05,
|
| 981 |
+
"loss": 0.3291,
|
| 982 |
+
"step": 69500
|
| 983 |
+
},
|
| 984 |
+
{
|
| 985 |
+
"epoch": 0.07941550190597205,
|
| 986 |
+
"grad_norm": 0.2007092386484146,
|
| 987 |
+
"learning_rate": 1.920584498094028e-05,
|
| 988 |
+
"loss": 0.3188,
|
| 989 |
+
"step": 70000
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 0.0799827554910147,
|
| 993 |
+
"grad_norm": 0.5300312638282776,
|
| 994 |
+
"learning_rate": 1.9200172445089854e-05,
|
| 995 |
+
"loss": 0.3126,
|
| 996 |
+
"step": 70500
|
| 997 |
+
},
|
| 998 |
+
{
|
| 999 |
+
"epoch": 0.08055000907605736,
|
| 1000 |
+
"grad_norm": 0.5101011991500854,
|
| 1001 |
+
"learning_rate": 1.9194499909239428e-05,
|
| 1002 |
+
"loss": 0.3205,
|
| 1003 |
+
"step": 71000
|
| 1004 |
+
},
|
| 1005 |
+
{
|
| 1006 |
+
"epoch": 0.08111726266110002,
|
| 1007 |
+
"grad_norm": 1.1713281869888306,
|
| 1008 |
+
"learning_rate": 1.9188827373389e-05,
|
| 1009 |
+
"loss": 0.3466,
|
| 1010 |
+
"step": 71500
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.08168451624614267,
|
| 1014 |
+
"grad_norm": 3.439805269241333,
|
| 1015 |
+
"learning_rate": 1.9183154837538576e-05,
|
| 1016 |
+
"loss": 0.3124,
|
| 1017 |
+
"step": 72000
|
| 1018 |
+
},
|
| 1019 |
+
{
|
| 1020 |
+
"epoch": 0.08225176983118533,
|
| 1021 |
+
"grad_norm": 0.778472363948822,
|
| 1022 |
+
"learning_rate": 1.917748230168815e-05,
|
| 1023 |
+
"loss": 0.3308,
|
| 1024 |
+
"step": 72500
|
| 1025 |
+
},
|
| 1026 |
+
{
|
| 1027 |
+
"epoch": 0.08281902341622799,
|
| 1028 |
+
"grad_norm": 0.433795303106308,
|
| 1029 |
+
"learning_rate": 1.917180976583772e-05,
|
| 1030 |
+
"loss": 0.3271,
|
| 1031 |
+
"step": 73000
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 0.08338627700127065,
|
| 1035 |
+
"grad_norm": 15.664637565612793,
|
| 1036 |
+
"learning_rate": 1.9166137229987294e-05,
|
| 1037 |
+
"loss": 0.3219,
|
| 1038 |
+
"step": 73500
|
| 1039 |
+
},
|
| 1040 |
+
{
|
| 1041 |
+
"epoch": 0.0839535305863133,
|
| 1042 |
+
"grad_norm": 0.6989821195602417,
|
| 1043 |
+
"learning_rate": 1.916046469413687e-05,
|
| 1044 |
+
"loss": 0.3095,
|
| 1045 |
+
"step": 74000
|
| 1046 |
+
},
|
| 1047 |
+
{
|
| 1048 |
+
"epoch": 0.08452078417135596,
|
| 1049 |
+
"grad_norm": 19.356645584106445,
|
| 1050 |
+
"learning_rate": 1.9154792158286443e-05,
|
| 1051 |
+
"loss": 0.3266,
|
| 1052 |
+
"step": 74500
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.08508803775639862,
|
| 1056 |
+
"grad_norm": 0.32333022356033325,
|
| 1057 |
+
"learning_rate": 1.9149119622436013e-05,
|
| 1058 |
+
"loss": 0.3219,
|
| 1059 |
+
"step": 75000
|
| 1060 |
+
},
|
| 1061 |
+
{
|
| 1062 |
+
"epoch": 0.08565529134144127,
|
| 1063 |
+
"grad_norm": 0.29713454842567444,
|
| 1064 |
+
"learning_rate": 1.914344708658559e-05,
|
| 1065 |
+
"loss": 0.3364,
|
| 1066 |
+
"step": 75500
|
| 1067 |
+
},
|
| 1068 |
+
{
|
| 1069 |
+
"epoch": 0.08622254492648393,
|
| 1070 |
+
"grad_norm": 1.616568684577942,
|
| 1071 |
+
"learning_rate": 1.913777455073516e-05,
|
| 1072 |
+
"loss": 0.3326,
|
| 1073 |
+
"step": 76000
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 0.08678979851152659,
|
| 1077 |
+
"grad_norm": 0.38876867294311523,
|
| 1078 |
+
"learning_rate": 1.9132102014884735e-05,
|
| 1079 |
+
"loss": 0.3363,
|
