Commit
•
c54d851
1
Parent(s):
f3d7d6e
Upload folder using huggingface_hub
Browse files- README.md +36 -0
- added_tokens.json +3 -0
- bpe.codes +0 -0
- checkpoint-804/config.json +103 -0
- checkpoint-804/model.safetensors +3 -0
- checkpoint-804/optimizer.pt +3 -0
- checkpoint-804/rng_state.pth +3 -0
- checkpoint-804/scheduler.pt +3 -0
- checkpoint-804/trainer_state.json +2964 -0
- checkpoint-804/training_args.bin +3 -0
- config.json +103 -0
- model.safetensors +3 -0
- runs/Sep04_02-07-36_r-cuwfnguyen-autotrain-advanced-tqduow70-6d178-7tndo/events.out.tfevents.1725415658.r-cuwfnguyen-autotrain-advanced-tqduow70-6d178-7tndo.102.0 +2 -2
- runs/Sep04_02-07-36_r-cuwfnguyen-autotrain-advanced-tqduow70-6d178-7tndo/events.out.tfevents.1725431613.r-cuwfnguyen-autotrain-advanced-tqduow70-6d178-7tndo.102.1 +3 -0
- special_tokens_map.json +9 -0
- tokenizer_config.json +54 -0
- training_args.bin +3 -0
- training_params.json +30 -0
- vocab.txt +0 -0
README.md
ADDED
|
@@ -0,0 +1,36 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
|
| 2 |
+
---
|
| 3 |
+
tags:
|
| 4 |
+
- autotrain
|
| 5 |
+
- text-classification
|
| 6 |
+
base_model: vinai/phobert-base
|
| 7 |
+
widget:
|
| 8 |
+
- text: "I love AutoTrain"
|
| 9 |
+
---
|
| 10 |
+
|
| 11 |
+
# Model Trained Using AutoTrain
|
| 12 |
+
|
| 13 |
+
- Problem type: Text Classification
|
| 14 |
+
|
| 15 |
+
## Validation Metrics
|
| 16 |
+
loss: 0.890067458152771
|
| 17 |
+
|
| 18 |
+
f1_macro: 0.7178212262554823
|
| 19 |
+
|
| 20 |
+
f1_micro: 0.76
|
| 21 |
+
|
| 22 |
+
f1_weighted: 0.7341365415630122
|
| 23 |
+
|
| 24 |
+
precision_macro: 0.7468954248366012
|
| 25 |
+
|
| 26 |
+
precision_micro: 0.76
|
| 27 |
+
|
| 28 |
+
precision_weighted: 0.766861111111111
|
| 29 |
+
|
| 30 |
+
recall_macro: 0.7448179271708684
|
| 31 |
+
|
| 32 |
+
recall_micro: 0.76
|
| 33 |
+
|
| 34 |
+
recall_weighted: 0.76
|
| 35 |
+
|
| 36 |
+
accuracy: 0.76
|
added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
checkpoint-804/config.json
ADDED
|
@@ -0,0 +1,103 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "vinai/phobert-base",
|
| 3 |
+
"_num_labels": 34,
|
| 4 |
+
"architectures": [
|
| 5 |
+
"RobertaForSequenceClassification"
|
| 6 |
+
],
|
| 7 |
+
"attention_probs_dropout_prob": 0.1,
|
| 8 |
+
"bos_token_id": 0,
|
| 9 |
+
"classifier_dropout": null,
|
| 10 |
+
"eos_token_id": 2,
|
| 11 |
+
"gradient_checkpointing": false,
|
| 12 |
+
"hidden_act": "gelu",
|
| 13 |
+
"hidden_dropout_prob": 0.1,
|
| 14 |
+
"hidden_size": 768,
|
| 15 |
+
"id2label": {
|
| 16 |
+
"0": " H\u1ecfi v\u1ec1 c\u00e1ch x\u1eed l\u00fd khi g\u1eb7p v\u1ea5n \u0111\u1ec1 v\u1edbi s\u1ea3n ph\u1ea9m b\u1ecb l\u1ed7i",
|
| 17 |
+
"1": " H\u1ecfi v\u1ec1 t\u00ecnh tr\u1ea1ng \u0111\u01a1n h\u00e0ng v\u00e0 c\u00e1ch th\u1ee9c h\u1ee7y \u0111\u01a1n h\u00e0ng",
|
| 18 |
+
"2": " Y\u00eau c\u1ea7u h\u1ed7 tr\u1ee3 khi g\u1eb7p v\u1ea5n \u0111\u1ec1 v\u1edbi Shopee Mall ho\u1eb7c c\u00e1c shop kh\u00e1c",
|
| 19 |
+
"3": "H\u1ecfi v\u1ec1 ch\u00ednh s\u00e1ch b\u1ea3o h\u00e0nh v\u00e0 h\u1ed7 tr\u1ee3 k\u1ef9 thu\u1eadt",
|
| 20 |
+
"4": "H\u1ecfi v\u1ec1 ch\u00ednh s\u00e1ch b\u1ea3o m\u1eadt th\u00f4ng tin kh\u00e1ch h\u00e0ng",
|
| 21 |
+
"5": "H\u1ecfi v\u1ec1 ch\u00ednh s\u00e1ch \u0111i\u1ec3m th\u01b0\u1edfng v\u00e0 c\u00e1ch s\u1eed d\u1ee5ng",
|
| 22 |
+
"6": "H\u1ecfi v\u1ec1 ch\u00ednh s\u00e1ch \u0111\u1ed1i v\u1edbi kh\u00e1ch h\u00e0ng th\u00e2n thi\u1ebft v\u00e0 \u01b0u \u0111\u00e3i sinh nh\u1eadt",
|
| 23 |
+
"7": "H\u1ecfi v\u1ec1 ch\u00ednh s\u00e1ch \u0111\u1ed5i tr\u1ea3 v\u00e0 ho\u00e0n ti\u1ec1n",
|
| 24 |
+
"8": "H\u1ecfi v\u1ec1 ch\u01b0\u01a1ng tr\u00ecnh khuy\u1ebfn m\u00e3i, \u01b0u \u0111\u00e3i v\u00e0 flash sale",
|
| 25 |
+
"9": "H\u1ecfi v\u1ec1 ch\u01b0\u01a1ng tr\u00ecnh \u01b0u \u0111\u00e3i combo v\u00e0 ph\u00ed v\u1eadn chuy\u1ec3n khi mua combo",
|
| 26 |
+
"10": "H\u1ecfi v\u1ec1 c\u00e1ch th\u1ee9c g\u00f3i h\u00e0ng, \u0111\u1eb7c bi\u1ec7t \u0111\u1ed1i v\u1edbi h\u00e0ng d\u1ec5 v\u1ee1",
|
| 27 |
+
"11": "H\u1ecfi v\u1ec1 c\u00e1ch th\u1ee9c nh\u1eadn h\u00e0ng t\u1ea1i c\u1eeda h\u00e0ng ho\u1eb7c \u0111\u1ecba \u0111i\u1ec3m kh\u00e1c",
|
| 28 |
+
"12": "H\u1ecfi v\u1ec1 c\u00e1ch th\u1ee9c nh\u1eadn h\u00f3a \u0111\u01a1n \u0111\u1ecf",
|
| 29 |
+
"13": "H\u1ecfi v\u1ec1 c\u00e1ch th\u1ee9c \u0111\u1ed5i qu\u00e0 t\u1eb7ng k\u00e8m ho\u1eb7c s\u1ea3n ph\u1ea9m khuy\u1ebfn m\u00e3i",
|
| 30 |
+
"14": "H\u1ecfi v\u1ec1 c\u00e1ch th\u1ee9c \u0111\u1ed5i, tr\u1ea3, v\u00e0 s\u1eeda \u0111\u1ed5i th\u00f4ng tin \u0111\u01a1n h\u00e0ng",
|
| 31 |
+
"15": "H\u1ecfi v\u1ec1 gi\u00e1 s\u1ea3n ph\u1ea9m",
|
| 32 |
+
"16": "H\u1ecfi v\u1ec1 h\u01b0\u1edbng d\u1eabn s\u1eed d\u1ee5ng v\u00e0 b\u1ea3o qu\u1ea3n s\u1ea3n ph\u1ea9m",
|
| 33 |
+
"17": "H\u1ecfi v\u1ec1 k\u00edch th\u01b0\u1edbc, m\u00e0u s\u1eafc v\u00e0 ch\u1ea5t li\u1ec7u s\u1ea3n ph\u1ea9m",
|
| 34 |
+
"18": "H\u1ecfi v\u1ec1 l\u1ecbch s\u1eed mua h\u00e0ng v\u00e0 ch\u01b0\u01a1ng tr\u00ecnh kh\u00e1ch h\u00e0ng th\u00e2n thi\u1ebft",
|
| 35 |
+
"19": "H\u1ecfi v\u1ec1 ngu\u1ed3n g\u1ed1c, xu\u1ea5t x\u1ee9 s\u1ea3n ph\u1ea9m",
|
| 36 |
+
"20": "H\u1ecfi v\u1ec1 ph\u00ed v\u00e0 th\u1eddi gian v\u1eadn chuy\u1ec3n",
|
| 37 |
+
"21": "H\u1ecfi v\u1ec1 ph\u00ed v\u00e0 \u0111i\u1ec1u ki\u1ec7n \u0111\u1eb7t h\u00e0ng tr\u01b0\u1edbc s\u1ea3n ph\u1ea9m ch\u01b0a ra m\u1eaft",
|
| 38 |
+
"22": "H\u1ecfi v\u1ec1 ph\u01b0\u01a1ng th\u1ee9c thanh to\u00e1n v\u00e0 h\u1ed7 tr\u1ee3 khi g\u1eb7p s\u1ef1 c\u1ed1 thanh to\u00e1n",
|
| 39 |
+
"23": "H\u1ecfi v\u1ec1 s\u1ea3n ph\u1ea9m m\u1edbi v\u1ec1, s\u1ea3n ph\u1ea9m b\u00e1n ch\u1ea1y nh\u1ea5t v\u00e0 hot trend",
|
| 40 |
+
"24": "H\u1ecfi v\u1ec1 s\u1ea3n ph\u1ea9m \u0111\u00e3 qua s\u1eed d\u1ee5ng ho\u1eb7c h\u00e0ng \u0111\u1ed5i tr\u1ea3",
|
| 41 |
+
"25": "H\u1ecfi v\u1ec1 th\u1eddi gian giao h\u00e0ng d\u1ef1 ki\u1ebfn",
|
| 42 |
+
"26": "H\u1ecfi v\u1ec1 t\u00ecnh tr\u1ea1ng c\u00f2n h\u00e0ng v\u00e0 s\u1ed1 l\u01b0\u1ee3ng mua t\u1ed1i \u0111a",
|
| 43 |
+
"27": "Y\u00eau c\u1ea7u gi\u1ea3m gi\u00e1 khuy\u1ebfn m\u00e3i ho\u1eb7c m\u00e3 gi\u1ea3m gi\u00e1",
|
| 44 |
+
"28": "Y\u00eau c\u1ea7u g\u1eedi h\u00ecnh \u1ea3nh th\u1eadt c\u1ee7a s\u1ea3n ph\u1ea9m",
|
| 45 |
+
"29": "Y\u00eau c\u1ea7u h\u1ed7 tr\u1ee3 khi kh\u00f4ng nh\u1eadn \u0111\u01b0\u1ee3c h\u00e0ng ho\u1eb7c h\u00e0ng kh\u00f4ng \u0111\u00fang m\u00f4 t\u1ea3",
|
| 46 |
+
"30": "Y\u00eau c\u1ea7u th\u00f4ng tin chi ti\u1ebft s\u1ea3n ph\u1ea9m",
|
| 47 |
+
"31": "Y\u00eau c\u1ea7u t\u01b0 v\u1ea5n ph\u1ed1i \u0111\u1ed3 ho\u1eb7c c\u00e1ch s\u1eed d\u1ee5ng s\u1ea3n ph\u1ea9m",
|
| 48 |
+
"32": "Y\u00eau c\u1ea7u t\u01b0 v\u1ea5n s\u1ea3n ph\u1ea9m ph\u00f9 h\u1ee3p",
|
| 49 |
+
"33": "Y\u00eau c\u1ea7u x\u00e1c nh\u1eadn th\u00f4ng tin \u0111\u01a1n h\u00e0ng ho\u1eb7c \u0111\u1ecba ch\u1ec9 giao h\u00e0ng"
|
| 50 |
+
},
|
| 51 |
+
"initializer_range": 0.02,
|
| 52 |
+
"intermediate_size": 3072,
|
| 53 |
+
"label2id": {
|
| 54 |
+
" H\u1ecfi v\u1ec1 c\u00e1ch x\u1eed l\u00fd khi g\u1eb7p v\u1ea5n \u0111\u1ec1 v\u1edbi s\u1ea3n ph\u1ea9m b\u1ecb l\u1ed7i": 0,
|
| 55 |
+
" H\u1ecfi v\u1ec1 t\u00ecnh tr\u1ea1ng \u0111\u01a1n h\u00e0ng v\u00e0 c\u00e1ch th\u1ee9c h\u1ee7y \u0111\u01a1n h\u00e0ng": 1,
|
| 56 |
+
" Y\u00eau c\u1ea7u h\u1ed7 tr\u1ee3 khi g\u1eb7p v\u1ea5n \u0111\u1ec1 v\u1edbi Shopee Mall ho\u1eb7c c\u00e1c shop kh\u00e1c": 2,
|
| 57 |
+
"H\u1ecfi v\u1ec1 ch\u00ednh s\u00e1ch b\u1ea3o h\u00e0nh v\u00e0 h\u1ed7 tr\u1ee3 k\u1ef9 thu\u1eadt": 3,
|
| 58 |
+
"H\u1ecfi v\u1ec1 ch\u00ednh s\u00e1ch b\u1ea3o m\u1eadt th\u00f4ng tin kh\u00e1ch h\u00e0ng": 4,
|
| 59 |
+
"H\u1ecfi v\u1ec1 ch\u00ednh s\u00e1ch \u0111i\u1ec3m th\u01b0\u1edfng v\u00e0 c\u00e1ch s\u1eed d\u1ee5ng": 5,
|
| 60 |
+
"H\u1ecfi v\u1ec1 ch\u00ednh s\u00e1ch \u0111\u1ed1i v\u1edbi kh\u00e1ch h\u00e0ng th\u00e2n thi\u1ebft v\u00e0 \u01b0u \u0111\u00e3i sinh nh\u1eadt": 6,
|
| 61 |
+
"H\u1ecfi v\u1ec1 ch\u00ednh s\u00e1ch \u0111\u1ed5i tr\u1ea3 v\u00e0 ho\u00e0n ti\u1ec1n": 7,
|
| 62 |
+
"H\u1ecfi v\u1ec1 ch\u01b0\u01a1ng tr\u00ecnh khuy\u1ebfn m\u00e3i, \u01b0u \u0111\u00e3i v\u00e0 flash sale": 8,
|
| 63 |
+
"H\u1ecfi v\u1ec1 ch\u01b0\u01a1ng tr\u00ecnh \u01b0u \u0111\u00e3i combo v\u00e0 ph\u00ed v\u1eadn chuy\u1ec3n khi mua combo": 9,
|
| 64 |
+
"H\u1ecfi v\u1ec1 c\u00e1ch th\u1ee9c g\u00f3i h\u00e0ng, \u0111\u1eb7c bi\u1ec7t \u0111\u1ed1i v\u1edbi h\u00e0ng d\u1ec5 v\u1ee1": 10,
|
| 65 |
+
"H\u1ecfi v\u1ec1 c\u00e1ch th\u1ee9c nh\u1eadn h\u00e0ng t\u1ea1i c\u1eeda h\u00e0ng ho\u1eb7c \u0111\u1ecba \u0111i\u1ec3m kh\u00e1c": 11,
|
| 66 |
+
"H\u1ecfi v\u1ec1 c\u00e1ch th\u1ee9c nh\u1eadn h\u00f3a \u0111\u01a1n \u0111\u1ecf": 12,
|
| 67 |
+
"H\u1ecfi v\u1ec1 c\u00e1ch th\u1ee9c \u0111\u1ed5i qu\u00e0 t\u1eb7ng k\u00e8m ho\u1eb7c s\u1ea3n ph\u1ea9m khuy\u1ebfn m\u00e3i": 13,
|
| 68 |
+
"H\u1ecfi v\u1ec1 c\u00e1ch th\u1ee9c \u0111\u1ed5i, tr\u1ea3, v\u00e0 s\u1eeda \u0111\u1ed5i th\u00f4ng tin \u0111\u01a1n h\u00e0ng": 14,
|
| 69 |
+
"H\u1ecfi v\u1ec1 gi\u00e1 s\u1ea3n ph\u1ea9m": 15,
|
| 70 |
+
"H\u1ecfi v\u1ec1 h\u01b0\u1edbng d\u1eabn s\u1eed d\u1ee5ng v\u00e0 b\u1ea3o qu\u1ea3n s\u1ea3n ph\u1ea9m": 16,
|
| 71 |
+
"H\u1ecfi v\u1ec1 k\u00edch th\u01b0\u1edbc, m\u00e0u s\u1eafc v\u00e0 ch\u1ea5t li\u1ec7u s\u1ea3n ph\u1ea9m": 17,
|
| 72 |
+
"H\u1ecfi v\u1ec1 l\u1ecbch s\u1eed mua h\u00e0ng v\u00e0 ch\u01b0\u01a1ng tr\u00ecnh kh\u00e1ch h\u00e0ng th\u00e2n thi\u1ebft": 18,
|
| 73 |
+
"H\u1ecfi v\u1ec1 ngu\u1ed3n g\u1ed1c, xu\u1ea5t x\u1ee9 s\u1ea3n ph\u1ea9m": 19,
|
| 74 |
+
"H\u1ecfi v\u1ec1 ph\u00ed v\u00e0 th\u1eddi gian v\u1eadn chuy\u1ec3n": 20,
|
| 75 |
+
"H\u1ecfi v\u1ec1 ph\u00ed v\u00e0 \u0111i\u1ec1u ki\u1ec7n \u0111\u1eb7t h\u00e0ng tr\u01b0\u1edbc s\u1ea3n ph\u1ea9m ch\u01b0a ra m\u1eaft": 21,
|
| 76 |
+
"H\u1ecfi v\u1ec1 ph\u01b0\u01a1ng th\u1ee9c thanh to\u00e1n v\u00e0 h\u1ed7 tr\u1ee3 khi g\u1eb7p s\u1ef1 c\u1ed1 thanh to\u00e1n": 22,
|
| 77 |
+
"H\u1ecfi v\u1ec1 s\u1ea3n ph\u1ea9m m\u1edbi v\u1ec1, s\u1ea3n ph\u1ea9m b\u00e1n ch\u1ea1y nh\u1ea5t v\u00e0 hot trend": 23,
|
| 78 |
+
"H\u1ecfi v\u1ec1 s\u1ea3n ph\u1ea9m \u0111\u00e3 qua s\u1eed d\u1ee5ng ho\u1eb7c h\u00e0ng \u0111\u1ed5i tr\u1ea3": 24,
|
| 79 |
+
"H\u1ecfi v\u1ec1 th\u1eddi gian giao h\u00e0ng d\u1ef1 ki\u1ebfn": 25,
|
| 80 |
+
"H\u1ecfi v\u1ec1 t\u00ecnh tr\u1ea1ng c\u00f2n h\u00e0ng v\u00e0 s\u1ed1 l\u01b0\u1ee3ng mua t\u1ed1i \u0111a": 26,
|
| 81 |
+
"Y\u00eau c\u1ea7u gi\u1ea3m gi\u00e1 khuy\u1ebfn m\u00e3i ho\u1eb7c m\u00e3 gi\u1ea3m gi\u00e1": 27,
|
| 82 |
+
"Y\u00eau c\u1ea7u g\u1eedi h\u00ecnh \u1ea3nh th\u1eadt c\u1ee7a s\u1ea3n ph\u1ea9m": 28,
|
| 83 |
+
"Y\u00eau c\u1ea7u h\u1ed7 tr\u1ee3 khi kh\u00f4ng nh\u1eadn \u0111\u01b0\u1ee3c h\u00e0ng ho\u1eb7c h\u00e0ng kh\u00f4ng \u0111\u00fang m\u00f4 t\u1ea3": 29,
|
| 84 |
+
"Y\u00eau c\u1ea7u th\u00f4ng tin chi ti\u1ebft s\u1ea3n ph\u1ea9m": 30,
|
| 85 |
+
"Y\u00eau c\u1ea7u t\u01b0 v\u1ea5n ph\u1ed1i \u0111\u1ed3 ho\u1eb7c c\u00e1ch s\u1eed d\u1ee5ng s\u1ea3n ph\u1ea9m": 31,
|
| 86 |
+
"Y\u00eau c\u1ea7u t\u01b0 v\u1ea5n s\u1ea3n ph\u1ea9m ph\u00f9 h\u1ee3p": 32,
|
| 87 |
+
"Y\u00eau c\u1ea7u x\u00e1c nh\u1eadn th\u00f4ng tin \u0111\u01a1n h\u00e0ng ho\u1eb7c \u0111\u1ecba ch\u1ec9 giao h\u00e0ng": 33
|
| 88 |
+
},
|
| 89 |
+
"layer_norm_eps": 1e-05,
|
| 90 |
+
"max_position_embeddings": 258,
|
| 91 |
+
"model_type": "roberta",
|
| 92 |
+
"num_attention_heads": 12,
|
| 93 |
+
"num_hidden_layers": 12,
|
| 94 |
+
"pad_token_id": 1,
|
| 95 |
+
"position_embedding_type": "absolute",
|
| 96 |
+
"problem_type": "single_label_classification",
|
| 97 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 98 |
+
"torch_dtype": "float32",
|
| 99 |
+
"transformers_version": "4.44.1",
|
| 100 |
+
"type_vocab_size": 1,
|
| 101 |
+
"use_cache": true,
|
| 102 |
+
"vocab_size": 64001
|
| 103 |
+
}
|
checkpoint-804/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5826d0f1dcc22361c7b62ea13a25df96f2f5df8fd171223ca9ed8a55fe74b661
|
| 3 |
+
size 540121816
|
checkpoint-804/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ffc80ac3cce3e9bb2f5a6e84ba6bc447360a8ce21768fa40321116492cc95a81
|
| 3 |
+
size 1080357882
|
checkpoint-804/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9f7ab35b74288490c67ff83c3396c36dfac7dd8561e58018df35cee5f14c27d7
|
| 3 |
+
size 13990
|
checkpoint-804/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:dc284fae0f6688769e6e852885ff2fcfd7422e360f0717b6b301d16e492c2b54
|
| 3 |
+
size 1064
|
checkpoint-804/trainer_state.json
ADDED
|
@@ -0,0 +1,2964 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.890067458152771,
|
| 3 |
+
"best_model_checkpoint": "autotrain-okjuj-buuqs/checkpoint-804",
|
| 4 |
+
"epoch": 6.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 804,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.014925373134328358,
|
| 13 |
+
"grad_norm": 3.468337297439575,
|
| 14 |
+
"learning_rate": 4.975124378109453e-07,
|
| 15 |
+
"loss": 3.5512,
|
| 16 |
+
"step": 2
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.029850746268656716,
|
| 20 |
+
"grad_norm": 3.3904953002929688,
|
| 21 |
+
"learning_rate": 9.950248756218907e-07,
|
| 22 |
+
"loss": 3.5495,
|
| 23 |
+
"step": 4
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.04477611940298507,
|
| 27 |
+
"grad_norm": 8.210745811462402,
|
| 28 |
+
"learning_rate": 1.4925373134328358e-06,
|
| 29 |
+
"loss": 3.5323,
|
| 30 |
+
"step": 6
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.05970149253731343,
|
| 34 |
+
"grad_norm": 3.321380376815796,
|
| 35 |
+
"learning_rate": 1.9900497512437813e-06,
|
| 36 |
+
"loss": 3.5311,
|
| 37 |
+
"step": 8
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.07462686567164178,
|
| 41 |
+
"grad_norm": 3.103438377380371,
|
| 42 |
+
"learning_rate": 2.4875621890547264e-06,
|
| 43 |
+
"loss": 3.5312,
|
| 44 |
+
"step": 10
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.08955223880597014,
|
| 48 |
+
"grad_norm": 3.2276968955993652,
|
| 49 |
+
"learning_rate": 2.9850746268656716e-06,
|
| 50 |
+
"loss": 3.5472,
|
| 51 |
+
"step": 12
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.1044776119402985,
|
| 55 |
+
"grad_norm": 3.511948585510254,
|
| 56 |
+
"learning_rate": 3.482587064676617e-06,
|
| 57 |
+
"loss": 3.5238,
|
| 58 |
+
"step": 14
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.11940298507462686,
|
| 62 |
+
"grad_norm": 3.459820508956909,
|
| 63 |
+
"learning_rate": 3.980099502487563e-06,
|
| 64 |
+
"loss": 3.5457,
|
| 65 |
+
"step": 16
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.13432835820895522,
|
| 69 |
+
"grad_norm": 3.187225103378296,
|
| 70 |
+
"learning_rate": 4.477611940298508e-06,
|
| 71 |
+
"loss": 3.5064,
|
| 72 |
+
"step": 18
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.14925373134328357,
|
| 76 |
+
"grad_norm": 3.3685667514801025,
|
| 77 |
+
"learning_rate": 4.975124378109453e-06,
|
| 78 |
+
"loss": 3.5019,
|
| 79 |
+
"step": 20
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.16417910447761194,
|
| 83 |
+
"grad_norm": 3.525916576385498,
|
| 84 |
+