| 1080 |
+
"step": 76500
|
| 1081 |
+
},
|
| 1082 |
+
{
|
| 1083 |
+
"epoch": 0.08735705209656926,
|
| 1084 |
+
"grad_norm": 0.2728395164012909,
|
| 1085 |
+
"learning_rate": 1.912642947903431e-05,
|
| 1086 |
+
"loss": 0.3174,
|
| 1087 |
+
"step": 77000
|
| 1088 |
+
},
|
| 1089 |
+
{
|
| 1090 |
+
"epoch": 0.08792430568161191,
|
| 1091 |
+
"grad_norm": 0.38797077536582947,
|
| 1092 |
+
"learning_rate": 1.9120756943183883e-05,
|
| 1093 |
+
"loss": 0.3341,
|
| 1094 |
+
"step": 77500
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.08849155926665457,
|
| 1098 |
+
"grad_norm": 0.9708689451217651,
|
| 1099 |
+
"learning_rate": 1.9115084407333454e-05,
|
| 1100 |
+
"loss": 0.3258,
|
| 1101 |
+
"step": 78000
|
| 1102 |
+
},
|
| 1103 |
+
{
|
| 1104 |
+
"epoch": 0.08905881285169723,
|
| 1105 |
+
"grad_norm": 0.5084453821182251,
|
| 1106 |
+
"learning_rate": 1.910941187148303e-05,
|
| 1107 |
+
"loss": 0.318,
|
| 1108 |
+
"step": 78500
|
| 1109 |
+
},
|
| 1110 |
+
{
|
| 1111 |
+
"epoch": 0.08962606643673988,
|
| 1112 |
+
"grad_norm": 10.395822525024414,
|
| 1113 |
+
"learning_rate": 1.9103739335632602e-05,
|
| 1114 |
+
"loss": 0.3267,
|
| 1115 |
+
"step": 79000
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 0.09019332002178254,
|
| 1119 |
+
"grad_norm": 56.53432846069336,
|
| 1120 |
+
"learning_rate": 1.9098066799782176e-05,
|
| 1121 |
+
"loss": 0.3206,
|
| 1122 |
+
"step": 79500
|
| 1123 |
+
},
|
| 1124 |
+
{
|
| 1125 |
+
"epoch": 0.0907605736068252,
|
| 1126 |
+
"grad_norm": 0.5847833752632141,
|
| 1127 |
+
"learning_rate": 1.909239426393175e-05,
|
| 1128 |
+
"loss": 0.3289,
|
| 1129 |
+
"step": 80000
|
| 1130 |
+
},
|
| 1131 |
+
{
|
| 1132 |
+
"epoch": 0.09132782719186786,
|
| 1133 |
+
"grad_norm": 0.6055426597595215,
|
| 1134 |
+
"learning_rate": 1.9086721728081324e-05,
|
| 1135 |
+
"loss": 0.3099,
|
| 1136 |
+
"step": 80500
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.09189508077691051,
|
| 1140 |
+
"grad_norm": 0.13933761417865753,
|
| 1141 |
+
"learning_rate": 1.9081049192230895e-05,
|
| 1142 |
+
"loss": 0.3094,
|
| 1143 |
+
"step": 81000
|
| 1144 |
+
},
|
| 1145 |
+
{
|
| 1146 |
+
"epoch": 0.09246233436195317,
|
| 1147 |
+
"grad_norm": 3.231118679046631,
|
| 1148 |
+
"learning_rate": 1.9075376656380472e-05,
|
| 1149 |
+
"loss": 0.3292,
|
| 1150 |
+
"step": 81500
|
| 1151 |
+
},
|
| 1152 |
+
{
|
| 1153 |
+
"epoch": 0.09302958794699583,
|
| 1154 |
+
"grad_norm": 11.789403915405273,
|
| 1155 |
+
"learning_rate": 1.9069704120530043e-05,
|
| 1156 |
+
"loss": 0.3284,
|
| 1157 |
+
"step": 82000
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 0.09359684153203848,
|
| 1161 |
+
"grad_norm": 0.47824761271476746,
|
| 1162 |
+
"learning_rate": 1.9064031584679617e-05,
|
| 1163 |
+
"loss": 0.316,
|
| 1164 |
+
"step": 82500
|
| 1165 |
+
},
|
| 1166 |
+
{
|
| 1167 |
+
"epoch": 0.09416409511708114,
|
| 1168 |
+
"grad_norm": 0.5431397557258606,
|
| 1169 |
+
"learning_rate": 1.9058359048829188e-05,
|
| 1170 |
+
"loss": 0.3316,
|
| 1171 |
+
"step": 83000
|
| 1172 |
+
},
|
| 1173 |
+
{
|
| 1174 |
+
"epoch": 0.0947313487021238,
|
| 1175 |
+
"grad_norm": 0.027491087093949318,
|
| 1176 |
+
"learning_rate": 1.9052686512978765e-05,
|
| 1177 |
+
"loss": 0.3226,
|
| 1178 |
+