"learning_rate": 5.472636815920398e-06,
|
| 85 |
+
"loss": 3.5202,
|
| 86 |
+
"step": 22
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.1791044776119403,
|
| 90 |
+
"grad_norm": 3.199352979660034,
|
| 91 |
+
"learning_rate": 5.970149253731343e-06,
|
| 92 |
+
"loss": 3.4878,
|
| 93 |
+
"step": 24
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.19402985074626866,
|
| 97 |
+
"grad_norm": 3.151981830596924,
|
| 98 |
+
"learning_rate": 6.467661691542288e-06,
|
| 99 |
+
"loss": 3.5254,
|
| 100 |
+
"step": 26
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.208955223880597,
|
| 104 |
+
"grad_norm": 3.353118896484375,
|
| 105 |
+
"learning_rate": 6.965174129353234e-06,
|
| 106 |
+
"loss": 3.5179,
|
| 107 |
+
"step": 28
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.22388059701492538,
|
| 111 |
+
"grad_norm": 5.352255344390869,
|
| 112 |
+
"learning_rate": 7.4626865671641785e-06,
|
| 113 |
+
"loss": 3.4951,
|
| 114 |
+
"step": 30
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.23880597014925373,
|
| 118 |
+
"grad_norm": 4.2452712059021,
|
| 119 |
+
"learning_rate": 7.960199004975125e-06,
|
| 120 |
+
"loss": 3.5144,
|
| 121 |
+
"step": 32
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.2537313432835821,
|
| 125 |
+
"grad_norm": 3.4914658069610596,
|
| 126 |
+
"learning_rate": 8.45771144278607e-06,
|
| 127 |
+
"loss": 3.4755,
|
| 128 |
+
"step": 34
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.26865671641791045,
|
| 132 |
+
"grad_norm": 3.413722038269043,
|
| 133 |
+
"learning_rate": 8.955223880597016e-06,
|
| 134 |
+
"loss": 3.5059,
|
| 135 |
+
"step": 36
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.2835820895522388,
|
| 139 |
+
"grad_norm": 3.5784690380096436,
|
| 140 |
+
"learning_rate": 9.45273631840796e-06,
|
| 141 |
+
"loss": 3.503,
|
| 142 |
+
"step": 38
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.29850746268656714,
|
| 146 |
+
"grad_norm": 3.3770291805267334,
|
| 147 |
+
"learning_rate": 9.950248756218906e-06,
|
| 148 |
+
"loss": 3.4968,
|
| 149 |
+
"step": 40
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.31343283582089554,
|
| 153 |
+
"grad_norm": 3.709470748901367,
|
| 154 |
+
"learning_rate": 1.0447761194029851e-05,
|
| 155 |
+
"loss": 3.5453,
|
| 156 |
+
"step": 42
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.3283582089552239,
|
| 160 |
+
"grad_norm": 3.529064655303955,
|
| 161 |
+
"learning_rate": 1.0945273631840796e-05,
|
| 162 |
+
"loss": 3.4461,
|
| 163 |
+
"step": 44
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.34328358208955223,
|
| 167 |
+
"grad_norm": 3.4094252586364746,
|
| 168 |
+
"learning_rate": 1.1442786069651743e-05,
|
| 169 |
+
"loss": 3.483,
|
| 170 |
+
"step": 46
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.3582089552238806,
|
| 174 |
+
"grad_norm": 3.385557174682617,
|
| 175 |
+
"learning_rate": 1.1940298507462686e-05,
|
| 176 |
+
"loss": 3.4522,
|
| 177 |
+
"step": 48
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.373134328358209,
|
| 181 |
+
"grad_norm": 3.4028918743133545,
|
| 182 |
+
"learning_rate": 1.2437810945273633e-05,
|
| 183 |
+
"loss": 3.4814,
|
| 184 |
+
"step": 50
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.3880597014925373,
|
| 188 |
+
"grad_norm": 3.482884168624878,
|
| 189 |
+
"learning_rate": 1.2935323383084577e-05,
|
| 190 |
+
"loss": 3.4917,
|
| 191 |
+
"step": 52
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.40298507462686567,
|
| 195 |
+
"grad_norm": 3.4329397678375244,
|
| 196 |
+
"learning_rate": 1.3432835820895523e-05,
|
| 197 |
+
"loss": 3.4945,
|
| 198 |
+
"step": 54
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.417910447761194,
|
| 202 |
+
"grad_norm": 3.6114747524261475,
|
| 203 |
+
"learning_rate": 1.3930348258706468e-05,
|
| 204 |
+
"loss": 3.3991,
|
| 205 |
+
"step": 56
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.43283582089552236,
|
| 209 |
+
"grad_norm": 3.743497848510742,
|
| 210 |
+
"learning_rate": 1.4427860696517415e-05,
|
| 211 |
+
"loss": 3.4314,
|
| 212 |
+
"step": 58
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.44776119402985076,
|
| 216 |
+
"grad_norm": 13.538641929626465,
|
| 217 |
+
"learning_rate": 1.4925373134328357e-05,
|
| 218 |
+
"loss": 3.4495,
|
| 219 |
+
"step": 60
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.4626865671641791,
|
| 223 |
+
"grad_norm": 3.536511182785034,
|
| 224 |
+
"learning_rate": 1.5422885572139304e-05,
|
| 225 |
+
"loss": 3.4579,
|
| 226 |
+
"step": 62
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.47761194029850745,
|
| 230 |
+
"grad_norm": 4.389700889587402,
|
| 231 |
+
"learning_rate": 1.592039800995025e-05,
|
| 232 |
+
"loss": 3.3908,
|
| 233 |
+
"step": 64
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.4925373134328358,
|
| 237 |
+
"grad_norm": 4.821465969085693,
|
| 238 |
+
"learning_rate": 1.6417910447761194e-05,
|
| 239 |
+
"loss": 3.4305,
|
| 240 |
+
"step": 66
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.5074626865671642,
|
| 244 |
+
"grad_norm": 3.5865774154663086,
|
| 245 |
+
"learning_rate": 1.691542288557214e-05,
|
| 246 |
+
"loss": 3.4297,
|
| 247 |
+
"step": 68
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.5223880597014925,
|
| 251 |
+
"grad_norm": 3.3377368450164795,
|
| 252 |
+
"learning_rate": 1.7412935323383088e-05,
|
| 253 |
+
"loss": 3.5506,
|
| 254 |
+
"step": 70
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.5373134328358209,
|
| 258 |
+
"grad_norm": 3.887446880340576,
|
| 259 |
+
"learning_rate": 1.791044776119403e-05,
|
| 260 |
+
"loss": 3.4623,
|
| 261 |
+
"step": 72
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.5522388059701493,
|
| 265 |
+
"grad_norm": 3.546586275100708,
|
| 266 |
+
"learning_rate": 1.8407960199004975e-05,
|
| 267 |
+
"loss": 3.4059,
|
| 268 |
+
"step": 74
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.5671641791044776,
|
| 272 |
+
"grad_norm": 3.8084304332733154,
|
| 273 |
+
"learning_rate": 1.890547263681592e-05,
|
| 274 |
+
"loss": 3.3708,
|
| 275 |
+
"step": 76
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.582089552238806,
|
| 279 |
+
"grad_norm": 3.5616061687469482,
|
| 280 |
+
"learning_rate": 1.9402985074626868e-05,
|
| 281 |
+
"loss": 3.4297,
|
| 282 |
+
"step": 78
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.5970149253731343,
|
| 286 |
+
"grad_norm": 3.7048122882843018,
|
| 287 |
+
"learning_rate": 1.990049751243781e-05,
|
| 288 |
+
"loss": 3.3541,
|
| 289 |
+
"step": 80
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.6119402985074627,
|
| 293 |
+
"grad_norm": 3.7918601036071777,
|
| 294 |
+
"learning_rate": 2.0398009950248755e-05,
|
| 295 |
+
"loss": 3.3734,
|
| 296 |
+
"step": 82
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.6268656716417911,
|
| 300 |
+
"grad_norm": 4.018467903137207,
|
| 301 |
+
"learning_rate": 2.0895522388059702e-05,
|
| 302 |
+
"loss": 3.3347,
|
| 303 |
+
"step": 84
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.6417910447761194,
|
| 307 |
+
"grad_norm": 3.6919217109680176,
|
| 308 |
+
"learning_rate": 2.139303482587065e-05,
|
| 309 |
+
"loss": 3.4747,
|
| 310 |
+
"step": 86
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.6567164179104478,
|
| 314 |
+
"grad_norm": 3.553607940673828,
|
| 315 |
+
"learning_rate": 2.1890547263681592e-05,
|
| 316 |
+
"loss": 3.346,
|
| 317 |
+
"step": 88
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.6716417910447762,
|
| 321 |
+
"grad_norm": 3.717123508453369,
|
| 322 |
+
"learning_rate": 2.238805970149254e-05,
|
| 323 |
+
"loss": 3.3681,
|
| 324 |
+
"step": 90
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.6865671641791045,
|
| 328 |
+
"grad_norm": 4.329309463500977,
|
| 329 |
+
"learning_rate": 2.2885572139303486e-05,
|
| 330 |
+
"loss": 3.2841,
|
| 331 |
+
"step": 92
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.7014925373134329,
|
| 335 |
+
"grad_norm": 3.9699785709381104,
|
| 336 |
+
"learning_rate": 2.338308457711443e-05,
|
| 337 |
+
"loss": 3.3339,
|
| 338 |
+
"step": 94
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.7164179104477612,
|
| 342 |
+
"grad_norm": 3.9800264835357666,
|
| 343 |
+
"learning_rate": 2.3880597014925373e-05,
|
| 344 |
+
"loss": 3.272,
|
| 345 |
+
"step": 96
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.7313432835820896,
|
| 349 |
+
"grad_norm": 3.881544589996338,
|
| 350 |
+
"learning_rate": 2.437810945273632e-05,
|
| 351 |
+
"loss": 3.3739,
|
| 352 |
+
"step": 98
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.746268656716418,
|
| 356 |
+
"grad_norm": 3.7055130004882812,
|
| 357 |
+
"learning_rate": 2.4875621890547266e-05,
|
| 358 |
+
"loss": 3.2847,
|
| 359 |
+
"step": 100
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.7611940298507462,
|
| 363 |
+
"grad_norm": 4.061646461486816,
|
| 364 |
+
"learning_rate": 2.537313432835821e-05,
|
| 365 |
+
"loss": 3.2573,
|
| 366 |
+
"step": 102
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.7761194029850746,
|
| 370 |
+
"grad_norm": 4.569900035858154,
|
| 371 |
+
"learning_rate": 2.5870646766169153e-05,
|
| 372 |
+
"loss": 3.2594,
|
| 373 |
+
"step": 104
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.7910447761194029,
|
| 377 |
+
"grad_norm": 4.053166389465332,
|
| 378 |
+
"learning_rate": 2.6368159203980103e-05,
|
| 379 |
+
"loss": 3.2739,
|
| 380 |
+
"step": 106
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.8059701492537313,
|
| 384 |
+
"grad_norm": 5.338119029998779,
|
| 385 |
+
"learning_rate": 2.6865671641791047e-05,
|
| 386 |
+
"loss": 3.1965,
|
| 387 |
+
"step": 108
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.8208955223880597,
|
| 391 |
+
"grad_norm": 4.987647533416748,
|
| 392 |
+
"learning_rate": 2.736318407960199e-05,
|
| 393 |
+
"loss": 3.1951,
|
| 394 |
+
"step": 110
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.835820895522388,
|
| 398 |
+
"grad_norm": 4.582657814025879,
|
| 399 |
+
"learning_rate": 2.7860696517412937e-05,
|
| 400 |
+
"loss": 3.1758,
|
| 401 |
+
"step": 112
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.8507462686567164,
|
| 405 |
+
"grad_norm": 4.130568027496338,
|
| 406 |
+
"learning_rate": 2.835820895522388e-05,
|
| 407 |
+
"loss": 3.2563,
|
| 408 |
+
"step": 114
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.8656716417910447,
|
| 412 |
+
"grad_norm": 3.6177594661712646,
|
| 413 |
+
"learning_rate": 2.885572139303483e-05,
|
| 414 |
+
"loss": 3.1394,
|
| 415 |
+
"step": 116
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.8805970149253731,
|
| 419 |
+
"grad_norm": 4.4392242431640625,
|
| 420 |
+
"learning_rate": 2.935323383084577e-05,
|
| 421 |
+
"loss": 2.9928,
|
| 422 |
+
"step": 118
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.8955223880597015,
|
| 426 |
+
"grad_norm": 4.2055439949035645,
|
| 427 |
+
"learning_rate": 2.9850746268656714e-05,
|
| 428 |
+
"loss": 3.0482,
|
| 429 |
+
"step": 120
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.9104477611940298,
|
| 433 |
+
"grad_norm": 4.438642978668213,
|
| 434 |
+
"learning_rate": 3.0348258706467664e-05,
|
| 435 |
+
"loss": 3.024,
|
| 436 |
+
"step": 122
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.9253731343283582,
|
| 440 |
+
"grad_norm": 4.41525936126709,
|
| 441 |
+
"learning_rate": 3.084577114427861e-05,
|
| 442 |
+
"loss": 3.1619,
|
| 443 |
+
"step": 124
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.9402985074626866,
|
| 447 |
+
"grad_norm": 4.188882827758789,
|
| 448 |
+
"learning_rate": 3.1343283582089554e-05,
|
| 449 |
+
"loss": 3.2697,
|
| 450 |
+
"step": 126
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.9552238805970149,
|
| 454 |
+
"grad_norm": 3.903090000152588,
|
| 455 |
+
"learning_rate": 3.18407960199005e-05,
|
| 456 |
+
"loss": 3.0515,
|
| 457 |
+
"step": 128
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.9701492537313433,
|
| 461 |
+
"grad_norm": 4.03048849105835,
|
| 462 |
+
"learning_rate": 3.233830845771145e-05,
|
| 463 |
+
"loss": 2.9312,
|
| 464 |
+
"step": 130
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.9850746268656716,
|
| 468 |
+
"grad_norm": 3.842494249343872,
|
| 469 |
+
"learning_rate": 3.283582089552239e-05,
|
| 470 |
+
"loss": 3.0714,
|
| 471 |
+
"step": 132
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 1.0,
|
| 475 |
+
"grad_norm": 3.9151153564453125,
|
| 476 |
+
"learning_rate": 3.3333333333333335e-05,
|
| 477 |
+
"loss": 2.8809,
|
| 478 |
+
"step": 134
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 1.0,
|
| 482 |
+
"eval_accuracy": 0.375,
|
| 483 |
+
"eval_f1_macro": 0.2794185541385505,
|
| 484 |
+
"eval_f1_micro": 0.375,
|
| 485 |
+
"eval_f1_weighted": 0.29179203240000456,
|
| 486 |
+
"eval_loss": 2.9781627655029297,
|
| 487 |
+
"eval_precision_macro": 0.26769010075416033,
|
| 488 |
+
"eval_precision_micro": 0.375,
|
| 489 |
+
"eval_precision_weighted": 0.28345060265850164,
|
| 490 |
+
"eval_recall_macro": 0.3689075630252101,
|
| 491 |
+
"eval_recall_micro": 0.375,
|
| 492 |
+
"eval_recall_weighted": 0.375,
|
| 493 |
+
"eval_runtime": 38.3444,
|
| 494 |
+
"eval_samples_per_second": 5.216,
|
| 495 |
+
"eval_steps_per_second": 0.183,
|
| 496 |
+
"step": 134
|
| 497 |
+
},
|
| 498 |
+
{
|
| 499 |
+
"epoch": 1.0149253731343284,
|
| 500 |
+
"grad_norm": 3.894672155380249,
|
| 501 |
+
"learning_rate": 3.383084577114428e-05,
|
| 502 |
+
"loss": 2.8126,
|
| 503 |
+
"step": 136
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 1.0298507462686568,
|
| 507 |
+
"grad_norm": 4.255545139312744,
|
| 508 |
+
"learning_rate": 3.432835820895522e-05,
|
| 509 |
+
"loss": 2.9291,
|
| 510 |
+
"step": 138
|
| 511 |
+
},
|
| 512 |
+
{
|
| 513 |
+
"epoch": 1.044776119402985,
|
| 514 |
+
"grad_norm": 4.275138854980469,
|
| 515 |
+
"learning_rate": 3.4825870646766175e-05,
|
| 516 |
+
"loss": 2.8673,
|
| 517 |
+
"step": 140
|
| 518 |
+
},
|
| 519 |
+
{
|
| 520 |
+
"epoch": 1.0597014925373134,
|
| 521 |
+
"grad_norm": 4.796700477600098,
|
| 522 |
+
"learning_rate": 3.5323383084577115e-05,
|
| 523 |
+
"loss": 2.8431,
|
| 524 |
+
"step": 142
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 1.0746268656716418,
|
| 528 |
+
"grad_norm": 4.217311859130859,
|
| 529 |
+
"learning_rate": 3.582089552238806e-05,
|
| 530 |
+
"loss": 2.8104,
|
| 531 |
+
"step": 144
|
| 532 |
+
},
|
| 533 |
+
{
|
| 534 |
+
"epoch": 1.0895522388059702,
|
| 535 |
+
"grad_norm": 4.417973041534424,
|
| 536 |
+
"learning_rate": 3.631840796019901e-05,
|
| 537 |
+
"loss": 2.7942,
|
| 538 |
+
"step": 146
|
| 539 |
+
},
|
| 540 |
+
{
|
| 541 |
+
"epoch": 1.1044776119402986,
|
| 542 |
+
"grad_norm": 3.8580806255340576,
|
| 543 |
+
"learning_rate": 3.681592039800995e-05,
|
| 544 |
+
"loss": 2.6514,
|
| 545 |
+
"step": 148
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 1.1194029850746268,
|
| 549 |
+
"grad_norm": 4.099730014801025,
|
| 550 |
+
"learning_rate": 3.73134328358209e-05,
|
| 551 |
+
"loss": 2.693,
|
| 552 |
+
"step": 150
|
| 553 |
+
},
|
| 554 |
+
{
|
| 555 |
+
"epoch": 1.1343283582089552,
|
| 556 |
+
"grad_norm": 4.598573684692383,
|
| 557 |
+
"learning_rate": 3.781094527363184e-05,
|
| 558 |
+
"loss": 2.7205,
|
| 559 |
+
"step": 152
|
| 560 |
+
},
|
| 561 |
+
{
|
| 562 |
+
"epoch": 1.1492537313432836,
|
| 563 |
+
"grad_norm": 4.880782604217529,
|
| 564 |
+
"learning_rate": 3.830845771144278e-05,
|
| 565 |
+
"loss": 2.8268,
|
| 566 |
+
"step": 154
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 1.164179104477612,
|
| 570 |
+
"grad_norm": 3.9455466270446777,
|
| 571 |
+
"learning_rate": 3.8805970149253736e-05,
|
| 572 |
+
"loss": 2.5482,
|
| 573 |
+
"step": 156
|
| 574 |
+
},
|
| 575 |
+
{
|
| 576 |
+
"epoch": 1.1791044776119404,
|
| 577 |
+
"grad_norm": 3.5259761810302734,
|
| 578 |
+
"learning_rate": 3.9303482587064676e-05,
|
| 579 |
+
"loss": 2.4929,
|
| 580 |
+
"step": 158
|
| 581 |
+
},
|
| 582 |
+
{
|
| 583 |
+
"epoch": 1.1940298507462686,
|
| 584 |
+
"grad_norm": 4.351325035095215,
|
| 585 |
+
"learning_rate": 3.980099502487562e-05,
|
| 586 |
+
"loss": 2.6261,
|
| 587 |
+
"step": 160
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 1.208955223880597,
|
| 591 |
+
"grad_norm": 3.8126604557037354,
|
| 592 |
+
"learning_rate": 4.029850746268657e-05,
|
| 593 |
+
"loss": 2.3924,
|
| 594 |
+
"step": 162
|
| 595 |
+
},
|
| 596 |
+
{
|
| 597 |
+
"epoch": 1.2238805970149254,
|
| 598 |
+
"grad_norm": 5.261939525604248,
|
| 599 |
+
"learning_rate": 4.079601990049751e-05,
|
| 600 |
+
"loss": 2.3364,
|
| 601 |
+
"step": 164
|
| 602 |
+
},
|
| 603 |
+
{
|
| 604 |
+
"epoch": 1.2388059701492538,
|
| 605 |
+
"grad_norm": 5.208514213562012,
|
| 606 |
+
"learning_rate": 4.1293532338308464e-05,
|
| 607 |
+
"loss": 2.3356,
|
| 608 |
+
"step": 166
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 1.2537313432835822,
|
| 612 |
+
"grad_norm": 4.369601249694824,
|
| 613 |
+
"learning_rate": 4.1791044776119404e-05,
|
| 614 |
+
"loss": 2.4191,
|
| 615 |
+
"step": 168
|
| 616 |
+
},
|
| 617 |
+
{
|
| 618 |
+
"epoch": 1.2686567164179103,
|
| 619 |
+
"grad_norm": 5.281370639801025,
|
| 620 |
+
"learning_rate": 4.228855721393035e-05,
|
| 621 |
+
"loss": 2.4544,
|
| 622 |
+
"step": 170
|
| 623 |
+
},
|
| 624 |
+
{
|
| 625 |
+
"epoch": 1.2835820895522387,
|
| 626 |
+
"grad_norm": 4.296084880828857,
|
| 627 |
+
"learning_rate": 4.27860696517413e-05,
|
| 628 |
+
"loss": 2.4446,
|
| 629 |
+
"step": 172
|
| 630 |
+
},
|
| 631 |
+
{
|
| 632 |
+
"epoch": 1.2985074626865671,
|
| 633 |
+
"grad_norm": 3.7815287113189697,
|
| 634 |
+
"learning_rate": 4.328358208955224e-05,
|
| 635 |
+
"loss": 2.3687,
|
| 636 |
+
"step": 174
|
| 637 |
+
},
|
| 638 |
+
{
|
| 639 |
+
"epoch": 1.3134328358208955,
|
| 640 |
+
"grad_norm": 7.161779403686523,
|
| 641 |
+
"learning_rate": 4.3781094527363184e-05,
|
| 642 |
+
"loss": 2.67,
|
| 643 |
+
"step": 176
|
| 644 |
+
},
|
| 645 |
+
{
|
| 646 |
+
"epoch": 1.328358208955224,
|
| 647 |
+
"grad_norm": 4.957999229431152,
|
| 648 |
+
"learning_rate": 4.427860696517413e-05,
|
| 649 |
+
"loss": 2.4644,
|
| 650 |
+
"step": 178
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 1.3432835820895521,
|
| 654 |
+
"grad_norm": 4.455275058746338,
|
| 655 |
+
"learning_rate": 4.477611940298508e-05,
|
| 656 |
+
"loss": 2.4389,
|
| 657 |
+
"step": 180
|
| 658 |
+
},
|
| 659 |
+
{
|
| 660 |
+
"epoch": 1.3582089552238805,
|
| 661 |
+
"grad_norm": 5.357694625854492,
|
| 662 |
+
"learning_rate": 4.5273631840796025e-05,
|
| 663 |
+
"loss": 2.1849,
|
| 664 |
+
"step": 182
|
| 665 |
+
},
|
| 666 |
+
{
|
| 667 |
+
"epoch": 1.373134328358209,
|
| 668 |
+
"grad_norm": 6.346301078796387,
|
| 669 |
+
"learning_rate": 4.577114427860697e-05,
|
| 670 |
+
"loss": 2.3134,
|
| 671 |
+
"step": 184
|
| 672 |
+
},
|
| 673 |
+
{
|
| 674 |
+
"epoch": 1.3880597014925373,
|
| 675 |
+
"grad_norm": 6.189851760864258,
|
| 676 |
+
"learning_rate": 4.626865671641791e-05,
|
| 677 |
+
"loss": 2.322,
|
| 678 |
+
"step": 186
|
| 679 |
+
},
|
| 680 |
+
{
|
| 681 |
+
"epoch": 1.4029850746268657,
|
| 682 |
+
"grad_norm": 3.729904890060425,
|
| 683 |
+
"learning_rate": 4.676616915422886e-05,
|
| 684 |
+
"loss": 2.2053,
|
| 685 |
+
"step": 188
|
| 686 |
+
},
|
| 687 |
+
{
|
| 688 |
+
"epoch": 1.417910447761194,
|
| 689 |
+
"grad_norm": 4.210063457489014,
|
| 690 |
+
"learning_rate": 4.7263681592039805e-05,
|
| 691 |
+
"loss": 1.9863,
|
| 692 |
+
"step": 190
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 1.4328358208955223,
|
| 696 |
+
"grad_norm": 4.447649002075195,
|
| 697 |
+
"learning_rate": 4.7761194029850745e-05,
|
| 698 |
+
"loss": 2.0302,
|
| 699 |
+
"step": 192
|
| 700 |
+
},
|
| 701 |
+
{
|
| 702 |
+
"epoch": 1.4477611940298507,
|
| 703 |
+
"grad_norm": 5.010406970977783,
|
| 704 |
+
"learning_rate": 4.82587064676617e-05,
|
| 705 |
+
"loss": 2.379,
|
| 706 |
+
"step": 194
|
| 707 |
+
},
|
| 708 |
+
{
|
| 709 |
+
"epoch": 1.462686567164179,
|
| 710 |
+
"grad_norm": 4.134596824645996,
|
| 711 |
+
"learning_rate": 4.875621890547264e-05,
|
| 712 |
+
"loss": 2.1104,
|
| 713 |
+
"step": 196
|
| 714 |
+
},
|
| 715 |
+
{
|
| 716 |
+
"epoch": 1.4776119402985075,
|
| 717 |
+
"grad_norm": 3.678053140640259,
|
| 718 |
+
"learning_rate": 4.9253731343283586e-05,
|
| 719 |
+
"loss": 2.0437,
|
| 720 |
+
"step": 198
|
| 721 |
+
},
|
| 722 |
+
{
|
| 723 |
+
"epoch": 1.4925373134328357,
|
| 724 |
+
"grad_norm": 6.969429016113281,
|
| 725 |
+
"learning_rate": 4.975124378109453e-05,
|
| 726 |
+
"loss": 2.1688,
|
| 727 |
+
"step": 200
|
| 728 |
+
},
|
| 729 |
+
{
|
| 730 |
+
"epoch": 1.5074626865671643,
|
| 731 |
+
"grad_norm": 5.489563941955566,
|
| 732 |
+
"learning_rate": 4.997236042012162e-05,
|
| 733 |
+
"loss": 1.9494,
|
| 734 |
+
"step": 202
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 1.5223880597014925,
|
| 738 |
+
"grad_norm": 3.8776581287384033,
|
| 739 |
+
"learning_rate": 4.991708126036485e-05,
|
| 740 |
+
"loss": 1.7246,
|
| 741 |
+
"step": 204
|
| 742 |
+
},
|
| 743 |
+
{
|
| 744 |
+
"epoch": 1.537313432835821,
|
| 745 |
+
"grad_norm": 3.915743350982666,
|
| 746 |
+
"learning_rate": 4.9861802100608076e-05,
|
| 747 |
+
"loss": 2.1206,
|
| 748 |
+
"step": 206
|
| 749 |
+
},
|
| 750 |
+
{
|
| 751 |
+
"epoch": 1.5522388059701493,
|
| 752 |
+
"grad_norm": 10.498429298400879,
|
| 753 |
+
"learning_rate": 4.9806522940851304e-05,
|
| 754 |
+
"loss": 1.8089,
|
| 755 |
+
"step": 208
|
| 756 |
+
},
|
| 757 |
+
{
|
| 758 |
+
"epoch": 1.5671641791044775,
|
| 759 |
+
"grad_norm": 5.074433326721191,
|
| 760 |
+
"learning_rate": 4.975124378109453e-05,
|
| 761 |
+
"loss": 2.2841,
|
| 762 |
+
"step": 210
|
| 763 |
+
},
|
| 764 |
+
{
|
| 765 |
+
"epoch": 1.582089552238806,
|
| 766 |
+
"grad_norm": 16.675743103027344,
|
| 767 |
+
"learning_rate": 4.969596462133776e-05,
|
| 768 |
+
"loss": 1.9404,
|
| 769 |
+
"step": 212
|
| 770 |
+
},
|
| 771 |
+
{
|
| 772 |
+
"epoch": 1.5970149253731343,
|
| 773 |
+
"grad_norm": 4.692626953125,
|
| 774 |
+
"learning_rate": 4.964068546158099e-05,
|
| 775 |
+
"loss": 1.9563,
|
| 776 |
+
"step": 214
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 1.6119402985074627,
|
| 780 |
+
"grad_norm": 4.603738307952881,
|
| 781 |
+
"learning_rate": 4.958540630182422e-05,
|
| 782 |
+
"loss": 2.4004,
|
| 783 |
+
"step": 216
|
| 784 |
+
},
|
| 785 |
+
{
|
| 786 |
+
"epoch": 1.626865671641791,
|
| 787 |
+
"grad_norm": 29.3167781829834,
|
| 788 |
+
"learning_rate": 4.9530127142067445e-05,
|
| 789 |
+
"loss": 2.1461,
|
| 790 |
+
"step": 218
|
| 791 |
+
},
|
| 792 |
+
{
|
| 793 |
+
"epoch": 1.6417910447761193,
|
| 794 |
+
"grad_norm": 4.929641246795654,
|
| 795 |
+
"learning_rate": 4.947484798231067e-05,
|
| 796 |
+
"loss": 1.6009,
|
| 797 |
+
"step": 220
|
| 798 |
+
},
|
| 799 |
+
{
|
| 800 |
+
"epoch": 1.6567164179104479,
|
| 801 |
+
"grad_norm": 7.109385967254639,
|
| 802 |
+
"learning_rate": 4.94195688225539e-05,
|
| 803 |
+
"loss": 1.884,
|
| 804 |
+
"step": 222
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"epoch": 1.671641791044776,
|
| 808 |
+
"grad_norm": 14.075770378112793,
|
| 809 |
+
"learning_rate": 4.936428966279712e-05,
|
| 810 |
+
"loss": 1.6543,
|
| 811 |
+
"step": 224
|
| 812 |
+
},
|
| 813 |
+
{
|
| 814 |
+
"epoch": 1.6865671641791045,
|
| 815 |
+
"grad_norm": 4.957328796386719,
|
| 816 |
+
"learning_rate": 4.930901050304035e-05,
|
| 817 |
+
"loss": 1.6223,
|
| 818 |
+
"step": 226
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 1.7014925373134329,
|
| 822 |
+
"grad_norm": 6.562300205230713,
|
| 823 |
+
"learning_rate": 4.9253731343283586e-05,
|
| 824 |
+
"loss": 1.4456,
|
| 825 |
+
"step": 228
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 1.716417910447761,
|
| 829 |
+
"grad_norm": 5.4999237060546875,
|
| 830 |
+
"learning_rate": 4.9198452183526814e-05,
|
| 831 |
+
"loss": 1.7814,
|
| 832 |
+
"step": 230
|
| 833 |
+
},
|
| 834 |
+
{
|
| 835 |
+
"epoch": 1.7313432835820897,
|
| 836 |
+
"grad_norm": 5.5423736572265625,
|
| 837 |
+
"learning_rate": 4.914317302377004e-05,
|
| 838 |
+
"loss": 1.884,
|
| 839 |
+
"step": 232
|
| 840 |
+
},
|
| 841 |
+
{
|
| 842 |
+
"epoch": 1.7462686567164178,
|
| 843 |
+
"grad_norm": 5.507659435272217,
|
| 844 |
+
"learning_rate": 4.908789386401327e-05,
|
| 845 |
+
"loss": 1.8302,
|
| 846 |
+
"step": 234
|
| 847 |
+
},
|
| 848 |
+
{
|
| 849 |
+
"epoch": 1.7611940298507462,
|
| 850 |
+
"grad_norm": 7.983649730682373,
|
| 851 |
+
"learning_rate": 4.90326147042565e-05,
|
| 852 |
+
"loss": 1.9323,
|
| 853 |
+
"step": 236
|
| 854 |
+
},
|
| 855 |
+
{
|
| 856 |
+
"epoch": 1.7761194029850746,
|
| 857 |
+
"grad_norm": 4.843520641326904,
|
| 858 |
+
"learning_rate": 4.8977335544499726e-05,
|
| 859 |
+
"loss": 1.6411,
|
| 860 |
+
"step": 238
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 1.7910447761194028,
|
| 864 |
+
"grad_norm": 4.717188835144043,
|
| 865 |
+
"learning_rate": 4.8922056384742954e-05,
|
| 866 |
+
"loss": 1.6714,
|
| 867 |
+
"step": 240
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 1.8059701492537314,
|
| 871 |
+
"grad_norm": 4.654526233673096,
|
| 872 |
+
"learning_rate": 4.886677722498618e-05,
|
| 873 |
+
"loss": 1.6478,
|
| 874 |
+
"step": 242
|
| 875 |
+
},
|
| 876 |
+
{
|
| 877 |
+
"epoch": 1.8208955223880596,
|
| 878 |
+
"grad_norm": 5.104541301727295,
|
| 879 |
+
"learning_rate": 4.881149806522941e-05,
|
| 880 |
+
"loss": 1.6125,
|
| 881 |
+
"step": 244
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"epoch": 1.835820895522388,
|
| 885 |
+
"grad_norm": 4.356812000274658,
|
| 886 |
+
"learning_rate": 4.875621890547264e-05,
|
| 887 |
+
"loss": 1.9471,
|
| 888 |
+
"step": 246
|
| 889 |
+
},
|
| 890 |
+
{
|
| 891 |
+
"epoch": 1.8507462686567164,
|
| 892 |
+
"grad_norm": 4.166391849517822,
|
| 893 |
+
"learning_rate": 4.870093974571587e-05,
|
| 894 |
+
"loss": 1.4783,
|
| 895 |
+
"step": 248
|
| 896 |
+
},
|
| 897 |
+
{
|
| 898 |
+
"epoch": 1.8656716417910446,
|
| 899 |
+
"grad_norm": 6.2490057945251465,
|
| 900 |
+
"learning_rate": 4.8645660585959095e-05,
|
| 901 |
+
"loss": 1.8157,
|
| 902 |
+
"step": 250
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 1.8805970149253732,
|
| 906 |
+
"grad_norm": 5.395322322845459,
|
| 907 |
+
"learning_rate": 4.859038142620232e-05,
|
| 908 |
+
"loss": 1.8702,
|
| 909 |
+
"step": 252
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 1.8955223880597014,
|
| 913 |
+
"grad_norm": 4.278140544891357,
|
| 914 |
+
"learning_rate": 4.853510226644555e-05,
|
| 915 |
+
"loss": 1.5588,
|
| 916 |
+
"step": 254
|
| 917 |
+
},
|
| 918 |
+
{
|
| 919 |
+
"epoch": 1.9104477611940298,
|
| 920 |
+
"grad_norm": 5.143709182739258,
|
| 921 |
+
"learning_rate": 4.847982310668878e-05,
|
| 922 |
+
"loss": 1.5783,
|
| 923 |
+
"step": 256
|
| 924 |
+
},
|
| 925 |
+
{
|
| 926 |
+
"epoch": 1.9253731343283582,
|
| 927 |
+
"grad_norm": 8.840173721313477,
|
| 928 |
+
"learning_rate": 4.842454394693201e-05,
|
| 929 |
+
"loss": 1.5149,
|
| 930 |
+
"step": 258
|
| 931 |
+
},
|
| 932 |
+
{
|
| 933 |
+
"epoch": 1.9402985074626866,
|
| 934 |
+
"grad_norm": 5.001150608062744,
|
| 935 |
+
"learning_rate": 4.836926478717524e-05,
|
| 936 |
+
"loss": 1.5529,
|
| 937 |
+
"step": 260
|
| 938 |
+
},
|
| 939 |
+
{
|
| 940 |
+
"epoch": 1.955223880597015,
|
| 941 |
+
"grad_norm": 4.259793281555176,
|
| 942 |
+
"learning_rate": 4.831398562741847e-05,
|
| 943 |
+
"loss": 1.7178,
|
| 944 |
+
"step": 262
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 1.9701492537313432,
|
| 948 |
+
"grad_norm": 3.2455430030822754,
|
| 949 |
+
"learning_rate": 4.82587064676617e-05,
|
| 950 |
+
"loss": 1.26,
|
| 951 |
+
"step": 264
|
| 952 |
+
},
|
| 953 |
+
{
|
| 954 |
+
"epoch": 1.9850746268656716,
|
| 955 |
+
"grad_norm": 3.899599313735962,
|
| 956 |
+
"learning_rate": 4.820342730790493e-05,
|
| 957 |
+
"loss": 1.5229,
|
| 958 |
+
"step": 266
|
| 959 |
+
},
|
| 960 |
+
{
|
| 961 |
+
"epoch": 2.0,
|
| 962 |
+
"grad_norm": 6.097535610198975,
|
| 963 |
+
"learning_rate": 4.814814814814815e-05,
|
| 964 |
+
"loss": 1.6973,
|
| 965 |
+
"step": 268
|
| 966 |
+
},
|
| 967 |
+
{
|
| 968 |
+
"epoch": 2.0,
|
| 969 |
+
"eval_accuracy": 0.63,
|
| 970 |
+
"eval_f1_macro": 0.569843507632617,
|
| 971 |
+
"eval_f1_micro": 0.63,
|
| 972 |
+
"eval_f1_weighted": 0.5873017265672064,
|
| 973 |
+
"eval_loss": 1.6746174097061157,
|
| 974 |
+
"eval_precision_macro": 0.5998444366091426,
|
| 975 |
+
"eval_precision_micro": 0.63,
|
| 976 |
+
"eval_precision_weighted": 0.623490675990676,
|
| 977 |
+
"eval_recall_macro": 0.6212885154061624,
|
| 978 |
+
"eval_recall_micro": 0.63,
|
| 979 |
+
"eval_recall_weighted": 0.63,
|
| 980 |
+
"eval_runtime": 35.3294,
|
| 981 |
+
"eval_samples_per_second": 5.661,
|
| 982 |
+
"eval_steps_per_second": 0.198,
|
| 983 |
+
"step": 268
|
| 984 |
+
},
|
| 985 |
+
{
|
| 986 |
+
"epoch": 2.014925373134328,
|
| 987 |
+
"grad_norm": 2.5601863861083984,
|
| 988 |
+
"learning_rate": 4.8092868988391376e-05,
|
| 989 |
+
"loss": 1.0241,
|
| 990 |
+
"step": 270
|
| 991 |
+
},
|
| 992 |
+
{
|
| 993 |
+
"epoch": 2.029850746268657,
|
| 994 |
+
"grad_norm": 4.754575252532959,
|
| 995 |
+
"learning_rate": 4.8037589828634604e-05,
|
| 996 |
+
"loss": 1.5737,
|
| 997 |
+
"step": 272
|
| 998 |
+
},
|
| 999 |
+
{
|
| 1000 |
+
"epoch": 2.044776119402985,
|
| 1001 |
+
"grad_norm": 4.370357990264893,
|
| 1002 |
+
"learning_rate": 4.798231066887783e-05,
|
| 1003 |
+
"loss": 1.4961,
|
| 1004 |
+
"step": 274
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 2.0597014925373136,
|
| 1008 |
+
"grad_norm": 8.790416717529297,
|
| 1009 |
+
"learning_rate": 4.792703150912106e-05,
|
| 1010 |
+
"loss": 1.6075,
|
| 1011 |
+
"step": 276
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 2.074626865671642,
|
| 1015 |
+
"grad_norm": 4.466557502746582,
|
| 1016 |
+
"learning_rate": 4.787175234936429e-05,
|
| 1017 |
+
"loss": 1.4832,
|
| 1018 |
+
"step": 278
|
| 1019 |
+
},
|
| 1020 |
+
{
|
| 1021 |
+
"epoch": 2.08955223880597,
|
| 1022 |
+
"grad_norm": 3.5914487838745117,
|
| 1023 |
+
"learning_rate": 4.781647318960752e-05,
|
| 1024 |
+
"loss": 1.1338,
|
| 1025 |
+
"step": 280
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 2.1044776119402986,
|
| 1029 |
+
"grad_norm": 3.8444533348083496,
|
| 1030 |
+
"learning_rate": 4.7761194029850745e-05,
|
| 1031 |
+
"loss": 1.2921,
|
| 1032 |
+
"step": 282
|
| 1033 |
+
},
|
| 1034 |
+
{
|
| 1035 |
+
"epoch": 2.1194029850746268,