"step": 83500
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.09529860228716645,
|
| 1182 |
+
"grad_norm": 0.167594775557518,
|
| 1183 |
+
"learning_rate": 1.9047013977128336e-05,
|
| 1184 |
+
"loss": 0.3265,
|
| 1185 |
+
"step": 84000
|
| 1186 |
+
},
|
| 1187 |
+
{
|
| 1188 |
+
"epoch": 0.09586585587220911,
|
| 1189 |
+
"grad_norm": 25.423133850097656,
|
| 1190 |
+
"learning_rate": 1.904134144127791e-05,
|
| 1191 |
+
"loss": 0.3212,
|
| 1192 |
+
"step": 84500
|
| 1193 |
+
},
|
| 1194 |
+
{
|
| 1195 |
+
"epoch": 0.09643310945725177,
|
| 1196 |
+
"grad_norm": 8.196709632873535,
|
| 1197 |
+
"learning_rate": 1.9035668905427484e-05,
|
| 1198 |
+
"loss": 0.3355,
|
| 1199 |
+
"step": 85000
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 0.09700036304229442,
|
| 1203 |
+
"grad_norm": 0.22309711575508118,
|
| 1204 |
+
"learning_rate": 1.9029996369577058e-05,
|
| 1205 |
+
"loss": 0.3287,
|
| 1206 |
+
"step": 85500
|
| 1207 |
+
},
|
| 1208 |
+
{
|
| 1209 |
+
"epoch": 0.09756761662733708,
|
| 1210 |
+
"grad_norm": 4.482788562774658,
|
| 1211 |
+
"learning_rate": 1.902432383372663e-05,
|
| 1212 |
+
"loss": 0.3315,
|
| 1213 |
+
"step": 86000
|
| 1214 |
+
},
|
| 1215 |
+
{
|
| 1216 |
+
"epoch": 0.09813487021237974,
|
| 1217 |
+
"grad_norm": 5.861471652984619,
|
| 1218 |
+
"learning_rate": 1.9018651297876206e-05,
|
| 1219 |
+
"loss": 0.3193,
|
| 1220 |
+
"step": 86500
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 0.0987021237974224,
|
| 1224 |
+
"grad_norm": 2.7159125804901123,
|
| 1225 |
+
"learning_rate": 1.9012978762025777e-05,
|
| 1226 |
+
"loss": 0.3163,
|
| 1227 |
+
"step": 87000
|
| 1228 |
+
},
|
| 1229 |
+
{
|
| 1230 |
+
"epoch": 0.09926937738246505,
|
| 1231 |
+
"grad_norm": 0.4765207767486572,
|
| 1232 |
+
"learning_rate": 1.900730622617535e-05,
|
| 1233 |
+
"loss": 0.3125,
|
| 1234 |
+
"step": 87500
|
| 1235 |
+
},
|
| 1236 |
+
{
|
| 1237 |
+
"epoch": 0.09983663096750771,
|
| 1238 |
+
"grad_norm": 0.6516081094741821,
|
| 1239 |
+
"learning_rate": 1.9001633690324925e-05,
|
| 1240 |
+
"loss": 0.3231,
|
| 1241 |
+
"step": 88000
|
| 1242 |
+
},
|
| 1243 |
+
{
|
| 1244 |
+
"epoch": 0.10040388455255037,
|
| 1245 |
+
"grad_norm": 6.820018291473389,
|
| 1246 |
+
"learning_rate": 1.89959611544745e-05,
|
| 1247 |
+
"loss": 0.3099,
|
| 1248 |
+
"step": 88500
|
| 1249 |
+
}
|
| 1250 |
+
],
|
| 1251 |
+
"logging_steps": 500,
|
| 1252 |
+
"max_steps": 1762880,
|
| 1253 |
+
"num_input_tokens_seen": 0,
|
| 1254 |
+
"num_train_epochs": 2,
|
| 1255 |
+
"save_steps": 500,
|
| 1256 |
+
"stateful_callbacks": {
|
| 1257 |
+
"TrainerControl": {
|
| 1258 |
+
"args": {
|
| 1259 |
+
"should_epoch_stop": false,
|
| 1260 |
+
"should_evaluate": false,
|
| 1261 |
+
"should_log": false,
|
| 1262 |
+
"should_save": true,
|
| 1263 |
+
"should_training_stop": false
|
| 1264 |
+
},
|
| 1265 |
+
"attributes": {}
|
| 1266 |
+
}
|
| 1267 |
+
},
|
| 1268 |
+
"total_flos": 9.3786918248448e+16,
|
| 1269 |
+
"train_batch_size": 8,
|
| 1270 |
+
"trial_name": null,
|
| 1271 |
+
"trial_params": null
|
| 1272 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:05340782a7a34442f1242993a00d5107b9d0d51678042cf3f76657e3198e8af4
|
| 3 |
+
size 5240
|
vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|