|
| 1036 |
+
"grad_norm": 7.032942295074463,
|
| 1037 |
+
"learning_rate": 4.770591487009397e-05,
|
| 1038 |
+
"loss": 1.5322,
|
| 1039 |
+
"step": 284
|
| 1040 |
+
},
|
| 1041 |
+
{
|
| 1042 |
+
"epoch": 2.1343283582089554,
|
| 1043 |
+
"grad_norm": 6.563926696777344,
|
| 1044 |
+
"learning_rate": 4.76506357103372e-05,
|
| 1045 |
+
"loss": 1.2677,
|
| 1046 |
+
"step": 286
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 2.1492537313432836,
|
| 1050 |
+
"grad_norm": 3.413158655166626,
|
| 1051 |
+
"learning_rate": 4.7595356550580436e-05,
|
| 1052 |
+
"loss": 1.2462,
|
| 1053 |
+
"step": 288
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 2.1641791044776117,
|
| 1057 |
+
"grad_norm": 4.282972812652588,
|
| 1058 |
+
"learning_rate": 4.7540077390823664e-05,
|
| 1059 |
+
"loss": 1.1632,
|
| 1060 |
+
"step": 290
|
| 1061 |
+
},
|
| 1062 |
+
{
|
| 1063 |
+
"epoch": 2.1791044776119404,
|
| 1064 |
+
"grad_norm": 11.003588676452637,
|
| 1065 |
+
"learning_rate": 4.748479823106689e-05,
|
| 1066 |
+
"loss": 1.2687,
|
| 1067 |
+
"step": 292
|
| 1068 |
+
},
|
| 1069 |
+
{
|
| 1070 |
+
"epoch": 2.1940298507462686,
|
| 1071 |
+
"grad_norm": 4.725019931793213,
|
| 1072 |
+
"learning_rate": 4.742951907131012e-05,
|
| 1073 |
+
"loss": 1.2352,
|
| 1074 |
+
"step": 294
|
| 1075 |
+
},
|
| 1076 |
+
{
|
| 1077 |
+
"epoch": 2.208955223880597,
|
| 1078 |
+
"grad_norm": 3.6079459190368652,
|
| 1079 |
+
"learning_rate": 4.737423991155335e-05,
|
| 1080 |
+
"loss": 1.4994,
|
| 1081 |
+
"step": 296
|
| 1082 |
+
},
|
| 1083 |
+
{
|
| 1084 |
+
"epoch": 2.2238805970149254,
|
| 1085 |
+
"grad_norm": 3.576037645339966,
|
| 1086 |
+
"learning_rate": 4.731896075179658e-05,
|
| 1087 |
+
"loss": 1.1007,
|
| 1088 |
+
"step": 298
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 2.2388059701492535,
|
| 1092 |
+
"grad_norm": 3.0327272415161133,
|
| 1093 |
+
"learning_rate": 4.7263681592039805e-05,
|
| 1094 |
+
"loss": 1.0624,
|
| 1095 |
+
"step": 300
|
| 1096 |
+
},
|
| 1097 |
+
{
|
| 1098 |
+
"epoch": 2.253731343283582,
|
| 1099 |
+
"grad_norm": 5.710484504699707,
|
| 1100 |
+
"learning_rate": 4.720840243228303e-05,
|
| 1101 |
+
"loss": 1.2262,
|
| 1102 |
+
"step": 302
|
| 1103 |
+
},
|
| 1104 |
+
{
|
| 1105 |
+
"epoch": 2.2686567164179103,
|
| 1106 |
+
"grad_norm": 3.059814453125,
|
| 1107 |
+
"learning_rate": 4.715312327252626e-05,
|
| 1108 |
+
"loss": 1.0292,
|
| 1109 |
+
"step": 304
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 2.283582089552239,
|
| 1113 |
+
"grad_norm": 4.122474193572998,
|
| 1114 |
+
"learning_rate": 4.709784411276949e-05,
|
| 1115 |
+
"loss": 1.1731,
|
| 1116 |
+
"step": 306
|
| 1117 |
+
},
|
| 1118 |
+
{
|
| 1119 |
+
"epoch": 2.298507462686567,
|
| 1120 |
+
"grad_norm": 8.998103141784668,
|
| 1121 |
+
"learning_rate": 4.704256495301272e-05,
|
| 1122 |
+
"loss": 1.5075,
|
| 1123 |
+
"step": 308
|
| 1124 |
+
},
|
| 1125 |
+
{
|
| 1126 |
+
"epoch": 2.3134328358208958,
|
| 1127 |
+
"grad_norm": 3.947554349899292,
|
| 1128 |
+
"learning_rate": 4.6987285793255946e-05,
|
| 1129 |
+
"loss": 1.1681,
|
| 1130 |
+
"step": 310
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 2.328358208955224,
|
| 1134 |
+
"grad_norm": 6.467041492462158,
|
| 1135 |
+
"learning_rate": 4.693200663349917e-05,
|
| 1136 |
+
"loss": 1.2688,
|
| 1137 |
+
"step": 312
|
| 1138 |
+
},
|
| 1139 |
+
{
|
| 1140 |
+
"epoch": 2.343283582089552,
|
| 1141 |
+
"grad_norm": 6.714489936828613,
|
| 1142 |
+
"learning_rate": 4.68767274737424e-05,
|
| 1143 |
+
"loss": 1.5061,
|
| 1144 |
+
"step": 314
|
| 1145 |
+
},
|
| 1146 |
+
{
|
| 1147 |
+
"epoch": 2.3582089552238807,
|
| 1148 |
+
"grad_norm": 4.802642822265625,
|
| 1149 |
+
"learning_rate": 4.682144831398563e-05,
|
| 1150 |
+
"loss": 1.0498,
|
| 1151 |
+
"step": 316
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 2.373134328358209,
|
| 1155 |
+
"grad_norm": 3.7944350242614746,
|
| 1156 |
+
"learning_rate": 4.676616915422886e-05,
|
| 1157 |
+
"loss": 1.1943,
|
| 1158 |
+
"step": 318
|
| 1159 |
+
},
|
| 1160 |
+
{
|
| 1161 |
+
"epoch": 2.388059701492537,
|
| 1162 |
+
"grad_norm": 6.245696067810059,
|
| 1163 |
+
"learning_rate": 4.6710889994472086e-05,
|
| 1164 |
+
"loss": 0.9827,
|
| 1165 |
+
"step": 320
|
| 1166 |
+
},
|
| 1167 |
+
{
|
| 1168 |
+
"epoch": 2.4029850746268657,
|
| 1169 |
+
"grad_norm": 3.385528326034546,
|
| 1170 |
+
"learning_rate": 4.6655610834715315e-05,
|
| 1171 |
+
"loss": 1.2411,
|
| 1172 |
+
"step": 322
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 2.417910447761194,
|
| 1176 |
+
"grad_norm": 4.5617828369140625,
|
| 1177 |
+
"learning_rate": 4.660033167495854e-05,
|
| 1178 |
+
"loss": 0.947,
|
| 1179 |
+
"step": 324
|
| 1180 |
+
},
|
| 1181 |
+
{
|
| 1182 |
+
"epoch": 2.4328358208955225,
|
| 1183 |
+
"grad_norm": 6.033749580383301,
|
| 1184 |
+
"learning_rate": 4.654505251520177e-05,
|
| 1185 |
+
"loss": 1.1135,
|
| 1186 |
+
"step": 326
|
| 1187 |
+
},
|
| 1188 |
+
{
|
| 1189 |
+
"epoch": 2.4477611940298507,
|
| 1190 |
+
"grad_norm": 4.091092586517334,
|
| 1191 |
+
"learning_rate": 4.6489773355445e-05,
|
| 1192 |
+
"loss": 1.2219,
|
| 1193 |
+
"step": 328
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 2.4626865671641793,
|
| 1197 |
+
"grad_norm": 4.028707027435303,
|
| 1198 |
+
"learning_rate": 4.643449419568823e-05,
|
| 1199 |
+
"loss": 0.8016,
|
| 1200 |
+
"step": 330
|
| 1201 |
+
},
|
| 1202 |
+
{
|
| 1203 |
+
"epoch": 2.4776119402985075,
|
| 1204 |
+
"grad_norm": 13.044775009155273,
|
| 1205 |
+
"learning_rate": 4.6379215035931455e-05,
|
| 1206 |
+
"loss": 1.0265,
|
| 1207 |
+
"step": 332
|
| 1208 |
+
},
|
| 1209 |
+
{
|
| 1210 |
+
"epoch": 2.4925373134328357,
|
| 1211 |
+
"grad_norm": 3.4753003120422363,
|
| 1212 |
+
"learning_rate": 4.632393587617468e-05,
|
| 1213 |
+
"loss": 1.1795,
|
| 1214 |
+
"step": 334
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 2.5074626865671643,
|
| 1218 |
+
"grad_norm": 10.754464149475098,
|
| 1219 |
+
"learning_rate": 4.626865671641791e-05,
|
| 1220 |
+
"loss": 1.0397,
|
| 1221 |
+
"step": 336
|
| 1222 |
+
},
|
| 1223 |
+
{
|
| 1224 |
+
"epoch": 2.5223880597014925,
|
| 1225 |
+
"grad_norm": 5.801613807678223,
|
| 1226 |
+
"learning_rate": 4.621337755666114e-05,
|
| 1227 |
+
"loss": 1.0849,
|
| 1228 |
+
"step": 338
|
| 1229 |
+
},
|
| 1230 |
+
{
|
| 1231 |
+
"epoch": 2.5373134328358207,
|
| 1232 |
+
"grad_norm": 7.185355186462402,
|
| 1233 |
+
"learning_rate": 4.615809839690437e-05,
|
| 1234 |
+
"loss": 1.3039,
|
| 1235 |
+
"step": 340
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 2.5522388059701493,
|
| 1239 |
+
"grad_norm": 5.9482622146606445,
|
| 1240 |
+
"learning_rate": 4.6102819237147596e-05,
|
| 1241 |
+
"loss": 0.9354,
|
| 1242 |
+
"step": 342
|
| 1243 |
+
},
|
| 1244 |
+
{
|
| 1245 |
+
"epoch": 2.5671641791044775,
|
| 1246 |
+
"grad_norm": 5.785240173339844,
|
| 1247 |
+
"learning_rate": 4.6047540077390824e-05,
|
| 1248 |
+
"loss": 1.1711,
|
| 1249 |
+
"step": 344
|
| 1250 |
+
},
|
| 1251 |
+
{
|
| 1252 |
+
"epoch": 2.582089552238806,
|
| 1253 |
+
"grad_norm": 4.588739395141602,
|
| 1254 |
+
"learning_rate": 4.599226091763406e-05,
|
| 1255 |
+
"loss": 1.2397,
|
| 1256 |
+
"step": 346
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 2.5970149253731343,
|
| 1260 |
+
"grad_norm": 9.58119010925293,
|
| 1261 |
+
"learning_rate": 4.593698175787729e-05,
|
| 1262 |
+
"loss": 1.0885,
|
| 1263 |
+
"step": 348
|
| 1264 |
+
},
|
| 1265 |
+
{
|
| 1266 |
+
"epoch": 2.611940298507463,
|
| 1267 |
+
"grad_norm": 7.116174221038818,
|
| 1268 |
+
"learning_rate": 4.5881702598120515e-05,
|
| 1269 |
+
"loss": 1.4472,
|
| 1270 |
+
"step": 350
|
| 1271 |
+
},
|
| 1272 |
+
{
|
| 1273 |
+
"epoch": 2.626865671641791,
|
| 1274 |
+
"grad_norm": 4.783851146697998,
|
| 1275 |
+
"learning_rate": 4.582642343836374e-05,
|
| 1276 |
+
"loss": 1.0596,
|
| 1277 |
+
"step": 352
|
| 1278 |
+
},
|
| 1279 |
+
{
|
| 1280 |
+
"epoch": 2.6417910447761193,
|
| 1281 |
+
"grad_norm": 8.053864479064941,
|
| 1282 |
+
"learning_rate": 4.577114427860697e-05,
|
| 1283 |
+
"loss": 1.1109,
|
| 1284 |
+
"step": 354
|
| 1285 |
+
},
|
| 1286 |
+
{
|
| 1287 |
+
"epoch": 2.656716417910448,
|
| 1288 |
+
"grad_norm": 7.039917945861816,
|
| 1289 |
+
"learning_rate": 4.571586511885019e-05,
|
| 1290 |
+
"loss": 1.0317,
|
| 1291 |
+
"step": 356
|
| 1292 |
+
},
|
| 1293 |
+
{
|
| 1294 |
+
"epoch": 2.671641791044776,
|
| 1295 |
+
"grad_norm": 10.328742980957031,
|
| 1296 |
+
"learning_rate": 4.566058595909342e-05,
|
| 1297 |
+
"loss": 1.3927,
|
| 1298 |
+
"step": 358
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 2.6865671641791042,
|
| 1302 |
+
"grad_norm": 6.044647693634033,
|
| 1303 |
+
"learning_rate": 4.560530679933665e-05,
|
| 1304 |
+
"loss": 1.6115,
|
| 1305 |
+
"step": 360
|
| 1306 |
+
},
|
| 1307 |
+
{
|
| 1308 |
+
"epoch": 2.701492537313433,
|
| 1309 |
+
"grad_norm": 5.099875450134277,
|
| 1310 |
+
"learning_rate": 4.555002763957988e-05,
|
| 1311 |
+
"loss": 0.7492,
|
| 1312 |
+
"step": 362
|
| 1313 |
+
},
|
| 1314 |
+
{
|
| 1315 |
+
"epoch": 2.716417910447761,
|
| 1316 |
+
"grad_norm": 5.763055324554443,
|
| 1317 |
+
"learning_rate": 4.5494748479823105e-05,
|
| 1318 |
+
"loss": 1.2244,
|
| 1319 |
+
"step": 364
|
| 1320 |
+
},
|
| 1321 |
+
{
|
| 1322 |
+
"epoch": 2.7313432835820897,
|
| 1323 |
+
"grad_norm": 7.4969987869262695,
|
| 1324 |
+
"learning_rate": 4.5439469320066333e-05,
|
| 1325 |
+
"loss": 0.9924,
|
| 1326 |
+
"step": 366
|
| 1327 |
+
},
|
| 1328 |
+
{
|
| 1329 |
+
"epoch": 2.746268656716418,
|
| 1330 |
+
"grad_norm": 11.879778861999512,
|
| 1331 |
+
"learning_rate": 4.538419016030956e-05,
|
| 1332 |
+
"loss": 1.3022,
|
| 1333 |
+
"step": 368
|
| 1334 |
+
},
|
| 1335 |
+
{
|
| 1336 |
+
"epoch": 2.7611940298507465,
|
| 1337 |
+
"grad_norm": 2.6563034057617188,
|
| 1338 |
+
"learning_rate": 4.532891100055279e-05,
|
| 1339 |
+
"loss": 1.0468,
|
| 1340 |
+
"step": 370
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 2.7761194029850746,
|
| 1344 |
+
"grad_norm": 4.352832317352295,
|
| 1345 |
+
"learning_rate": 4.5273631840796025e-05,
|
| 1346 |
+
"loss": 0.7805,
|
| 1347 |
+
"step": 372
|
| 1348 |
+
},
|
| 1349 |
+
{
|
| 1350 |
+
"epoch": 2.791044776119403,
|
| 1351 |
+
"grad_norm": 5.752820014953613,
|
| 1352 |
+
"learning_rate": 4.521835268103925e-05,
|
| 1353 |
+
"loss": 1.0971,
|
| 1354 |
+
"step": 374
|
| 1355 |
+
},
|
| 1356 |
+
{
|
| 1357 |
+
"epoch": 2.8059701492537314,
|
| 1358 |
+
"grad_norm": 6.028427600860596,
|
| 1359 |
+
"learning_rate": 4.516307352128248e-05,
|
| 1360 |
+
"loss": 0.9887,
|
| 1361 |
+
"step": 376
|
| 1362 |
+
},
|
| 1363 |
+
{
|
| 1364 |
+
"epoch": 2.8208955223880596,
|
| 1365 |
+
"grad_norm": 5.541314601898193,
|
| 1366 |
+
"learning_rate": 4.510779436152571e-05,
|
| 1367 |
+
"loss": 1.0044,
|
| 1368 |
+
"step": 378
|
| 1369 |
+
},
|
| 1370 |
+
{
|
| 1371 |
+
"epoch": 2.835820895522388,
|
| 1372 |
+
"grad_norm": 6.708818435668945,
|
| 1373 |
+
"learning_rate": 4.505251520176894e-05,
|
| 1374 |
+
"loss": 0.9756,
|
| 1375 |
+
"step": 380
|
| 1376 |
+
},
|
| 1377 |
+
{
|
| 1378 |
+
"epoch": 2.8507462686567164,
|
| 1379 |
+
"grad_norm": 5.191244602203369,
|
| 1380 |
+
"learning_rate": 4.4997236042012165e-05,
|
| 1381 |
+
"loss": 1.0194,
|
| 1382 |
+
"step": 382
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 2.8656716417910446,
|
| 1386 |
+
"grad_norm": 5.064216613769531,
|
| 1387 |
+
"learning_rate": 4.4941956882255393e-05,
|
| 1388 |
+
"loss": 1.1003,
|
| 1389 |
+
"step": 384
|
| 1390 |
+
},
|
| 1391 |
+
{
|
| 1392 |
+
"epoch": 2.8805970149253732,
|
| 1393 |
+
"grad_norm": 4.351062774658203,
|
| 1394 |
+
"learning_rate": 4.488667772249862e-05,
|
| 1395 |
+
"loss": 1.0402,
|
| 1396 |
+
"step": 386
|
| 1397 |
+
},
|
| 1398 |
+
{
|
| 1399 |
+
"epoch": 2.8955223880597014,
|
| 1400 |
+
"grad_norm": 9.046034812927246,
|
| 1401 |
+
"learning_rate": 4.483139856274185e-05,
|
| 1402 |
+
"loss": 1.0788,
|
| 1403 |
+
"step": 388
|
| 1404 |
+
},
|
| 1405 |
+
{
|
| 1406 |
+
"epoch": 2.91044776119403,
|
| 1407 |
+
"grad_norm": 9.672553062438965,
|
| 1408 |
+
"learning_rate": 4.477611940298508e-05,
|
| 1409 |
+
"loss": 1.1218,
|
| 1410 |
+
"step": 390
|
| 1411 |
+
},
|
| 1412 |
+
{
|
| 1413 |
+
"epoch": 2.925373134328358,
|
| 1414 |
+
"grad_norm": 6.787928581237793,
|
| 1415 |
+
"learning_rate": 4.4720840243228306e-05,
|
| 1416 |
+
"loss": 0.9067,
|
| 1417 |
+
"step": 392
|
| 1418 |
+
},
|
| 1419 |
+
{
|
| 1420 |
+
"epoch": 2.9402985074626864,
|
| 1421 |
+
"grad_norm": 7.9477152824401855,
|
| 1422 |
+
"learning_rate": 4.4665561083471534e-05,
|
| 1423 |
+
"loss": 1.0954,
|
| 1424 |
+
"step": 394
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 2.955223880597015,
|
| 1428 |
+
"grad_norm": 6.69403600692749,
|
| 1429 |
+
"learning_rate": 4.461028192371476e-05,
|
| 1430 |
+
"loss": 1.2409,
|
| 1431 |
+
"step": 396
|
| 1432 |
+
},
|
| 1433 |
+
{
|
| 1434 |
+
"epoch": 2.970149253731343,
|
| 1435 |
+
"grad_norm": 6.419997692108154,
|
| 1436 |
+
"learning_rate": 4.455500276395799e-05,
|
| 1437 |
+
"loss": 0.7133,
|
| 1438 |
+
"step": 398
|
| 1439 |
+
},
|
| 1440 |
+
{
|
| 1441 |
+
"epoch": 2.9850746268656714,
|
| 1442 |
+
"grad_norm": 6.382747173309326,
|
| 1443 |
+
"learning_rate": 4.449972360420122e-05,
|
| 1444 |
+
"loss": 1.3368,
|
| 1445 |
+
"step": 400
|
| 1446 |
+
},
|
| 1447 |
+
{
|
| 1448 |
+
"epoch": 3.0,
|
| 1449 |
+
"grad_norm": 6.085859298706055,
|
| 1450 |
+
"learning_rate": 4.4444444444444447e-05,
|
| 1451 |
+
"loss": 0.8309,
|
| 1452 |
+
"step": 402
|
| 1453 |
+
},
|
| 1454 |
+
{
|
| 1455 |
+
"epoch": 3.0,
|
| 1456 |
+
"eval_accuracy": 0.7,
|
| 1457 |
+
"eval_f1_macro": 0.6561413388812769,
|
| 1458 |
+
"eval_f1_micro": 0.7,
|
| 1459 |
+
"eval_f1_weighted": 0.6696798888246258,
|
| 1460 |
+
"eval_loss": 1.1434708833694458,
|
| 1461 |
+
"eval_precision_macro": 0.6706204141575659,
|
| 1462 |
+
"eval_precision_micro": 0.7,
|
| 1463 |
+
"eval_precision_weighted": 0.6840218575284365,
|
| 1464 |
+
"eval_recall_macro": 0.6900560224089635,
|
| 1465 |
+
"eval_recall_micro": 0.7,
|
| 1466 |
+
"eval_recall_weighted": 0.7,
|
| 1467 |
+
"eval_runtime": 36.7029,
|
| 1468 |
+
"eval_samples_per_second": 5.449,
|
| 1469 |
+
"eval_steps_per_second": 0.191,
|
| 1470 |
+
"step": 402
|
| 1471 |
+
},
|
| 1472 |
+
{
|
| 1473 |
+
"epoch": 3.014925373134328,
|
| 1474 |
+
"grad_norm": 8.274943351745605,
|
| 1475 |
+
"learning_rate": 4.4389165284687675e-05,
|
| 1476 |
+
"loss": 0.728,
|
| 1477 |
+
"step": 404
|
| 1478 |
+
},
|
| 1479 |
+
{
|
| 1480 |
+
"epoch": 3.029850746268657,
|
| 1481 |
+
"grad_norm": 3.896862745285034,
|
| 1482 |
+
"learning_rate": 4.43338861249309e-05,
|
| 1483 |
+
"loss": 0.8955,
|
| 1484 |
+
"step": 406
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 3.044776119402985,
|
| 1488 |
+
"grad_norm": 3.3480184078216553,
|
| 1489 |
+
"learning_rate": 4.427860696517413e-05,
|
| 1490 |
+
"loss": 0.84,
|
| 1491 |
+
"step": 408
|
| 1492 |
+
},
|
| 1493 |
+
{
|
| 1494 |
+
"epoch": 3.0597014925373136,
|
| 1495 |
+
"grad_norm": 9.12331485748291,
|
| 1496 |
+
"learning_rate": 4.422332780541736e-05,
|
| 1497 |
+
"loss": 0.8026,
|
| 1498 |
+
"step": 410
|
| 1499 |
+
},
|
| 1500 |
+
{
|
| 1501 |
+
"epoch": 3.074626865671642,
|
| 1502 |
+
"grad_norm": 6.628831386566162,
|
| 1503 |
+
"learning_rate": 4.416804864566059e-05,
|
| 1504 |
+
"loss": 1.0466,
|
| 1505 |
+
"step": 412
|
| 1506 |
+
},
|
| 1507 |
+
{
|
| 1508 |
+
"epoch": 3.08955223880597,
|
| 1509 |
+
"grad_norm": 2.638613224029541,
|
| 1510 |
+
"learning_rate": 4.4112769485903815e-05,
|
| 1511 |
+
"loss": 0.6834,
|
| 1512 |
+
"step": 414
|
| 1513 |
+
},
|
| 1514 |
+
{
|
| 1515 |
+
"epoch": 3.1044776119402986,
|
| 1516 |
+
"grad_norm": 3.1941006183624268,
|
| 1517 |
+
"learning_rate": 4.4057490326147043e-05,
|
| 1518 |
+
"loss": 0.9236,
|
| 1519 |
+
"step": 416
|
| 1520 |
+
},
|
| 1521 |
+
{
|
| 1522 |
+
"epoch": 3.1194029850746268,
|
| 1523 |
+
"grad_norm": 4.337372779846191,
|
| 1524 |
+
"learning_rate": 4.400221116639027e-05,
|
| 1525 |
+
"loss": 0.8447,
|
| 1526 |
+
"step": 418
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 3.1343283582089554,
|
| 1530 |
+
"grad_norm": 8.465522766113281,
|
| 1531 |
+
"learning_rate": 4.39469320066335e-05,
|
| 1532 |
+
"loss": 0.8768,
|
| 1533 |
+
"step": 420
|
| 1534 |
+
},
|
| 1535 |
+
{
|
| 1536 |
+
"epoch": 3.1492537313432836,
|
| 1537 |
+
"grad_norm": 4.252687454223633,
|
| 1538 |
+
"learning_rate": 4.389165284687673e-05,
|
| 1539 |
+
"loss": 0.7148,
|
| 1540 |
+
"step": 422
|
| 1541 |
+
},
|
| 1542 |
+
{
|
| 1543 |
+
"epoch": 3.1641791044776117,
|
| 1544 |
+
"grad_norm": 5.22479772567749,
|
| 1545 |
+
"learning_rate": 4.3836373687119956e-05,
|
| 1546 |
+
"loss": 0.7375,
|
| 1547 |
+
"step": 424
|
| 1548 |
+
},
|
| 1549 |
+
{
|
| 1550 |
+
"epoch": 3.1791044776119404,
|
| 1551 |
+
"grad_norm": 3.702615737915039,
|
| 1552 |
+
"learning_rate": 4.3781094527363184e-05,
|
| 1553 |
+
"loss": 0.8488,
|
| 1554 |
+
"step": 426
|
| 1555 |
+
},
|
| 1556 |
+
{
|
| 1557 |
+
"epoch": 3.1940298507462686,
|
| 1558 |
+
"grad_norm": 3.2345383167266846,
|
| 1559 |
+
"learning_rate": 4.372581536760641e-05,
|
| 1560 |
+
"loss": 0.7861,
|
| 1561 |
+
"step": 428
|
| 1562 |
+
},
|
| 1563 |
+
{
|
| 1564 |
+
"epoch": 3.208955223880597,
|
| 1565 |
+
"grad_norm": 7.831496238708496,
|
| 1566 |
+
"learning_rate": 4.367053620784965e-05,
|
| 1567 |
+
"loss": 1.117,
|
| 1568 |
+
"step": 430
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 3.2238805970149254,
|
| 1572 |
+
"grad_norm": 2.40031361579895,
|
| 1573 |
+
"learning_rate": 4.3615257048092875e-05,
|
| 1574 |
+
"loss": 0.7079,
|
| 1575 |
+
"step": 432
|
| 1576 |
+
},
|
| 1577 |
+
{
|
| 1578 |
+
"epoch": 3.2388059701492535,
|
| 1579 |
+
"grad_norm": 4.2407050132751465,
|
| 1580 |
+
"learning_rate": 4.3559977888336103e-05,
|
| 1581 |
+
"loss": 0.7778,
|
| 1582 |
+
"step": 434
|
| 1583 |
+
},
|
| 1584 |
+
{
|
| 1585 |
+
"epoch": 3.253731343283582,
|
| 1586 |
+
"grad_norm": 3.885312795639038,
|
| 1587 |
+
"learning_rate": 4.350469872857933e-05,
|
| 1588 |
+
"loss": 1.0815,
|
| 1589 |
+
"step": 436
|
| 1590 |
+
},
|
| 1591 |
+
{
|
| 1592 |
+
"epoch": 3.2686567164179103,
|
| 1593 |
+
"grad_norm": 3.0180788040161133,
|
| 1594 |
+
"learning_rate": 4.344941956882256e-05,
|
| 1595 |
+
"loss": 0.6417,
|
| 1596 |
+
"step": 438
|
| 1597 |
+
},
|
| 1598 |
+
{
|
| 1599 |
+
"epoch": 3.283582089552239,
|
| 1600 |
+
"grad_norm": 6.578612804412842,
|
| 1601 |
+
"learning_rate": 4.339414040906579e-05,
|
| 1602 |
+
"loss": 0.9152,
|
| 1603 |
+
"step": 440
|
| 1604 |
+
},
|
| 1605 |
+
{
|
| 1606 |
+
"epoch": 3.298507462686567,
|
| 1607 |
+
"grad_norm": 4.349058151245117,
|
| 1608 |
+
"learning_rate": 4.3338861249309016e-05,
|
| 1609 |
+
"loss": 0.9936,
|
| 1610 |
+
"step": 442
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 3.3134328358208958,
|
| 1614 |
+
"grad_norm": 7.187576770782471,
|
| 1615 |
+
"learning_rate": 4.328358208955224e-05,
|
| 1616 |
+
"loss": 0.9757,
|
| 1617 |
+
"step": 444
|
| 1618 |
+
},
|
| 1619 |
+
{
|
| 1620 |
+
"epoch": 3.328358208955224,
|
| 1621 |
+
"grad_norm": 3.981358528137207,
|
| 1622 |
+
"learning_rate": 4.3228302929795465e-05,
|
| 1623 |
+
"loss": 0.6527,
|
| 1624 |
+
"step": 446
|
| 1625 |
+
},
|
| 1626 |
+
{
|
| 1627 |
+
"epoch": 3.343283582089552,
|
| 1628 |
+
"grad_norm": 4.734504222869873,
|
| 1629 |
+
"learning_rate": 4.3173023770038694e-05,
|
| 1630 |
+
"loss": 0.917,
|
| 1631 |
+
"step": 448
|
| 1632 |
+
},
|
| 1633 |
+
{
|
| 1634 |
+
"epoch": 3.3582089552238807,
|
| 1635 |
+
"grad_norm": 7.110349655151367,
|
| 1636 |
+
"learning_rate": 4.311774461028192e-05,
|
| 1637 |
+
"loss": 0.7495,
|
| 1638 |
+
"step": 450
|
| 1639 |
+
},
|
| 1640 |
+
{
|
| 1641 |
+
"epoch": 3.373134328358209,
|
| 1642 |
+
"grad_norm": 4.323439121246338,
|
| 1643 |
+
"learning_rate": 4.306246545052515e-05,
|
| 1644 |
+
"loss": 0.7577,
|
| 1645 |
+
"step": 452
|
| 1646 |
+
},
|
| 1647 |
+
{
|
| 1648 |
+
"epoch": 3.388059701492537,
|
| 1649 |
+
"grad_norm": 6.730025768280029,
|
| 1650 |
+
"learning_rate": 4.300718629076838e-05,
|
| 1651 |
+
"loss": 1.2703,
|
| 1652 |
+
"step": 454
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 3.4029850746268657,
|
| 1656 |
+
"grad_norm": 10.35930347442627,
|
| 1657 |
+
"learning_rate": 4.2951907131011606e-05,
|
| 1658 |
+
"loss": 0.4539,
|
| 1659 |
+
"step": 456
|
| 1660 |
+
},
|
| 1661 |
+
{
|
| 1662 |
+
"epoch": 3.417910447761194,
|
| 1663 |
+
"grad_norm": 3.9190192222595215,
|
| 1664 |
+
"learning_rate": 4.289662797125484e-05,
|
| 1665 |
+
"loss": 0.8863,
|
| 1666 |
+
"step": 458
|
| 1667 |
+
},
|
| 1668 |
+
{
|
| 1669 |
+
"epoch": 3.4328358208955225,
|
| 1670 |
+
"grad_norm": 6.045052528381348,
|
| 1671 |
+
"learning_rate": 4.284134881149807e-05,
|
| 1672 |
+
"loss": 0.8221,
|
| 1673 |
+
"step": 460
|
| 1674 |
+
},
|
| 1675 |
+
{
|
| 1676 |
+
"epoch": 3.4477611940298507,
|
| 1677 |
+
"grad_norm": 9.53062629699707,
|
| 1678 |
+
"learning_rate": 4.27860696517413e-05,
|
| 1679 |
+
"loss": 0.9652,
|
| 1680 |
+
"step": 462
|
| 1681 |
+
},
|
| 1682 |
+
{
|
| 1683 |
+
"epoch": 3.4626865671641793,
|
| 1684 |
+
"grad_norm": 5.4133710861206055,
|
| 1685 |
+
"learning_rate": 4.2730790491984525e-05,
|
| 1686 |
+
"loss": 0.6179,
|
| 1687 |
+
"step": 464
|
| 1688 |
+
},
|
| 1689 |
+
{
|
| 1690 |
+
"epoch": 3.4776119402985075,
|
| 1691 |
+
"grad_norm": 5.08922004699707,
|
| 1692 |
+
"learning_rate": 4.2675511332227754e-05,
|
| 1693 |
+
"loss": 0.8349,
|
| 1694 |
+
"step": 466
|
| 1695 |
+
},
|
| 1696 |
+
{
|
| 1697 |
+
"epoch": 3.4925373134328357,
|
| 1698 |
+
"grad_norm": 9.440414428710938,
|
| 1699 |
+
"learning_rate": 4.262023217247098e-05,
|
| 1700 |
+
"loss": 0.7352,
|
| 1701 |
+
"step": 468
|
| 1702 |
+
},
|
| 1703 |
+
{
|
| 1704 |
+
"epoch": 3.5074626865671643,
|
| 1705 |
+
"grad_norm": 5.086453437805176,
|
| 1706 |
+
"learning_rate": 4.256495301271421e-05,
|
| 1707 |
+
"loss": 0.6674,
|
| 1708 |
+
"step": 470
|
| 1709 |
+
},
|
| 1710 |
+
{
|
| 1711 |
+
"epoch": 3.5223880597014925,
|
| 1712 |
+
"grad_norm": 5.283850193023682,
|
| 1713 |
+
"learning_rate": 4.250967385295744e-05,
|
| 1714 |
+
"loss": 0.9129,
|
| 1715 |
+
"step": 472
|
| 1716 |
+
},
|
| 1717 |
+
{
|
| 1718 |
+
"epoch": 3.5373134328358207,
|
| 1719 |
+
"grad_norm": 6.586909770965576,
|
| 1720 |
+
"learning_rate": 4.2454394693200666e-05,
|
| 1721 |
+
"loss": 0.4512,
|
| 1722 |
+
"step": 474
|
| 1723 |
+
},
|
| 1724 |
+
{
|
| 1725 |
+
"epoch": 3.5522388059701493,
|
| 1726 |
+
"grad_norm": 6.328141212463379,
|
| 1727 |
+
"learning_rate": 4.2399115533443894e-05,
|
| 1728 |
+
"loss": 0.9986,
|
| 1729 |
+
"step": 476
|
| 1730 |
+
},
|
| 1731 |
+
{
|
| 1732 |
+
"epoch": 3.5671641791044775,
|
| 1733 |
+
"grad_norm": 11.822186470031738,
|
| 1734 |
+
"learning_rate": 4.234383637368712e-05,
|
| 1735 |
+
"loss": 0.6238,
|
| 1736 |
+
"step": 478
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"epoch": 3.582089552238806,
|
| 1740 |
+
"grad_norm": 12.303596496582031,
|
| 1741 |
+
"learning_rate": 4.228855721393035e-05,
|
| 1742 |
+
"loss": 0.666,
|
| 1743 |
+
"step": 480
|
| 1744 |
+
},
|
| 1745 |
+
{
|
| 1746 |
+
"epoch": 3.5970149253731343,
|
| 1747 |
+
"grad_norm": 10.991721153259277,
|
| 1748 |
+
"learning_rate": 4.223327805417358e-05,
|
| 1749 |
+
"loss": 0.8176,
|
| 1750 |
+
"step": 482
|
| 1751 |
+
},
|
| 1752 |
+
{
|
| 1753 |
+
"epoch": 3.611940298507463,
|
| 1754 |
+
"grad_norm": 4.936182022094727,
|
| 1755 |
+
"learning_rate": 4.217799889441681e-05,
|
| 1756 |
+
"loss": 0.6964,
|
| 1757 |
+
"step": 484
|
| 1758 |
+
},
|
| 1759 |
+
{
|
| 1760 |
+
"epoch": 3.626865671641791,
|
| 1761 |
+
"grad_norm": 4.455245494842529,
|
| 1762 |
+
"learning_rate": 4.2122719734660035e-05,
|
| 1763 |
+
"loss": 0.8681,
|
| 1764 |
+
"step": 486
|
| 1765 |
+
},
|
| 1766 |
+
{
|
| 1767 |
+
"epoch": 3.6417910447761193,
|
| 1768 |
+
"grad_norm": 4.1669206619262695,
|
| 1769 |
+
"learning_rate": 4.206744057490326e-05,
|
| 1770 |
+
"loss": 0.8556,
|
| 1771 |
+
"step": 488
|
| 1772 |
+
},
|
| 1773 |
+
{
|
| 1774 |
+
"epoch": 3.656716417910448,
|
| 1775 |
+
"grad_norm": 2.757246732711792,
|
| 1776 |
+
"learning_rate": 4.201216141514649e-05,
|
| 1777 |
+
"loss": 0.6557,
|
| 1778 |
+
"step": 490
|
| 1779 |
+
},
|
| 1780 |
+
{
|
| 1781 |
+
"epoch": 3.671641791044776,
|
| 1782 |
+
"grad_norm": 4.065201282501221,
|
| 1783 |
+
"learning_rate": 4.195688225538972e-05,
|
| 1784 |
+
"loss": 0.7899,
|
| 1785 |
+
"step": 492
|
| 1786 |
+
},
|
| 1787 |
+
{
|
| 1788 |
+
"epoch": 3.6865671641791042,
|
| 1789 |
+
"grad_norm": 5.666295528411865,
|
| 1790 |
+
"learning_rate": 4.190160309563295e-05,
|
| 1791 |
+
"loss": 0.7589,
|
| 1792 |
+
"step": 494
|
| 1793 |
+
},
|
| 1794 |
+
{
|
| 1795 |
+
"epoch": 3.701492537313433,
|
| 1796 |
+
"grad_norm": 4.929719924926758,
|
| 1797 |
+
"learning_rate": 4.1846323935876176e-05,
|
| 1798 |
+
"loss": 0.6709,
|
| 1799 |
+
"step": 496
|
| 1800 |
+
},
|
| 1801 |
+
{
|
| 1802 |
+
"epoch": 3.716417910447761,
|
| 1803 |
+
"grad_norm": 6.261268138885498,
|
| 1804 |
+
"learning_rate": 4.1791044776119404e-05,
|
| 1805 |
+
"loss": 0.4849,
|
| 1806 |
+
"step": 498
|
| 1807 |
+
},
|
| 1808 |
+
{
|
| 1809 |
+
"epoch": 3.7313432835820897,
|
| 1810 |
+
"grad_norm": 1.9559329748153687,
|
| 1811 |
+
"learning_rate": 4.173576561636263e-05,
|
| 1812 |
+
"loss": 0.4897,
|
| 1813 |
+
"step": 500
|
| 1814 |
+
},
|
| 1815 |
+
{
|
| 1816 |
+
"epoch": 3.746268656716418,
|
| 1817 |
+
"grad_norm": 3.6171419620513916,
|
| 1818 |
+
"learning_rate": 4.168048645660586e-05,
|
| 1819 |
+
"loss": 0.4541,
|
| 1820 |
+
"step": 502
|
| 1821 |
+
},
|
| 1822 |
+
{
|
| 1823 |
+
"epoch": 3.7611940298507465,
|
| 1824 |
+
"grad_norm": 5.1620988845825195,
|
| 1825 |
+
"learning_rate": 4.162520729684909e-05,
|
| 1826 |
+
"loss": 0.4952,
|
| 1827 |
+
"step": 504
|
| 1828 |
+
},
|
| 1829 |
+
{
|
| 1830 |
+
"epoch": 3.7761194029850746,
|
| 1831 |
+
"grad_norm": 6.901045322418213,
|
| 1832 |
+
"learning_rate": 4.1569928137092316e-05,
|
| 1833 |
+
"loss": 0.8728,
|
| 1834 |
+
"step": 506
|
| 1835 |
+
},
|
| 1836 |
+
{
|
| 1837 |
+
"epoch": 3.791044776119403,
|
| 1838 |
+
"grad_norm": 6.784924507141113,
|
| 1839 |
+
"learning_rate": 4.1514648977335544e-05,
|
| 1840 |
+
"loss": 1.0379,
|
| 1841 |
+
"step": 508
|
| 1842 |
+
},
|
| 1843 |
+
{
|
| 1844 |
+
"epoch": 3.8059701492537314,
|
| 1845 |
+
"grad_norm": 3.829332113265991,
|
| 1846 |
+
"learning_rate": 4.145936981757877e-05,
|
| 1847 |
+
"loss": 0.8488,
|
| 1848 |
+
"step": 510
|
| 1849 |
+
},
|
| 1850 |
+
{
|
| 1851 |
+
"epoch": 3.8208955223880596,
|
| 1852 |
+
"grad_norm": 8.228715896606445,
|
| 1853 |
+
"learning_rate": 4.1404090657822e-05,
|
| 1854 |
+
"loss": 0.4737,
|
| 1855 |
+
"step": 512
|
| 1856 |
+
},
|
| 1857 |
+
{
|
| 1858 |
+
"epoch": 3.835820895522388,
|
| 1859 |
+
"grad_norm": 7.089122295379639,
|
| 1860 |
+
"learning_rate": 4.134881149806523e-05,
|
| 1861 |
+
"loss": 0.8644,
|
| 1862 |
+
"step": 514
|
| 1863 |
+
},
|
| 1864 |
+
{
|
| 1865 |
+
"epoch": 3.8507462686567164,
|
| 1866 |
+
"grad_norm": 6.432941913604736,
|
| 1867 |
+
"learning_rate": 4.1293532338308464e-05,
|
| 1868 |
+
"loss": 0.5153,
|
| 1869 |
+
"step": 516
|
| 1870 |
+
},
|
| 1871 |
+
{
|
| 1872 |
+
"epoch": 3.8656716417910446,
|
| 1873 |
+
"grad_norm": 8.107361793518066,
|
| 1874 |
+
"learning_rate": 4.123825317855169e-05,
|
| 1875 |
+
"loss": 0.6265,
|
| 1876 |
+
"step": 518
|
| 1877 |
+
},
|
| 1878 |
+
{
|
| 1879 |
+
"epoch": 3.8805970149253732,
|
| 1880 |
+
"grad_norm": 4.366127014160156,
|
| 1881 |
+
"learning_rate": 4.118297401879492e-05,
|
| 1882 |
+
"loss": 0.9007,
|
| 1883 |
+
"step": 520
|
| 1884 |
+
},
|
| 1885 |
+
{
|
| 1886 |
+
"epoch": 3.8955223880597014,
|
| 1887 |
+
"grad_norm": 3.596374988555908,
|
| 1888 |
+
"learning_rate": 4.112769485903815e-05,
|
| 1889 |
+
"loss": 0.7653,
|
| 1890 |
+
"step": 522
|
| 1891 |
+
},
|
| 1892 |
+
{
|
| 1893 |
+
"epoch": 3.91044776119403,
|
| 1894 |
+
"grad_norm": 6.282411098480225,
|
| 1895 |
+
"learning_rate": 4.1072415699281376e-05,
|
| 1896 |
+
"loss": 0.8757,
|
| 1897 |
+
"step": 524
|
| 1898 |
+
},
|
| 1899 |
+
{
|
| 1900 |
+
"epoch": 3.925373134328358,
|
| 1901 |
+
"grad_norm": 6.818746566772461,
|
| 1902 |
+
"learning_rate": 4.1017136539524604e-05,
|
| 1903 |
+
"loss": 0.7569,
|
| 1904 |
+
"step": 526
|
| 1905 |
+
},
|
| 1906 |
+
{
|
| 1907 |
+
"epoch": 3.9402985074626864,
|
| 1908 |
+
"grad_norm": 6.6983113288879395,
|
| 1909 |
+
"learning_rate": 4.096185737976783e-05,
|
| 1910 |
+
"loss": 0.79,
|
| 1911 |
+
"step": 528
|
| 1912 |
+
},
|
| 1913 |
+
{
|
| 1914 |
+
"epoch": 3.955223880597015,
|
| 1915 |
+
"grad_norm": 3.532710075378418,
|
| 1916 |
+
"learning_rate": 4.090657822001106e-05,
|
| 1917 |
+
"loss": 0.6981,
|
| 1918 |
+
"step": 530
|
| 1919 |
+
},
|
| 1920 |
+
{
|
| 1921 |
+
"epoch": 3.970149253731343,
|
| 1922 |
+
"grad_norm": 6.195641040802002,
|
| 1923 |
+
"learning_rate": 4.085129906025428e-05,
|
| 1924 |
+
"loss": 0.8261,
|
| 1925 |
+
"step": 532
|
| 1926 |
+
},
|
| 1927 |
+
{
|
| 1928 |
+
"epoch": 3.9850746268656714,
|
| 1929 |
+
"grad_norm": 8.55378246307373,
|
| 1930 |
+
"learning_rate": 4.079601990049751e-05,
|
| 1931 |
+
"loss": 0.7682,
|
| 1932 |
+
"step": 534
|
| 1933 |
+
},
|
| 1934 |
+
{
|
| 1935 |
+
"epoch": 4.0,
|
| 1936 |
+
"grad_norm": 7.6657891273498535,
|
| 1937 |
+
"learning_rate": 4.074074074074074e-05,
|
| 1938 |
+
"loss": 0.6755,
|
| 1939 |
+
"step": 536
|
| 1940 |
+
},
|
| 1941 |
+
{
|
| 1942 |
+
"epoch": 4.0,
|
| 1943 |
+
"eval_accuracy": 0.74,
|
| 1944 |
+
"eval_f1_macro": 0.6978798162621692,
|
| 1945 |
+
"eval_f1_micro": 0.74,
|
| 1946 |
+
"eval_f1_weighted": 0.7101990093240095,
|
| 1947 |
+
"eval_loss": 0.93646240234375,
|
| 1948 |
+
"eval_precision_macro": 0.7079046770223241,
|
| 1949 |
+
"eval_precision_micro": 0.74,
|
| 1950 |
+
"eval_precision_weighted": 0.719476911976912,
|
| 1951 |
+
"eval_recall_macro": 0.7278711484593837,
|
| 1952 |
+
"eval_recall_micro": 0.74,
|
| 1953 |
+
"eval_recall_weighted": 0.74,
|
| 1954 |
+
"eval_runtime": 40.7905,
|
| 1955 |
+
"eval_samples_per_second": 4.903,
|
| 1956 |
+
"eval_steps_per_second": 0.172,
|
| 1957 |
+
"step": 536
|
| 1958 |
+
},
|
| 1959 |
+
{
|
| 1960 |
+
"epoch": 4.014925373134329,
|
| 1961 |
+
"grad_norm": 3.920659065246582,
|
| 1962 |
+
"learning_rate": 4.0685461580983966e-05,
|
| 1963 |
+
"loss": 0.5528,
|
| 1964 |
+
"step": 538
|
| 1965 |
+
},
|
| 1966 |
+
{
|
| 1967 |
+
"epoch": 4.029850746268656,
|
| 1968 |
+
"grad_norm": 5.147222995758057,
|
| 1969 |
+
"learning_rate": 4.0630182421227194e-05,
|
| 1970 |
+
"loss": 1.0055,
|
| 1971 |
+
"step": 540
|
| 1972 |
+
},
|
| 1973 |
+
{
|
| 1974 |
+
"epoch": 4.044776119402985,
|
| 1975 |
+
"grad_norm": 2.260542869567871,
|
| 1976 |
+
"learning_rate": 4.057490326147043e-05,
|
| 1977 |
+
"loss": 0.6422,
|
| 1978 |
+
"step": 542
|
| 1979 |
+
},
|
| 1980 |
+
{
|
| 1981 |
+
"epoch": 4.059701492537314,
|
| 1982 |
+
"grad_norm": 5.286315441131592,
|
| 1983 |
+
"learning_rate": 4.051962410171366e-05,
|
| 1984 |
+
"loss": 0.5399,
|
| 1985 |
+
"step": 544
|
| 1986 |
+
},
|
| 1987 |
+
{
|
| 1988 |
+
"epoch": 4.074626865671641,
|
| 1989 |
+
"grad_norm": 8.412203788757324,
|
| 1990 |
+
"learning_rate": 4.0464344941956886e-05,
|
| 1991 |
+
"loss": 0.6129,
|
| 1992 |
+
"step": 546
|
| 1993 |
+
},
|
| 1994 |
+
{
|
| 1995 |
+
"epoch": 4.08955223880597,
|
| 1996 |
+
"grad_norm": 6.18348503112793,
|
| 1997 |
+
"learning_rate": 4.0409065782200114e-05,
|
| 1998 |
+
"loss": 0.5879,
|
| 1999 |
+
"step": 548
|
| 2000 |
+
},
|
| 2001 |
+
{
|
| 2002 |
+
"epoch": 4.104477611940299,
|
| 2003 |
+
"grad_norm": 2.4661457538604736,
|
| 2004 |
+
"learning_rate": 4.035378662244334e-05,
|
| 2005 |
+
"loss": 0.5941,
|
| 2006 |
+
"step": 550
|
| 2007 |
+
},
|
| 2008 |
+
{
|
| 2009 |
+
"epoch": 4.119402985074627,
|
| 2010 |
+
"grad_norm": 4.261096477508545,
|
| 2011 |
+
"learning_rate": 4.029850746268657e-05,
|
| 2012 |
+
"loss": 0.9773,
|
| 2013 |
+
"step": 552
|
| 2014 |
+
},
|
| 2015 |
+
{
|
| 2016 |
+
"epoch": 4.134328358208955,
|
| 2017 |
+
"grad_norm": 6.717421054840088,
|
| 2018 |
+
"learning_rate": 4.02432283029298e-05,
|
| 2019 |
+
"loss": 0.7255,
|
| 2020 |
+
"step": 554
|
| 2021 |
+
},
|
| 2022 |
+
{
|
| 2023 |
+
"epoch": 4.149253731343284,
|
| 2024 |
+
"grad_norm": 4.847568511962891,
|
| 2025 |
+
"learning_rate": 4.0187949143173026e-05,
|
| 2026 |
+
"loss": 0.6345,
|
| 2027 |
+
"step": 556
|
| 2028 |
+
},
|
| 2029 |
+
{
|
| 2030 |
+
"epoch": 4.164179104477612,
|
| 2031 |
+
"grad_norm": 3.6066529750823975,
|
| 2032 |
+
"learning_rate": 4.0132669983416254e-05,
|
| 2033 |
+
"loss": 0.4784,
|
| 2034 |
+
"step": 558
|
| 2035 |
+
},
|
| 2036 |
+
{
|
| 2037 |
+
"epoch": 4.17910447761194,
|
| 2038 |
+
"grad_norm": 1.8522593975067139,
|
| 2039 |
+
"learning_rate": 4.007739082365948e-05,
|
| 2040 |
+
"loss": 0.4678,
|
| 2041 |
+
"step": 560
|
| 2042 |
+
},
|
| 2043 |
+
{
|
| 2044 |
+
"epoch": 4.1940298507462686,
|
| 2045 |
+
"grad_norm": 7.491433620452881,
|
| 2046 |
+
"learning_rate": 4.002211166390271e-05,
|
| 2047 |
+
"loss": 0.6955,
|
| 2048 |
+
"step": 562
|
| 2049 |
+
},
|
| 2050 |
+
{
|
| 2051 |
+
"epoch": 4.208955223880597,
|
| 2052 |
+
"grad_norm": 3.9711689949035645,
|
| 2053 |
+
"learning_rate": 3.996683250414594e-05,
|
| 2054 |
+
"loss": 0.7432,
|
| 2055 |
+
"step": 564
|
| 2056 |
+
},
|
| 2057 |
+
{
|
| 2058 |
+
"epoch": 4.223880597014926,
|
| 2059 |
+
"grad_norm": 3.5602264404296875,
|
| 2060 |
+
"learning_rate": 3.991155334438917e-05,
|
| 2061 |
+
"loss": 0.6618,
|
| 2062 |
+
"step": 566
|
| 2063 |
+
},
|
| 2064 |
+
{
|
| 2065 |
+
"epoch": 4.2388059701492535,
|
| 2066 |
+
"grad_norm": 3.383795976638794,
|
| 2067 |
+
"learning_rate": 3.9856274184632395e-05,
|
| 2068 |
+
"loss": 0.6079,
|
| 2069 |
+
"step": 568
|
| 2070 |
+
},
|
| 2071 |
+
{
|
| 2072 |
+
"epoch": 4.253731343283582,
|
| 2073 |
+
"grad_norm": 2.320563554763794,
|
| 2074 |
+
"learning_rate": 3.980099502487562e-05,
|
| 2075 |
+
"loss": 0.5248,
|
| 2076 |
+
"step": 570
|
| 2077 |
+
},
|
| 2078 |
+
{
|
| 2079 |
+
"epoch": 4.268656716417911,
|
| 2080 |
+
"grad_norm": 3.110429048538208,
|
| 2081 |
+
"learning_rate": 3.974571586511885e-05,
|
| 2082 |
+
"loss": 0.3517,
|
| 2083 |
+
"step": 572
|
| 2084 |
+
},
|
| 2085 |
+
{
|
| 2086 |
+
"epoch": 4.2835820895522385,
|
| 2087 |
+
"grad_norm": 6.524468421936035,
|
| 2088 |
+
"learning_rate": 3.9690436705362086e-05,
|
| 2089 |
+
"loss": 0.6481,
|
| 2090 |
+
"step": 574
|
| 2091 |
+
},
|
| 2092 |
+
{
|
| 2093 |
+
"epoch": 4.298507462686567,
|
| 2094 |
+
"grad_norm": 3.3383541107177734,
|
| 2095 |
+
"learning_rate": 3.9635157545605314e-05,
|
| 2096 |
+
"loss": 0.6297,
|
| 2097 |
+
"step": 576
|
| 2098 |
+
},
|
| 2099 |
+
{
|
| 2100 |
+
"epoch": 4.313432835820896,
|
| 2101 |
+
"grad_norm": 1.686207890510559,
|
| 2102 |
+
"learning_rate": 3.9579878385848536e-05,
|
| 2103 |
+
"loss": 0.5142,
|
| 2104 |
+
"step": 578
|
| 2105 |
+
},
|
| 2106 |
+
{
|
| 2107 |
+
"epoch": 4.3283582089552235,
|
| 2108 |
+
"grad_norm": 9.0740385055542,
|
| 2109 |
+
"learning_rate": 3.9524599226091764e-05,
|
| 2110 |
+
"loss": 0.478,
|
| 2111 |
+
"step": 580
|
| 2112 |
+
},
|
| 2113 |
+
{
|
| 2114 |
+
"epoch": 4.343283582089552,
|
| 2115 |
+
"grad_norm": 8.83989143371582,
|
| 2116 |
+
"learning_rate": 3.946932006633499e-05,
|
| 2117 |
+
"loss": 0.7348,
|
| 2118 |
+
"step": 582
|
| 2119 |
+
},
|
| 2120 |
+
{
|
| 2121 |
+
"epoch": 4.358208955223881,
|
| 2122 |
+
"grad_norm": 9.292248725891113,
|
| 2123 |
+
"learning_rate": 3.941404090657822e-05,
|
| 2124 |
+
"loss": 0.6026,
|
| 2125 |
+
"step": 584
|
| 2126 |
+
},
|
| 2127 |
+
{
|
| 2128 |
+
"epoch": 4.373134328358209,
|
| 2129 |
+
"grad_norm": 4.79518461227417,
|
| 2130 |
+
"learning_rate": 3.935876174682145e-05,
|
| 2131 |
+
"loss": 0.818,
|
| 2132 |
+
"step": 586
|
| 2133 |
+
},
|
| 2134 |
+
{
|
| 2135 |
+
"epoch": 4.388059701492537,
|
| 2136 |
+
"grad_norm": 6.292643070220947,
|
| 2137 |
+
"learning_rate": 3.9303482587064676e-05,
|
| 2138 |
+
"loss": 0.5522,
|
| 2139 |
+
"step": 588
|
| 2140 |
+
},
|
| 2141 |
+
{
|
| 2142 |
+
"epoch": 4.402985074626866,
|
| 2143 |
+
"grad_norm": 4.075753688812256,
|
| 2144 |
+
"learning_rate": 3.9248203427307905e-05,
|
| 2145 |
+
"loss": 0.5259,
|
| 2146 |
+
"step": 590
|
| 2147 |
+
},
|
| 2148 |
+
{
|
| 2149 |
+
"epoch": 4.417910447761194,
|
| 2150 |
+
"grad_norm": 2.985786199569702,
|
| 2151 |
+
"learning_rate": 3.919292426755113e-05,
|
| 2152 |
+
"loss": 0.4218,
|
| 2153 |
+
"step": 592
|
| 2154 |
+
},
|
| 2155 |
+
{
|
| 2156 |
+
"epoch": 4.432835820895522,
|
| 2157 |
+
"grad_norm": 4.570261478424072,
|
| 2158 |
+
"learning_rate": 3.913764510779436e-05,
|
| 2159 |
+
"loss": 0.4735,
|
| 2160 |
+
"step": 594
|
| 2161 |
+
},
|
| 2162 |
+
{
|
| 2163 |
+
"epoch": 4.447761194029851,
|
| 2164 |
+
"grad_norm": 2.8057727813720703,
|
| 2165 |
+
"learning_rate": 3.908236594803759e-05,
|
| 2166 |
+
"loss": 0.4241,
|
| 2167 |
+
"step": 596
|
| 2168 |
+
},
|
| 2169 |
+
{
|
| 2170 |
+
"epoch": 4.462686567164179,
|
| 2171 |
+
"grad_norm": 7.44076681137085,
|
| 2172 |
+
"learning_rate": 3.902708678828082e-05,
|
| 2173 |
+
"loss": 0.4751,
|
| 2174 |
+
"step": 598
|
| 2175 |
+
},
|
| 2176 |
+
{
|
| 2177 |
+
"epoch": 4.477611940298507,
|
| 2178 |
+
"grad_norm": 4.970695972442627,
|
| 2179 |
+
"learning_rate": 3.8971807628524045e-05,
|
| 2180 |
+
"loss": 0.5297,
|
| 2181 |
+
"step": 600
|
| 2182 |
+
},
|
| 2183 |
+
{
|
| 2184 |
+
"epoch": 4.492537313432836,
|
| 2185 |
+
"grad_norm": 3.2911581993103027,
|
| 2186 |
+
"learning_rate": 3.891652846876728e-05,
|
| 2187 |
+
"loss": 0.4756,
|
| 2188 |
+
"step": 602
|
| 2189 |
+
},
|
| 2190 |
+
{
|
| 2191 |
+
"epoch": 4.507462686567164,
|
| 2192 |
+
"grad_norm": 6.148906230926514,
|
| 2193 |
+
"learning_rate": 3.886124930901051e-05,
|
| 2194 |
+
"loss": 0.5251,
|
| 2195 |
+
"step": 604
|
| 2196 |
+
},
|
| 2197 |
+
{
|
| 2198 |
+
"epoch": 4.522388059701493,
|
| 2199 |
+
"grad_norm": 4.234598636627197,
|
| 2200 |
+
"learning_rate": 3.8805970149253736e-05,
|
| 2201 |
+
"loss": 0.3342,
|
| 2202 |
+
"step": 606
|
| 2203 |
+
},
|
| 2204 |
+
{
|
| 2205 |
+
"epoch": 4.537313432835821,
|
| 2206 |
+
"grad_norm": 4.786794662475586,
|
| 2207 |
+
"learning_rate": 3.8750690989496964e-05,
|
| 2208 |
+
"loss": 0.6528,
|
| 2209 |
+
"step": 608
|
| 2210 |
+
},
|
| 2211 |
+
{
|
| 2212 |
+
"epoch": 4.552238805970149,
|
| 2213 |
+
"grad_norm": 8.322453498840332,
|
| 2214 |
+
"learning_rate": 3.869541182974019e-05,
|
| 2215 |
+
"loss": 0.5658,
|
| 2216 |
+
"step": 610
|
| 2217 |
+
},
|
| 2218 |
+
{
|
| 2219 |
+
"epoch": 4.567164179104478,
|
| 2220 |
+
"grad_norm": 6.955239772796631,
|
| 2221 |
+
"learning_rate": 3.864013266998342e-05,
|
| 2222 |
+
"loss": 0.5496,
|
| 2223 |
+
"step": 612
|
| 2224 |
+
},
|
| 2225 |
+
{
|
| 2226 |
+
"epoch": 4.582089552238806,
|
| 2227 |
+
"grad_norm": 4.89069938659668,
|
| 2228 |
+
"learning_rate": 3.858485351022665e-05,
|
| 2229 |
+
"loss": 0.4191,
|
| 2230 |
+
"step": 614
|
| 2231 |
+
},
|
| 2232 |
+
{
|
| 2233 |
+
"epoch": 4.597014925373134,
|
| 2234 |
+
"grad_norm": 9.381403923034668,
|
| 2235 |
+
"learning_rate": 3.852957435046988e-05,
|
| 2236 |
+
"loss": 0.5993,
|
| 2237 |
+
"step": 616
|
| 2238 |
+
},
|
| 2239 |
+
{
|
| 2240 |
+
"epoch": 4.611940298507463,
|
| 2241 |
+
"grad_norm": 4.977263450622559,
|
| 2242 |
+
"learning_rate": 3.8474295190713105e-05,
|
| 2243 |
+
"loss": 0.569,
|
| 2244 |
+
"step": 618
|
| 2245 |
+
},
|
| 2246 |
+
{
|
| 2247 |
+
"epoch": 4.6268656716417915,
|
| 2248 |
+
"grad_norm": 3.183396816253662,
|
| 2249 |
+
"learning_rate": 3.841901603095633e-05,
|
| 2250 |
+
"loss": 0.5183,
|
| 2251 |
+
"step": 620
|
| 2252 |
+
},
|
| 2253 |
+
{
|
| 2254 |
+
"epoch": 4.641791044776119,
|
| 2255 |
+
"grad_norm": 3.1676368713378906,
|
| 2256 |
+
"learning_rate": 3.8363736871199555e-05,
|
| 2257 |
+
"loss": 0.4647,
|
| 2258 |
+
"step": 622
|
| 2259 |
+
},
|
| 2260 |
+
{
|
| 2261 |
+
"epoch": 4.656716417910448,
|
| 2262 |
+
"grad_norm": 9.969623565673828,
|
| 2263 |
+
"learning_rate": 3.830845771144278e-05,
|
| 2264 |
+
"loss": 0.4468,
|
| 2265 |
+
"step": 624
|
| 2266 |
+
},
|
| 2267 |
+
{
|
| 2268 |
+
"epoch": 4.6716417910447765,
|
| 2269 |
+
"grad_norm": 11.773579597473145,
|
| 2270 |
+
"learning_rate": 3.825317855168601e-05,
|
| 2271 |
+
"loss": 0.477,
|
| 2272 |
+
"step": 626
|
| 2273 |
+
},
|
| 2274 |
+
{
|
| 2275 |
+
"epoch": 4.686567164179104,
|
| 2276 |
+
"grad_norm": 3.361043691635132,
|
| 2277 |
+
"learning_rate": 3.8197899391929246e-05,
|
| 2278 |
+
"loss": 0.4533,
|
| 2279 |
+
"step": 628
|
| 2280 |
+
},
|
| 2281 |
+
{
|
| 2282 |
+
"epoch": 4.701492537313433,
|
| 2283 |
+
"grad_norm": 6.481675148010254,
|
| 2284 |
+
"learning_rate": 3.8142620232172474e-05,
|
| 2285 |
+
"loss": 0.5998,
|
| 2286 |
+
"step": 630
|
| 2287 |
+
},
|
| 2288 |
+
{
|
| 2289 |
+
"epoch": 4.7164179104477615,
|
| 2290 |
+
"grad_norm": 10.961061477661133,
|
| 2291 |
+
"learning_rate": 3.80873410724157e-05,
|
| 2292 |
+
"loss": 0.4246,
|
| 2293 |
+
"step": 632
|
| 2294 |
+
},
|
| 2295 |
+
{
|
| 2296 |
+
"epoch": 4.731343283582089,
|
| 2297 |
+
"grad_norm": 11.879002571105957,
|
| 2298 |
+
"learning_rate": 3.803206191265893e-05,
|
| 2299 |
+
"loss": 0.4113,
|
| 2300 |
+
"step": 634
|
| 2301 |
+
},
|
| 2302 |
+
{
|
| 2303 |
+
"epoch": 4.746268656716418,
|
| 2304 |
+
"grad_norm": 6.598226547241211,
|
| 2305 |
+
"learning_rate": 3.797678275290216e-05,
|
| 2306 |
+
"loss": 0.4986,
|
| 2307 |
+
"step": 636
|
| 2308 |
+
},
|
| 2309 |
+
{
|
| 2310 |
+
"epoch": 4.7611940298507465,
|
| 2311 |
+
"grad_norm": 9.366608619689941,
|
| 2312 |
+
"learning_rate": 3.7921503593145386e-05,
|
| 2313 |
+
"loss": 0.3439,
|
| 2314 |
+
"step": 638
|
| 2315 |
+
},
|
| 2316 |
+
{
|
| 2317 |
+
"epoch": 4.776119402985074,
|
| 2318 |
+
"grad_norm": 11.350964546203613,
|
| 2319 |
+
"learning_rate": 3.7866224433388615e-05,
|
| 2320 |
+
"loss": 0.8422,
|
| 2321 |
+
"step": 640
|
| 2322 |
+
},
|
| 2323 |
+
{
|
| 2324 |
+
"epoch": 4.791044776119403,
|
| 2325 |
+
"grad_norm": 8.099052429199219,
|
| 2326 |
+
"learning_rate": 3.781094527363184e-05,
|
| 2327 |
+
"loss": 0.687,
|
| 2328 |
+
"step": 642
|
| 2329 |
+
},
|
| 2330 |
+
{
|
| 2331 |
+
"epoch": 4.8059701492537314,
|
| 2332 |
+
"grad_norm": 5.031801223754883,
|
| 2333 |
+
"learning_rate": 3.775566611387507e-05,
|
| 2334 |
+
"loss": 0.484,
|
| 2335 |
+
"step": 644
|
| 2336 |
+
},
|
| 2337 |
+
{
|
| 2338 |
+
"epoch": 4.82089552238806,
|
| 2339 |
+
"grad_norm": 3.5851833820343018,
|
| 2340 |
+
"learning_rate": 3.77003869541183e-05,
|
| 2341 |
+
"loss": 0.6037,
|
| 2342 |
+
"step": 646
|
| 2343 |
+
},
|
| 2344 |
+
{
|
| 2345 |
+
"epoch": 4.835820895522388,
|
| 2346 |
+
"grad_norm": 14.049162864685059,
|
| 2347 |
+
"learning_rate": 3.764510779436153e-05,
|
| 2348 |
+
"loss": 0.5521,
|
| 2349 |
+
"step": 648
|
| 2350 |
+
},
|
| 2351 |
+
{
|
| 2352 |
+
"epoch": 4.850746268656716,
|
| 2353 |
+
"grad_norm": 5.086363315582275,
|
| 2354 |
+
"learning_rate": 3.7589828634604755e-05,
|
| 2355 |
+
"loss": 0.5173,
|
| 2356 |
+
"step": 650
|
| 2357 |
+
},
|
| 2358 |
+
{
|
| 2359 |
+
"epoch": 4.865671641791045,
|
| 2360 |
+
"grad_norm": 2.833794355392456,
|
| 2361 |
+
"learning_rate": 3.753454947484798e-05,
|
| 2362 |
+
"loss": 0.7762,
|
| 2363 |
+
"step": 652
|
| 2364 |
+
},
|
| 2365 |
+
{
|
| 2366 |
+
"epoch": 4.880597014925373,
|
| 2367 |
+
"grad_norm": 2.7670750617980957,
|
| 2368 |
+
"learning_rate": 3.747927031509121e-05,
|
| 2369 |
+
"loss": 0.5592,
|
| 2370 |
+
"step": 654
|
| 2371 |
+
},
|
| 2372 |
+
{
|
| 2373 |
+
"epoch": 4.895522388059701,
|
| 2374 |
+
"grad_norm": 2.6060335636138916,
|
| 2375 |
+
"learning_rate": 3.742399115533444e-05,
|
| 2376 |
+
"loss": 0.2484,
|
| 2377 |
+
"step": 656
|
| 2378 |
+
},
|
| 2379 |
+
{
|
| 2380 |
+
"epoch": 4.91044776119403,
|
| 2381 |
+
"grad_norm": 5.139571666717529,
|
| 2382 |
+
"learning_rate": 3.736871199557767e-05,
|
| 2383 |
+
"loss": 0.5157,
|
| 2384 |
+
"step": 658
|
| 2385 |
+
},
|
| 2386 |
+
{
|
| 2387 |
+
"epoch": 4.925373134328359,
|
| 2388 |
+
"grad_norm": 3.047414779663086,
|
| 2389 |
+
"learning_rate": 3.73134328358209e-05,
|
| 2390 |
+
"loss": 0.4967,
|
| 2391 |
+
"step": 660
|
| 2392 |
+
},
|
| 2393 |
+
{
|
| 2394 |
+
"epoch": 4.940298507462686,
|
| 2395 |
+
"grad_norm": 7.455503940582275,
|
| 2396 |
+
"learning_rate": 3.725815367606413e-05,
|
| 2397 |
+
"loss": 0.4034,
|
| 2398 |
+
"step": 662
|
| 2399 |
+
},
|
| 2400 |
+
{
|
| 2401 |
+
"epoch": 4.955223880597015,
|
| 2402 |
+
"grad_norm": 12.100015640258789,
|
| 2403 |
+
"learning_rate": 3.720287451630736e-05,
|
| 2404 |
+
"loss": 0.5245,
|
| 2405 |
+
"step": 664
|
| 2406 |
+
},
|
| 2407 |
+
{
|
| 2408 |
+
"epoch": 4.970149253731344,
|
| 2409 |
+
"grad_norm": 3.1610209941864014,
|
| 2410 |
+
"learning_rate": 3.714759535655058e-05,
|
| 2411 |
+
"loss": 0.6381,
|
| 2412 |
+
"step": 666
|
| 2413 |
+
},
|
| 2414 |
+
{
|
| 2415 |
+
"epoch": 4.985074626865671,
|
| 2416 |
+
"grad_norm": 3.781731367111206,
|
| 2417 |
+
"learning_rate": 3.709231619679381e-05,
|
| 2418 |
+
"loss": 0.5727,
|
| 2419 |
+
"step": 668
|
| 2420 |
+
},
|
| 2421 |
+
{
|
| 2422 |
+
"epoch": 5.0,
|
| 2423 |
+
"grad_norm": 1.7648042440414429,
|
| 2424 |
+
"learning_rate": 3.7037037037037037e-05,
|
| 2425 |
+
"loss": 0.3107,
|
| 2426 |
+
"step": 670
|
| 2427 |
+
},
|
| 2428 |
+
{
|
| 2429 |
+
"epoch": 5.0,
|
| 2430 |
+
"eval_accuracy": 0.725,
|
| 2431 |
+
"eval_f1_macro": 0.67604038118744,
|
| 2432 |
+
"eval_f1_micro": 0.725,
|
| 2433 |
+
"eval_f1_weighted": 0.6927145770895771,
|
| 2434 |
+
"eval_loss": 1.0104252099990845,
|
| 2435 |
+
"eval_precision_macro": 0.7111726508785332,
|
| 2436 |
+
"eval_precision_micro": 0.725,
|
| 2437 |
+
"eval_precision_weighted": 0.7247546897546897,
|
| 2438 |
+
"eval_recall_macro": 0.7092436974789916,
|
| 2439 |
+
"eval_recall_micro": 0.725,
|
| 2440 |
+
"eval_recall_weighted": 0.725,
|
| 2441 |
+
"eval_runtime": 39.3596,
|
| 2442 |
+
"eval_samples_per_second": 5.081,
|
| 2443 |
+
"eval_steps_per_second": 0.178,
|
| 2444 |
+
"step": 670
|
| 2445 |
+
},
|
| 2446 |
+
{
|
| 2447 |
+
"epoch": 5.014925373134329,
|
| 2448 |
+
"grad_norm": 1.6856744289398193,
|
| 2449 |
+
"learning_rate": 3.6981757877280265e-05,
|
| 2450 |
+
"loss": 0.289,
|
| 2451 |
+
"step": 672
|
| 2452 |
+
},
|
| 2453 |
+
{
|
| 2454 |
+
"epoch": 5.029850746268656,
|
| 2455 |
+
"grad_norm": 2.8615660667419434,
|
| 2456 |
+
"learning_rate": 3.692647871752349e-05,
|
| 2457 |
+
"loss": 0.4746,
|
| 2458 |
+
"step": 674
|
| 2459 |
+
},
|
| 2460 |
+
{
|
| 2461 |
+
"epoch": 5.044776119402985,
|
| 2462 |
+
"grad_norm": 6.883438587188721,
|
| 2463 |
+
"learning_rate": 3.687119955776672e-05,
|
| 2464 |
+
"loss": 0.6167,
|
| 2465 |
+
"step": 676
|
| 2466 |
+
},
|
| 2467 |
+
{
|
| 2468 |
+
"epoch": 5.059701492537314,
|
| 2469 |
+
"grad_norm": 7.769782543182373,
|
| 2470 |
+
"learning_rate": 3.681592039800995e-05,
|
| 2471 |
+
"loss": 0.3471,
|
| 2472 |
+
"step": 678
|
| 2473 |
+
},
|
| 2474 |
+
{
|
| 2475 |
+
"epoch": 5.074626865671641,
|
| 2476 |
+
"grad_norm": 3.5327107906341553,
|
| 2477 |
+
"learning_rate": 3.676064123825318e-05,
|
| 2478 |
+
"loss": 0.4453,
|
| 2479 |
+
"step": 680
|
| 2480 |
+
},
|
| 2481 |
+
{
|
| 2482 |
+
"epoch": 5.08955223880597,
|
| 2483 |
+
"grad_norm": 3.9162540435791016,
|
| 2484 |
+
"learning_rate": 3.6705362078496405e-05,
|
| 2485 |
+
"loss": 0.4471,
|
| 2486 |
+
"step": 682
|
| 2487 |
+
},
|
| 2488 |
+
{
|
| 2489 |
+
"epoch": 5.104477611940299,
|
| 2490 |
+
"grad_norm": 2.9827005863189697,
|
| 2491 |
+
"learning_rate": 3.6650082918739633e-05,
|
| 2492 |
+
"loss": 0.344,
|
| 2493 |
+
"step": 684
|
| 2494 |
+
},
|
| 2495 |
+
{
|
| 2496 |
+
"epoch": 5.119402985074627,
|
| 2497 |
+
"grad_norm": 5.053178310394287,
|
| 2498 |
+
"learning_rate": 3.659480375898287e-05,
|
| 2499 |
+
"loss": 0.3973,
|
| 2500 |
+
"step": 686
|
| 2501 |
+
},
|
| 2502 |
+
{
|
| 2503 |
+
"epoch": 5.134328358208955,
|
| 2504 |
+
"grad_norm": 5.1265716552734375,
|
| 2505 |
+
"learning_rate": 3.6539524599226097e-05,
|
| 2506 |
+
"loss": 0.3783,
|
| 2507 |
+
"step": 688
|
| 2508 |
+
},
|
| 2509 |
+
{
|
| 2510 |
+
"epoch": 5.149253731343284,
|
| 2511 |
+
"grad_norm": 3.368522882461548,
|
| 2512 |
+
"learning_rate": 3.6484245439469325e-05,
|
| 2513 |
+
"loss": 0.5253,
|
| 2514 |
+
"step": 690
|
| 2515 |
+
},
|
| 2516 |
+
{
|
| 2517 |
+
"epoch": 5.164179104477612,
|
| 2518 |
+
"grad_norm": 5.44252872467041,
|
| 2519 |
+
"learning_rate": 3.642896627971255e-05,
|
| 2520 |
+
"loss": 0.54,
|
| 2521 |
+
"step": 692
|
| 2522 |
+
},
|
| 2523 |
+
{
|
| 2524 |
+
"epoch": 5.17910447761194,
|
| 2525 |
+
"grad_norm": 7.313326358795166,
|
| 2526 |
+
"learning_rate": 3.637368711995578e-05,
|
| 2527 |
+
"loss": 0.3734,
|
| 2528 |
+
"step": 694
|
| 2529 |
+
},
|
| 2530 |
+
{
|
| 2531 |
+
"epoch": 5.1940298507462686,
|
| 2532 |
+
"grad_norm": 3.947737693786621,
|
| 2533 |
+
"learning_rate": 3.631840796019901e-05,
|
| 2534 |
+
"loss": 0.4204,
|
| 2535 |
+
"step": 696
|
| 2536 |
+
},
|
| 2537 |
+
{
|
| 2538 |
+
"epoch": 5.208955223880597,
|
| 2539 |
+
"grad_norm": 2.3334567546844482,
|
| 2540 |
+
"learning_rate": 3.626312880044224e-05,
|
| 2541 |
+
"loss": 0.3622,
|
| 2542 |
+
"step": 698
|
| 2543 |
+
},
|
| 2544 |
+
{
|
| 2545 |
+
"epoch": 5.223880597014926,
|
| 2546 |
+
"grad_norm": 4.052284240722656,
|
| 2547 |
+
"learning_rate": 3.6207849640685465e-05,
|
| 2548 |
+
"loss": 0.2778,
|
| 2549 |
+
"step": 700
|
| 2550 |
+
},
|
| 2551 |
+
{
|
| 2552 |
+
"epoch": 5.2388059701492535,
|
| 2553 |
+
"grad_norm": 11.430601119995117,
|
| 2554 |
+
"learning_rate": 3.6152570480928693e-05,
|
| 2555 |
+
"loss": 0.6734,
|
| 2556 |
+
"step": 702
|
| 2557 |
+
},
|
| 2558 |
+
{
|
| 2559 |
+
"epoch": 5.253731343283582,
|
| 2560 |
+
"grad_norm": 5.606107234954834,
|
| 2561 |
+
"learning_rate": 3.609729132117192e-05,
|
| 2562 |
+
"loss": 0.4028,
|
| 2563 |
+
"step": 704
|
| 2564 |
+
},
|
| 2565 |
+
{
|
| 2566 |
+
"epoch": 5.268656716417911,
|
| 2567 |
+
"grad_norm": 13.239790916442871,
|
| 2568 |
+
"learning_rate": 3.604201216141515e-05,
|
| 2569 |
+
"loss": 0.6932,
|
| 2570 |
+
"step": 706
|
| 2571 |
+
},
|
| 2572 |
+
{
|
| 2573 |
+
"epoch": 5.2835820895522385,
|
| 2574 |
+
"grad_norm": 2.727085590362549,
|
| 2575 |
+
"learning_rate": 3.598673300165838e-05,
|
| 2576 |
+
"loss": 0.6325,
|
| 2577 |
+
"step": 708
|
| 2578 |
+
},
|
| 2579 |
+
{
|
| 2580 |
+
"epoch": 5.298507462686567,
|
| 2581 |
+
"grad_norm": 4.920370101928711,
|
| 2582 |
+
"learning_rate": 3.59314538419016e-05,
|
| 2583 |
+
"loss": 0.215,
|
| 2584 |
+
"step": 710
|
| 2585 |
+
},
|
| 2586 |
+
{
|
| 2587 |
+
"epoch": 5.313432835820896,
|
| 2588 |
+
"grad_norm": 3.629782199859619,
|
| 2589 |
+
"learning_rate": 3.5876174682144834e-05,
|
| 2590 |
+
"loss": 0.5572,
|
| 2591 |
+
"step": 712
|
| 2592 |
+
},
|
| 2593 |
+
{
|
| 2594 |
+
"epoch": 5.3283582089552235,
|
| 2595 |
+
"grad_norm": 5.6093668937683105,
|
| 2596 |
+
"learning_rate": 3.582089552238806e-05,
|
| 2597 |
+
"loss": 0.7121,
|
| 2598 |
+
"step": 714
|
| 2599 |
+
},
|
| 2600 |
+
{
|
| 2601 |
+
"epoch": 5.343283582089552,
|
| 2602 |
+
"grad_norm": 6.661886215209961,
|
| 2603 |
+
"learning_rate": 3.576561636263129e-05,
|
| 2604 |
+
"loss": 0.4479,
|
| 2605 |
+
"step": 716
|
| 2606 |
+
},
|
| 2607 |
+
{
|
| 2608 |
+
"epoch": 5.358208955223881,
|
| 2609 |
+
"grad_norm": 10.350639343261719,
|
| 2610 |
+
"learning_rate": 3.571033720287452e-05,
|
| 2611 |
+
"loss": 0.5578,
|
| 2612 |
+
"step": 718
|
| 2613 |
+
},
|
| 2614 |
+
{
|
| 2615 |
+
"epoch": 5.373134328358209,
|
| 2616 |
+
"grad_norm": 6.086325168609619,
|
| 2617 |
+
"learning_rate": 3.565505804311775e-05,
|
| 2618 |
+
"loss": 0.4269,
|
| 2619 |
+
"step": 720
|
| 2620 |
+
},
|
| 2621 |
+
{
|
| 2622 |
+
"epoch": 5.388059701492537,
|
| 2623 |
+
"grad_norm": 6.263159275054932,
|
| 2624 |
+
"learning_rate": 3.5599778883360975e-05,
|
| 2625 |
+
"loss": 0.5466,
|
| 2626 |
+
"step": 722
|
| 2627 |
+
},
|
| 2628 |
+
{
|
| 2629 |
+
"epoch": 5.402985074626866,
|
| 2630 |
+
"grad_norm": 6.2910990715026855,
|
| 2631 |
+
"learning_rate": 3.55444997236042e-05,
|
| 2632 |
+
"loss": 0.3659,
|
| 2633 |
+
"step": 724
|
| 2634 |
+
},
|
| 2635 |
+
{
|
| 2636 |
+
"epoch": 5.417910447761194,
|
| 2637 |
+
"grad_norm": 4.90096378326416,
|
| 2638 |
+
"learning_rate": 3.548922056384743e-05,
|
| 2639 |
+
"loss": 0.6204,
|
| 2640 |
+
"step": 726
|
| 2641 |
+
},
|
| 2642 |
+
{
|
| 2643 |
+
"epoch": 5.432835820895522,
|
| 2644 |
+
"grad_norm": 3.6040308475494385,
|
| 2645 |
+
"learning_rate": 3.543394140409066e-05,
|
| 2646 |
+
"loss": 0.5589,
|
| 2647 |
+
"step": 728
|
| 2648 |
+
},
|
| 2649 |
+
{
|
| 2650 |
+
"epoch": 5.447761194029851,
|
| 2651 |
+
"grad_norm": 14.020445823669434,
|
| 2652 |
+
"learning_rate": 3.537866224433389e-05,
|
| 2653 |
+
"loss": 0.5483,
|
| 2654 |
+
"step": 730
|
| 2655 |
+
},
|
| 2656 |
+
{
|
| 2657 |
+
"epoch": 5.462686567164179,
|
| 2658 |
+
"grad_norm": 2.3177435398101807,
|
| 2659 |
+
"learning_rate": 3.5323383084577115e-05,
|
| 2660 |
+
"loss": 0.3787,
|
| 2661 |
+
"step": 732
|
| 2662 |
+
},
|
| 2663 |
+
{
|
| 2664 |
+
"epoch": 5.477611940298507,
|
| 2665 |
+
"grad_norm": 2.891357183456421,
|
| 2666 |
+
"learning_rate": 3.5268103924820344e-05,
|
| 2667 |
+
"loss": 0.347,
|
| 2668 |
+
"step": 734
|
| 2669 |
+
},
|
| 2670 |
+
{
|
| 2671 |
+
"epoch": 5.492537313432836,
|
| 2672 |
+
"grad_norm": 11.490731239318848,
|
| 2673 |
+
"learning_rate": 3.521282476506357e-05,
|
| 2674 |
+
"loss": 0.5475,
|
| 2675 |
+
"step": 736
|
| 2676 |
+
},
|
| 2677 |
+
{
|
| 2678 |
+
"epoch": 5.507462686567164,
|
| 2679 |
+
"grad_norm": 2.3620495796203613,
|
| 2680 |
+
"learning_rate": 3.51575456053068e-05,
|
| 2681 |
+
"loss": 0.4033,
|
| 2682 |
+
"step": 738
|
| 2683 |
+
},
|
| 2684 |
+
{
|
| 2685 |
+
"epoch": 5.522388059701493,
|
| 2686 |
+
"grad_norm": 3.9230000972747803,
|
| 2687 |
+
"learning_rate": 3.510226644555003e-05,
|
| 2688 |
+
"loss": 0.4246,
|
| 2689 |
+
"step": 740
|
| 2690 |
+
},
|
| 2691 |
+
{
|
| 2692 |
+
"epoch": 5.537313432835821,
|
| 2693 |
+
"grad_norm": 3.205672025680542,
|
| 2694 |
+
"learning_rate": 3.5046987285793256e-05,
|
| 2695 |
+
"loss": 0.4462,
|
| 2696 |
+
"step": 742
|
| 2697 |
+
},
|
| 2698 |
+
{
|
| 2699 |
+
"epoch": 5.552238805970149,
|
| 2700 |
+
"grad_norm": 5.685773849487305,
|
| 2701 |
+
"learning_rate": 3.499170812603649e-05,
|
| 2702 |
+
"loss": 0.379,
|
| 2703 |
+
"step": 744
|
| 2704 |
+
},
|
| 2705 |
+
{
|
| 2706 |
+
"epoch": 5.567164179104478,
|
| 2707 |
+
"grad_norm": 5.982370376586914,
|
| 2708 |
+
"learning_rate": 3.493642896627972e-05,
|
| 2709 |
+
"loss": 0.4962,
|
| 2710 |
+
"step": 746
|
| 2711 |
+
},
|
| 2712 |
+
{
|
| 2713 |
+
"epoch": 5.582089552238806,
|
| 2714 |
+
"grad_norm": 13.228385925292969,
|
| 2715 |
+
"learning_rate": 3.488114980652295e-05,
|
| 2716 |
+
"loss": 0.3345,
|
| 2717 |
+
"step": 748
|
| 2718 |
+
},
|
| 2719 |
+
{
|
| 2720 |
+
"epoch": 5.597014925373134,
|
| 2721 |
+
"grad_norm": 5.765232086181641,
|
| 2722 |
+
"learning_rate": 3.4825870646766175e-05,
|
| 2723 |
+
"loss": 0.4973,
|
| 2724 |
+
"step": 750
|
| 2725 |
+
},
|
| 2726 |
+
{
|
| 2727 |
+
"epoch": 5.611940298507463,
|
| 2728 |
+
"grad_norm": 1.701096534729004,
|
| 2729 |
+
"learning_rate": 3.4770591487009404e-05,
|
| 2730 |
+
"loss": 0.3608,
|
| 2731 |
+
"step": 752
|
| 2732 |
+
},
|
| 2733 |
+
{
|
| 2734 |
+
"epoch": 5.6268656716417915,
|
| 2735 |
+
"grad_norm": 14.079666137695312,
|
| 2736 |
+
"learning_rate": 3.4715312327252625e-05,
|
| 2737 |
+
"loss": 0.5854,
|
| 2738 |
+
"step": 754
|
| 2739 |
+
},
|
| 2740 |
+
{
|
| 2741 |
+
"epoch": 5.641791044776119,
|
| 2742 |
+
"grad_norm": 3.6387109756469727,
|
| 2743 |
+
"learning_rate": 3.466003316749585e-05,
|
| 2744 |
+
"loss": 0.2784,
|
| 2745 |
+
"step": 756
|
| 2746 |
+
},
|
| 2747 |
+
{
|
| 2748 |
+
"epoch": 5.656716417910448,
|
| 2749 |
+
"grad_norm": 1.6772111654281616,
|
| 2750 |
+
"learning_rate": 3.460475400773908e-05,
|
| 2751 |
+
"loss": 0.2205,
|
| 2752 |
+
"step": 758
|
| 2753 |
+
},
|
| 2754 |
+
{
|
| 2755 |
+
"epoch": 5.6716417910447765,
|
| 2756 |
+
"grad_norm": 5.0109405517578125,
|
| 2757 |
+
"learning_rate": 3.454947484798231e-05,
|
| 2758 |
+
"loss": 0.4189,
|
| 2759 |
+
"step": 760
|
| 2760 |
+
},
|
| 2761 |
+
{
|
| 2762 |
+
"epoch": 5.686567164179104,
|
| 2763 |
+
"grad_norm": 4.438418865203857,
|
| 2764 |
+
"learning_rate": 3.449419568822554e-05,
|
| 2765 |
+
"loss": 0.1976,
|
| 2766 |
+
"step": 762
|
| 2767 |
+
},
|
| 2768 |
+
{
|
| 2769 |
+
"epoch": 5.701492537313433,
|
| 2770 |
+
"grad_norm": 4.297623157501221,
|
| 2771 |
+
"learning_rate": 3.4438916528468766e-05,
|
| 2772 |
+
"loss": 0.3569,
|
| 2773 |
+
"step": 764
|
| 2774 |
+
},
|
| 2775 |
+
{
|
| 2776 |
+
"epoch": 5.7164179104477615,
|
| 2777 |
+
"grad_norm": 7.163605213165283,
|
| 2778 |
+
"learning_rate": 3.4383637368711994e-05,
|
| 2779 |
+
"loss": 0.2972,
|
| 2780 |
+
"step": 766
|
| 2781 |
+
},
|
| 2782 |
+
{
|
| 2783 |
+
"epoch": 5.731343283582089,
|
| 2784 |
+
"grad_norm": 3.6532185077667236,
|
| 2785 |
+
"learning_rate": 3.432835820895522e-05,
|
| 2786 |
+
"loss": 0.1953,
|
| 2787 |
+
"step": 768
|
| 2788 |
+
},
|
| 2789 |
+
{
|
| 2790 |
+
"epoch": 5.746268656716418,
|
| 2791 |
+
"grad_norm": 10.143959045410156,
|
| 2792 |
+
"learning_rate": 3.427307904919845e-05,
|
| 2793 |
+
"loss": 0.4292,
|
| 2794 |
+
"step": 770
|
| 2795 |
+
},
|
| 2796 |
+
{
|
| 2797 |
+
"epoch": 5.7611940298507465,
|
| 2798 |
+
"grad_norm": 7.471317768096924,
|
| 2799 |
+
"learning_rate": 3.4217799889441685e-05,
|
| 2800 |
+
"loss": 0.472,
|
| 2801 |
+
"step": 772
|
| 2802 |
+
},
|
| 2803 |
+
{
|
| 2804 |
+
"epoch": 5.776119402985074,
|
| 2805 |
+
"grad_norm": 3.14628529548645,
|
| 2806 |
+
"learning_rate": 3.416252072968491e-05,
|
| 2807 |
+
"loss": 0.4863,
|
| 2808 |
+
"step": 774
|
| 2809 |
+
},
|
| 2810 |
+
{
|
| 2811 |
+
"epoch": 5.791044776119403,
|
| 2812 |
+
"grad_norm": 2.2698168754577637,
|
| 2813 |
+
"learning_rate": 3.410724156992814e-05,
|
| 2814 |
+
"loss": 0.2093,
|
| 2815 |
+
"step": 776
|
| 2816 |
+
},
|
| 2817 |
+
{
|
| 2818 |
+
"epoch": 5.8059701492537314,
|
| 2819 |
+
"grad_norm": 3.6853973865509033,
|
| 2820 |
+
"learning_rate": 3.405196241017137e-05,
|
| 2821 |
+
"loss": 0.5749,
|
| 2822 |
+
"step": 778
|
| 2823 |
+
},
|
| 2824 |
+
{
|
| 2825 |
+
"epoch": 5.82089552238806,
|
| 2826 |
+
"grad_norm": 9.027390480041504,
|
| 2827 |
+
"learning_rate": 3.39966832504146e-05,
|
| 2828 |
+
"loss": 0.8224,
|
| 2829 |
+
"step": 780
|
| 2830 |
+
},
|
| 2831 |
+
{
|
| 2832 |
+
"epoch": 5.835820895522388,
|
| 2833 |
+
"grad_norm": 4.245992660522461,
|
| 2834 |
+
"learning_rate": 3.3941404090657825e-05,
|
| 2835 |
+
"loss": 0.2923,
|
| 2836 |
+
"step": 782
|
| 2837 |
+
},
|
| 2838 |
+
{
|
| 2839 |
+
"epoch": 5.850746268656716,
|
| 2840 |
+
"grad_norm": 10.722375869750977,
|
| 2841 |
+
"learning_rate": 3.3886124930901054e-05,
|
| 2842 |
+
"loss": 0.7797,
|
| 2843 |
+
"step": 784
|
| 2844 |
+
},
|
| 2845 |
+
{
|
| 2846 |
+
"epoch": 5.865671641791045,
|
| 2847 |
+
"grad_norm": 4.297248363494873,
|
| 2848 |
+
"learning_rate": 3.383084577114428e-05,
|
| 2849 |
+
"loss": 0.5619,
|
| 2850 |
+
"step": 786
|
| 2851 |
+
},
|
| 2852 |
+
{
|
| 2853 |
+
"epoch": 5.880597014925373,
|
| 2854 |
+
"grad_norm": 4.837419509887695,
|
| 2855 |
+
"learning_rate": 3.377556661138751e-05,
|
| 2856 |
+
"loss": 0.3356,
|
| 2857 |
+
"step": 788
|
| 2858 |
+
},
|
| 2859 |
+
{
|
| 2860 |
+
"epoch": 5.895522388059701,
|
| 2861 |
+
"grad_norm": 1.2375975847244263,
|
| 2862 |
+
"learning_rate": 3.372028745163074e-05,
|
| 2863 |
+
"loss": 0.1544,
|
| 2864 |
+
"step": 790
|
| 2865 |
+
},
|
| 2866 |
+
{
|
| 2867 |
+
"epoch": 5.91044776119403,
|
| 2868 |
+
"grad_norm": 12.756596565246582,
|
| 2869 |
+
"learning_rate": 3.3665008291873966e-05,
|
| 2870 |
+
"loss": 0.3809,
|
| 2871 |
+
"step": 792
|
| 2872 |
+
},
|
| 2873 |
+
{
|
| 2874 |
+
"epoch": 5.925373134328359,
|
| 2875 |
+
"grad_norm": 3.4860129356384277,
|
| 2876 |
+
"learning_rate": 3.3609729132117194e-05,
|
| 2877 |
+
"loss": 0.3144,
|
| 2878 |
+
"step": 794
|
| 2879 |
+
},
|
| 2880 |
+
{
|
| 2881 |
+
"epoch": 5.940298507462686,
|
| 2882 |
+
"grad_norm": 7.757724761962891,
|
| 2883 |
+
"learning_rate": 3.355444997236042e-05,
|
| 2884 |
+
"loss": 0.3695,
|
| 2885 |
+
"step": 796
|
| 2886 |
+
},
|
| 2887 |
+
{
|
| 2888 |
+
"epoch": 5.955223880597015,
|
| 2889 |
+
"grad_norm": 2.4491782188415527,
|
| 2890 |
+
"learning_rate": 3.349917081260365e-05,
|
| 2891 |
+
"loss": 0.275,
|
| 2892 |
+
"step": 798
|
| 2893 |
+
},
|
| 2894 |
+
{
|
| 2895 |
+
"epoch": 5.970149253731344,
|
| 2896 |
+
"grad_norm": 3.097658395767212,
|
| 2897 |
+
"learning_rate": 3.344389165284688e-05,
|
| 2898 |
+
"loss": 0.5223,
|
| 2899 |
+
"step": 800
|
| 2900 |
+
},
|
| 2901 |
+
{
|
| 2902 |
+
"epoch": 5.985074626865671,
|
| 2903 |
+
"grad_norm": 3.9302120208740234,
|
| 2904 |
+
"learning_rate": 3.338861249309011e-05,
|
| 2905 |
+
"loss": 0.2202,
|
| 2906 |
+
"step": 802
|
| 2907 |
+
},
|
| 2908 |
+
{
|
| 2909 |
+
"epoch": 6.0,
|
| 2910 |
+
"grad_norm": 4.732783794403076,
|
| 2911 |
+
"learning_rate": 3.3333333333333335e-05,
|
| 2912 |
+
"loss": 0.4422,
|
| 2913 |
+
"step": 804
|
| 2914 |
+
},
|
| 2915 |
+
{
|
| 2916 |
+
"epoch": 6.0,
|
| 2917 |
+
"eval_accuracy": 0.76,
|
| 2918 |
+
"eval_f1_macro": 0.7178212262554823,
|
| 2919 |
+
"eval_f1_micro": 0.76,
|
| 2920 |
+
"eval_f1_weighted": 0.7341365415630122,
|
| 2921 |
+
"eval_loss": 0.890067458152771,
|
| 2922 |
+
"eval_precision_macro": 0.7468954248366012,
|
| 2923 |
+
"eval_precision_micro": 0.76,
|
| 2924 |
+
"eval_precision_weighted": 0.766861111111111,
|
| 2925 |
+
"eval_recall_macro": 0.7448179271708684,
|
| 2926 |
+
"eval_recall_micro": 0.76,
|
| 2927 |
+
"eval_recall_weighted": 0.76,
|
| 2928 |
+
"eval_runtime": 37.7852,
|
| 2929 |
+
"eval_samples_per_second": 5.293,
|
| 2930 |
+
"eval_steps_per_second": 0.185,
|
| 2931 |
+
"step": 804
|
| 2932 |
+
}
|
| 2933 |
+
],
|
| 2934 |
+
"logging_steps": 2,
|
| 2935 |
+
"max_steps": 2010,
|
| 2936 |
+
"num_input_tokens_seen": 0,
|
| 2937 |
+
"num_train_epochs": 15,
|
| 2938 |
+
"save_steps": 500,
|
| 2939 |
+
"stateful_callbacks": {
|
| 2940 |
+
"EarlyStoppingCallback": {
|
| 2941 |
+
"args": {
|
| 2942 |
+
"early_stopping_patience": 5,
|
| 2943 |
+
"early_stopping_threshold": 0.01
|
| 2944 |
+
},
|
| 2945 |
+
"attributes": {
|
| 2946 |
+
"early_stopping_patience_counter": 0
|
| 2947 |
+
}
|
| 2948 |
+
},
|
| 2949 |
+
"TrainerControl": {
|
| 2950 |
+
"args": {
|
| 2951 |
+
"should_epoch_stop": false,
|
| 2952 |
+
"should_evaluate": false,
|
| 2953 |
+
"should_log": false,
|
| 2954 |
+
"should_save": true,
|
| 2955 |
+
"should_training_stop": false
|
| 2956 |
+
},
|
| 2957 |
+
"attributes": {}
|
| 2958 |
+
}
|
| 2959 |
+
},
|
| 2960 |
+
"total_flos": 844829150146560.0,
|
| 2961 |
+
"train_batch_size": 16,
|
| 2962 |
+
"trial_name": null,
|
| 2963 |
+
"trial_params": null
|
| 2964 |
+
}
|
checkpoint-804/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:da8fd87379f92b44bfe24ad17cf7970f1ba9767b514a253b1ea5d776a21c8229
|
| 3 |
+
size 5240
|
config.json
ADDED
|
@@ -0,0 +1,103 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "vinai/phobert-base",
|
| 3 |
+
"_num_labels": 34,
|
| 4 |
+
"architectures": [
|
| 5 |
+
"RobertaForSequenceClassification"
|
| 6 |
+
],
|
| 7 |
+
"attention_probs_dropout_prob": 0.1,
|
| 8 |
+
"bos_token_id": 0,
|
| 9 |
+
"classifier_dropout": null,
|
| 10 |
+
"eos_token_id": 2,
|
| 11 |
+
"gradient_checkpointing": false,
|
| 12 |
+
"hidden_act": "gelu",
|
| 13 |
+
"hidden_dropout_prob": 0.1,
|
| 14 |
+
"hidden_size": 768,
|
| 15 |
+
"id2label": {
|
| 16 |
+
"0": " H\u1ecfi v\u1ec1 c\u00e1ch x\u1eed l\u00fd khi g\u1eb7p v\u1ea5n \u0111\u1ec1 v\u1edbi s\u1ea3n ph\u1ea9m b\u1ecb l\u1ed7i",
|
| 17 |
+
"1": " H\u1ecfi v\u1ec1 t\u00ecnh tr\u1ea1ng \u0111\u01a1n h\u00e0ng v\u00e0 c\u00e1ch th\u1ee9c h\u1ee7y \u0111\u01a1n h\u00e0ng",
|
| 18 |
+
"2": " Y\u00eau c\u1ea7u h\u1ed7 tr\u1ee3 khi g\u1eb7p v\u1ea5n \u0111\u1ec1 v\u1edbi Shopee Mall ho\u1eb7c c\u00e1c shop kh\u00e1c",
|
| 19 |
+
"3": "H\u1ecfi v\u1ec1 ch\u00ednh s\u00e1ch b\u1ea3o h\u00e0nh v\u00e0 h\u1ed7 tr\u1ee3 k\u1ef9 thu\u1eadt",
|
| 20 |
+
"4": "H\u1ecfi v\u1ec1 ch\u00ednh s\u00e1ch b\u1ea3o m\u1eadt th\u00f4ng tin kh\u00e1ch h\u00e0ng",
|
| 21 |
+
"5": "H\u1ecfi v\u1ec1 ch\u00ednh s\u00e1ch \u0111i\u1ec3m th\u01b0\u1edfng v\u00e0 c\u00e1ch s\u1eed d\u1ee5ng",
|
| 22 |
+
"6": "H\u1ecfi v\u1ec1 ch\u00ednh s\u00e1ch \u0111\u1ed1i v\u1edbi kh\u00e1ch h\u00e0ng th\u00e2n thi\u1ebft v\u00e0 \u01b0u \u0111\u00e3i sinh nh\u1eadt",
|
| 23 |
+
"7": "H\u1ecfi v\u1ec1 ch\u00ednh s\u00e1ch \u0111\u1ed5i tr\u1ea3 v\u00e0 ho\u00e0n ti\u1ec1n",
|
| 24 |
+
"8": "H\u1ecfi v\u1ec1 ch\u01b0\u01a1ng tr\u00ecnh khuy\u1ebfn m\u00e3i, \u01b0u \u0111\u00e3i v\u00e0 flash sale",
|
| 25 |
+
"9": "H\u1ecfi v\u1ec1 ch\u01b0\u01a1ng tr\u00ecnh \u01b0u \u0111\u00e3i combo v\u00e0 ph\u00ed v\u1eadn chuy\u1ec3n khi mua combo",
|
| 26 |
+
"10": "H\u1ecfi v\u1ec1 c\u00e1ch th\u1ee9c g\u00f3i h\u00e0ng, \u0111\u1eb7c bi\u1ec7t \u0111\u1ed1i v\u1edbi h\u00e0ng d\u1ec5 v\u1ee1",
|
| 27 |
+
"11": "H\u1ecfi v\u1ec1 c\u00e1ch th\u1ee9c nh\u1eadn h\u00e0ng t\u1ea1i c\u1eeda h\u00e0ng ho\u1eb7c \u0111\u1ecba \u0111i\u1ec3m kh\u00e1c",
|
| 28 |
+
"12": "H\u1ecfi v\u1ec1 c\u00e1ch th\u1ee9c nh\u1eadn h\u00f3a \u0111\u01a1n \u0111\u1ecf",
|
| 29 |
+
"13": "H\u1ecfi v\u1ec1 c\u00e1ch th\u1ee9c \u0111\u1ed5i qu\u00e0 t\u1eb7ng k\u00e8m ho\u1eb7c s\u1ea3n ph\u1ea9m khuy\u1ebfn m\u00e3i",
|
| 30 |
+
"14": "H\u1ecfi v\u1ec1 c\u00e1ch th\u1ee9c \u0111\u1ed5i, tr\u1ea3, v\u00e0 s\u1eeda \u0111\u1ed5i th\u00f4ng tin \u0111\u01a1n h\u00e0ng",
|
| 31 |
+
"15": "H\u1ecfi v\u1ec1 gi\u00e1 s\u1ea3n ph\u1ea9m",
|
| 32 |
+
"16": "H\u1ecfi v\u1ec1 h\u01b0\u1edbng d\u1eabn s\u1eed d\u1ee5ng v\u00e0 b\u1ea3o qu\u1ea3n s\u1ea3n ph\u1ea9m",
|
| 33 |
+
"17": "H\u1ecfi v\u1ec1 k\u00edch th\u01b0\u1edbc, m\u00e0u s\u1eafc v\u00e0 ch\u1ea5t li\u1ec7u s\u1ea3n ph\u1ea9m",
|
| 34 |
+
"18": "H\u1ecfi v\u1ec1 l\u1ecbch s\u1eed mua h\u00e0ng v\u00e0 ch\u01b0\u01a1ng tr\u00ecnh kh\u00e1ch h\u00e0ng th\u00e2n thi\u1ebft",
|
| 35 |
+
"19": "H\u1ecfi v\u1ec1 ngu\u1ed3n g\u1ed1c, xu\u1ea5t x\u1ee9 s\u1ea3n ph\u1ea9m",
|
| 36 |
+
"20": "H\u1ecfi v\u1ec1 ph\u00ed v\u00e0 th\u1eddi gian v\u1eadn chuy\u1ec3n",
|
| 37 |
+
"21": "H\u1ecfi v\u1ec1 ph\u00ed v\u00e0 \u0111i\u1ec1u ki\u1ec7n \u0111\u1eb7t h\u00e0ng tr\u01b0\u1edbc s\u1ea3n ph\u1ea9m ch\u01b0a ra m\u1eaft",
|
| 38 |
+
"22": "H\u1ecfi v\u1ec1 ph\u01b0\u01a1ng th\u1ee9c thanh to\u00e1n v\u00e0 h\u1ed7 tr\u1ee3 khi g\u1eb7p s\u1ef1 c\u1ed1 thanh to\u00e1n",
|
| 39 |
+
"23": "H\u1ecfi v\u1ec1 s\u1ea3n ph\u1ea9m m\u1edbi v\u1ec1, s\u1ea3n ph\u1ea9m b\u00e1n ch\u1ea1y nh\u1ea5t v\u00e0 hot trend",
|
| 40 |
+
"24": "H\u1ecfi v\u1ec1 s\u1ea3n ph\u1ea9m \u0111\u00e3 qua s\u1eed d\u1ee5ng ho\u1eb7c h\u00e0ng \u0111\u1ed5i tr\u1ea3",
|
| 41 |
+
"25": "H\u1ecfi v\u1ec1 th\u1eddi gian giao h\u00e0ng d\u1ef1 ki\u1ebfn",
|
| 42 |
+
"26": "H\u1ecfi v\u1ec1 t\u00ecnh tr\u1ea1ng c\u00f2n h\u00e0ng v\u00e0 s\u1ed1 l\u01b0\u1ee3ng mua t\u1ed1i \u0111a",
|
| 43 |
+
"27": "Y\u00eau c\u1ea7u gi\u1ea3m gi\u00e1 khuy\u1ebfn m\u00e3i ho\u1eb7c m\u00e3 gi\u1ea3m gi\u00e1",
|
| 44 |
+
"28": "Y\u00eau c\u1ea7u g\u1eedi h\u00ecnh \u1ea3nh th\u1eadt c\u1ee7a s\u1ea3n ph\u1ea9m",
|
| 45 |
+
"29": "Y\u00eau c\u1ea7u h\u1ed7 tr\u1ee3 khi kh\u00f4ng nh\u1eadn \u0111\u01b0\u1ee3c h\u00e0ng ho\u1eb7c h\u00e0ng kh\u00f4ng \u0111\u00fang m\u00f4 t\u1ea3",
|
| 46 |
+
"30": "Y\u00eau c\u1ea7u th\u00f4ng tin chi ti\u1ebft s\u1ea3n ph\u1ea9m",
|
| 47 |
+
"31": "Y\u00eau c\u1ea7u t\u01b0 v\u1ea5n ph\u1ed1i \u0111\u1ed3 ho\u1eb7c c\u00e1ch s\u1eed d\u1ee5ng s\u1ea3n ph\u1ea9m",
|
| 48 |
+
"32": "Y\u00eau c\u1ea7u t\u01b0 v\u1ea5n s\u1ea3n ph\u1ea9m ph\u00f9 h\u1ee3p",
|
| 49 |
+
"33": "Y\u00eau c\u1ea7u x\u00e1c nh\u1eadn th\u00f4ng tin \u0111\u01a1n h\u00e0ng ho\u1eb7c \u0111\u1ecba ch\u1ec9 giao h\u00e0ng"
|
| 50 |
+
},
|
| 51 |
+
"initializer_range": 0.02,
|
| 52 |
+
"intermediate_size": 3072,
|
| 53 |
+
"label2id": {
|
| 54 |
+
" H\u1ecfi v\u1ec1 c\u00e1ch x\u1eed l\u00fd khi g\u1eb7p v\u1ea5n \u0111\u1ec1 v\u1edbi s\u1ea3n ph\u1ea9m b\u1ecb l\u1ed7i": 0,
|
| 55 |
+
" H\u1ecfi v\u1ec1 t\u00ecnh tr\u1ea1ng \u0111\u01a1n h\u00e0ng v\u00e0 c\u00e1ch th\u1ee9c h\u1ee7y \u0111\u01a1n h\u00e0ng": 1,
|
| 56 |
+
" Y\u00eau c\u1ea7u h\u1ed7 tr\u1ee3 khi g\u1eb7p v\u1ea5n \u0111\u1ec1 v\u1edbi Shopee Mall ho\u1eb7c c\u00e1c shop kh\u00e1c": 2,
|
| 57 |
+
"H\u1ecfi v\u1ec1 ch\u00ednh s\u00e1ch b\u1ea3o h\u00e0nh v\u00e0 h\u1ed7 tr\u1ee3 k\u1ef9 thu\u1eadt": 3,
|
| 58 |
+
"H\u1ecfi v\u1ec1 ch\u00ednh s\u00e1ch b\u1ea3o m\u1eadt th\u00f4ng tin kh\u00e1ch h\u00e0ng": 4,
|
| 59 |
+
"H\u1ecfi v\u1ec1 ch\u00ednh s\u00e1ch \u0111i\u1ec3m th\u01b0\u1edfng v\u00e0 c\u00e1ch s\u1eed d\u1ee5ng": 5,
|
| 60 |
+
"H\u1ecfi v\u1ec1 ch\u00ednh s\u00e1ch \u0111\u1ed1i v\u1edbi kh\u00e1ch h\u00e0ng th\u00e2n thi\u1ebft v\u00e0 \u01b0u \u0111\u00e3i sinh nh\u1eadt": 6,
|
| 61 |
+
"H\u1ecfi v\u1ec1 ch\u00ednh s\u00e1ch \u0111\u1ed5i tr\u1ea3 v\u00e0 ho\u00e0n ti\u1ec1n": 7,
|
| 62 |
+
"H\u1ecfi v\u1ec1 ch\u01b0\u01a1ng tr\u00ecnh khuy\u1ebfn m\u00e3i, \u01b0u \u0111\u00e3i v\u00e0 flash sale": 8,
|
| 63 |
+
"H\u1ecfi v\u1ec1 ch\u01b0\u01a1ng tr\u00ecnh \u01b0u \u0111\u00e3i combo v\u00e0 ph\u00ed v\u1eadn chuy\u1ec3n khi mua combo": 9,
|
| 64 |
+
"H\u1ecfi v\u1ec1 c\u00e1ch th\u1ee9c g\u00f3i h\u00e0ng, \u0111\u1eb7c bi\u1ec7t \u0111\u1ed1i v\u1edbi h\u00e0ng d\u1ec5 v\u1ee1": 10,
|
| 65 |
+
"H\u1ecfi v\u1ec1 c\u00e1ch th\u1ee9c nh\u1eadn h\u00e0ng t\u1ea1i c\u1eeda h\u00e0ng ho\u1eb7c \u0111\u1ecba \u0111i\u1ec3m kh\u00e1c": 11,
|
| 66 |
+
"H\u1ecfi v\u1ec1 c\u00e1ch th\u1ee9c nh\u1eadn h\u00f3a \u0111\u01a1n \u0111\u1ecf": 12,
|
| 67 |
+
"H\u1ecfi v\u1ec1 c\u00e1ch th\u1ee9c \u0111\u1ed5i qu\u00e0 t\u1eb7ng k\u00e8m ho\u1eb7c s\u1ea3n ph\u1ea9m khuy\u1ebfn m\u00e3i": 13,
|
| 68 |
+
"H\u1ecfi v\u1ec1 c\u00e1ch th\u1ee9c \u0111\u1ed5i, tr\u1ea3, v\u00e0 s\u1eeda \u0111\u1ed5i th\u00f4ng tin \u0111\u01a1n h\u00e0ng": 14,
|
| 69 |
+
"H\u1ecfi v\u1ec1 gi\u00e1 s\u1ea3n ph\u1ea9m": 15,
|
| 70 |
+
"H\u1ecfi v\u1ec1 h\u01b0\u1edbng d\u1eabn s\u1eed d\u1ee5ng v\u00e0 b\u1ea3o qu\u1ea3n s\u1ea3n ph\u1ea9m": 16,
|
| 71 |
+
"H\u1ecfi v\u1ec1 k\u00edch th\u01b0\u1edbc, m\u00e0u s\u1eafc v\u00e0 ch\u1ea5t li\u1ec7u s\u1ea3n ph\u1ea9m": 17,
|
| 72 |
+
"H\u1ecfi v\u1ec1 l\u1ecbch s\u1eed mua h\u00e0ng v\u00e0 ch\u01b0\u01a1ng tr\u00ecnh kh\u00e1ch h\u00e0ng th\u00e2n thi\u1ebft": 18,
|
| 73 |
+
"H\u1ecfi v\u1ec1 ngu\u1ed3n g\u1ed1c, xu\u1ea5t x\u1ee9 s\u1ea3n ph\u1ea9m": 19,
|
| 74 |
+
"H\u1ecfi v\u1ec1 ph\u00ed v\u00e0 th\u1eddi gian v\u1eadn chuy\u1ec3n": 20,
|
| 75 |
+
"H\u1ecfi v\u1ec1 ph\u00ed v\u00e0 \u0111i\u1ec1u ki\u1ec7n \u0111\u1eb7t h\u00e0ng tr\u01b0\u1edbc s\u1ea3n ph\u1ea9m ch\u01b0a ra m\u1eaft": 21,
|
| 76 |
+
"H\u1ecfi v\u1ec1 ph\u01b0\u01a1ng th\u1ee9c thanh to\u00e1n v\u00e0 h\u1ed7 tr\u1ee3 khi g\u1eb7p s\u1ef1 c\u1ed1 thanh to\u00e1n": 22,
|
| 77 |
+
"H\u1ecfi v\u1ec1 s\u1ea3n ph\u1ea9m m\u1edbi v\u1ec1, s\u1ea3n ph\u1ea9m b\u00e1n ch\u1ea1y nh\u1ea5t v\u00e0 hot trend": 23,
|
| 78 |
+
"H\u1ecfi v\u1ec1 s\u1ea3n ph\u1ea9m \u0111\u00e3 qua s\u1eed d\u1ee5ng ho\u1eb7c h\u00e0ng \u0111\u1ed5i tr\u1ea3": 24,
|
| 79 |
+
"H\u1ecfi v\u1ec1 th\u1eddi gian giao h\u00e0ng d\u1ef1 ki\u1ebfn": 25,
|
| 80 |
+
"H\u1ecfi v\u1ec1 t\u00ecnh tr\u1ea1ng c\u00f2n h\u00e0ng v\u00e0 s\u1ed1 l\u01b0\u1ee3ng mua t\u1ed1i \u0111a": 26,
|
| 81 |
+
"Y\u00eau c\u1ea7u gi\u1ea3m gi\u00e1 khuy\u1ebfn m\u00e3i ho\u1eb7c m\u00e3 gi\u1ea3m gi\u00e1": 27,
|
| 82 |
+
"Y\u00eau c\u1ea7u g\u1eedi h\u00ecnh \u1ea3nh th\u1eadt c\u1ee7a s\u1ea3n ph\u1ea9m": 28,
|
| 83 |
+
"Y\u00eau c\u1ea7u h\u1ed7 tr\u1ee3 khi kh\u00f4ng nh\u1eadn \u0111\u01b0\u1ee3c h\u00e0ng ho\u1eb7c h\u00e0ng kh\u00f4ng \u0111\u00fang m\u00f4 t\u1ea3": 29,
|
| 84 |
+
"Y\u00eau c\u1ea7u th\u00f4ng tin chi ti\u1ebft s\u1ea3n ph\u1ea9m": 30,
|
| 85 |
+
"Y\u00eau c\u1ea7u t\u01b0 v\u1ea5n ph\u1ed1i \u0111\u1ed3 ho\u1eb7c c\u00e1ch s\u1eed d\u1ee5ng s\u1ea3n ph\u1ea9m": 31,
|
| 86 |
+
"Y\u00eau c\u1ea7u t\u01b0 v\u1ea5n s\u1ea3n ph\u1ea9m ph\u00f9 h\u1ee3p": 32,
|
| 87 |
+
"Y\u00eau c\u1ea7u x\u00e1c nh\u1eadn th\u00f4ng tin \u0111\u01a1n h\u00e0ng ho\u1eb7c \u0111\u1ecba ch\u1ec9 giao h\u00e0ng": 33
|
| 88 |
+
},
|
| 89 |
+
"layer_norm_eps": 1e-05,
|
| 90 |
+
"max_position_embeddings": 258,
|
| 91 |
+
"model_type": "roberta",
|
| 92 |
+
"num_attention_heads": 12,
|
| 93 |
+
"num_hidden_layers": 12,
|
| 94 |
+
"pad_token_id": 1,
|
| 95 |
+
"position_embedding_type": "absolute",
|
| 96 |
+
"problem_type": "single_label_classification",
|
| 97 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 98 |
+
"torch_dtype": "float32",
|
| 99 |
+
"transformers_version": "4.44.1",
|
| 100 |
+
"type_vocab_size": 1,
|
| 101 |
+
"use_cache": true,
|
| 102 |
+
"vocab_size": 64001
|
| 103 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5826d0f1dcc22361c7b62ea13a25df96f2f5df8fd171223ca9ed8a55fe74b661
|
| 3 |
+
size 540121816
|
runs/Sep04_02-07-36_r-cuwfnguyen-autotrain-advanced-tqduow70-6d178-7tndo/events.out.tfevents.1725415658.r-cuwfnguyen-autotrain-advanced-tqduow70-6d178-7tndo.102.0
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0ec5b4d817e2fecec592f3f01a248ad828739355f818f116bbe73f2b8140b883
|
| 3 |
+
size 178006
|
runs/Sep04_02-07-36_r-cuwfnguyen-autotrain-advanced-tqduow70-6d178-7tndo/events.out.tfevents.1725431613.r-cuwfnguyen-autotrain-advanced-tqduow70-6d178-7tndo.102.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e9040a9b95016ed6efc609a507370796454c738c24933c913ad18cec75d1a52e
|
| 3 |
+
size 921
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": "<s>",
|
| 3 |
+
"cls_token": "<s>",
|
| 4 |
+
"eos_token": "</s>",
|
| 5 |
+
"mask_token": "<mask>",
|
| 6 |
+
"pad_token": "<pad>",
|
| 7 |
+
"sep_token": "</s>",
|
| 8 |
+
"unk_token": "<unk>"
|
| 9 |
+
}
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,54 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"64000": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": true,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"mask_token": "<mask>",
|
| 49 |
+
"model_max_length": 1000000000000000019884624838656,
|
| 50 |
+
"pad_token": "<pad>",
|
| 51 |
+
"sep_token": "</s>",
|
| 52 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 53 |
+
"unk_token": "<unk>"
|
| 54 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:da8fd87379f92b44bfe24ad17cf7970f1ba9767b514a253b1ea5d776a21c8229
|
| 3 |
+
size 5240
|
training_params.json
ADDED
|
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"data_path": "autotrain-okjuj-buuqs/autotrain-data",
|
| 3 |
+
"model": "vinai/phobert-base",
|
| 4 |
+
"lr": 5e-05,
|
| 5 |
+
"epochs": 15,
|
| 6 |
+
"max_seq_length": 128,
|
| 7 |
+
"batch_size": 16,
|
| 8 |
+
"warmup_ratio": 0.1,
|
| 9 |
+
"gradient_accumulation": 1,
|
| 10 |
+
"optimizer": "adamw_torch",
|
| 11 |
+
"scheduler": "linear",
|
| 12 |
+
"weight_decay": 0.0,
|
| 13 |
+
"max_grad_norm": 1.0,
|
| 14 |
+
"seed": 42,
|
| 15 |
+
"train_split": "train",
|
| 16 |
+
"valid_split": "validation",
|
| 17 |
+
"text_column": "autotrain_text",
|
| 18 |
+
"target_column": "autotrain_label",
|
| 19 |
+
"logging_steps": -1,
|
| 20 |
+
"project_name": "autotrain-okjuj-buuqs",
|
| 21 |
+
"auto_find_batch_size": false,
|
| 22 |
+
"mixed_precision": "fp16",
|
| 23 |
+
"save_total_limit": 1,
|
| 24 |
+
"push_to_hub": true,
|
| 25 |
+
"eval_strategy": "epoch",
|
| 26 |
+
"username": "cuwfnguyen",
|
| 27 |
+
"log": "tensorboard",
|
| 28 |
+
"early_stopping_patience": 5,
|
| 29 |
+
"early_stopping_threshold": 0.01
|
| 30 |
+
}
|
vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|