Upload additional checkpoints
Browse filesThis view is limited to 50 files because it contains too many changes. See raw diff
- ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/model.safetensors +3 -0
- ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1161/added_tokens.json +3 -0
- ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1161/bpe.codes +0 -0
- ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1161/model.safetensors +3 -0
- ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1161/optimizer.pt +3 -0
- ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1161/rng_state.pth +3 -0
- ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1161/scheduler.pt +3 -0
- ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1161/special_tokens_map.json +9 -0
- ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1161/tokenizer_config.json +55 -0
- ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1161/trainer_state.json +464 -0
- ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1161/training_args.bin +3 -0
- ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1161/vocab.txt +0 -0
- ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1548/added_tokens.json +3 -0
- ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1548/bpe.codes +0 -0
- ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1548/model.safetensors +3 -0
- ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1548/optimizer.pt +3 -0
- ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1548/rng_state.pth +3 -0
- ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1548/scheduler.pt +3 -0
- ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1548/special_tokens_map.json +9 -0
- ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1548/tokenizer_config.json +55 -0
- ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1548/trainer_state.json +602 -0
- ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1548/training_args.bin +3 -0
- ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1548/vocab.txt +0 -0
- ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/model.safetensors +3 -0
- ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-2592/added_tokens.json +3 -0
- ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-2592/bpe.codes +0 -0
- ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-2592/model.safetensors +3 -0
- ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-2592/optimizer.pt +3 -0
- ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-2592/rng_state.pth +3 -0
- ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-2592/scheduler.pt +3 -0
- ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-2592/special_tokens_map.json +9 -0
- ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-2592/tokenizer_config.json +55 -0
- ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-2592/trainer_state.json +830 -0
- ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-2592/training_args.bin +3 -0
- ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-2592/vocab.txt +0 -0
- ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-3888/added_tokens.json +3 -0
- ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-3888/bpe.codes +0 -0
- ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-3888/model.safetensors +3 -0
- ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-3888/optimizer.pt +3 -0
- ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-3888/rng_state.pth +3 -0
- ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-3888/scheduler.pt +3 -0
- ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-3888/special_tokens_map.json +9 -0
- ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-3888/tokenizer_config.json +55 -0
- ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-3888/trainer_state.json +1227 -0
- ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-3888/training_args.bin +3 -0
- ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-3888/vocab.txt +0 -0
- ood-uit_vsfc-phobert-topic-domain-03-s44-20260812T044906Z-6beb231d/model.safetensors +3 -0
- ood-uit_vsfc-phobert-topic-domain-03-s44-20260812T044906Z-6beb231d/trainer/checkpoint-1850/added_tokens.json +3 -0
- ood-uit_vsfc-phobert-topic-domain-03-s44-20260812T044906Z-6beb231d/trainer/checkpoint-1850/bpe.codes +0 -0
- ood-uit_vsfc-phobert-topic-domain-03-s44-20260812T044906Z-6beb231d/trainer/checkpoint-1850/model.safetensors +3 -0
ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7b01430813911af9dc9343d0bce7a034a76c0e23c959cebfdfef8d6fcb718657
|
| 3 |
+
size 540026404
|
ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1161/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1161/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1161/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a4851c672a49399a9fa2e0e1f6b8f5ee330da38b054666c94269c8567effdad1
|
| 3 |
+
size 540026436
|
ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1161/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c7d1ffe639a790344a6bd0e270941c0e59b1153440163dc52ef4ad2ff18627a7
|
| 3 |
+
size 1075446330
|
ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1161/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:98a4dcb9648968ceae65df7b713c379e0ba718e9d62190431de288a87706beb1
|
| 3 |
+
size 14244
|
ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1161/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c298f638e7632fed37bb986facf0a51195136ddc2dcf875aa03de77ff22f3c0f
|
| 3 |
+
size 1064
|
ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1161/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 |
+
}
|
ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1161/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1161/trainer_state.json
ADDED
|
@@ -0,0 +1,464 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1161,
|
| 3 |
+
"best_metric": 0.8947346271312796,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1161",
|
| 5 |
+
"epoch": 9.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1161,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.19455252918287938,
|
| 14 |
+
"grad_norm": 12.175275802612305,
|
| 15 |
+
"learning_rate": 1.86046511627907e-06,
|
| 16 |
+
"loss": 4.6741,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.38910505836575876,
|
| 21 |
+
"grad_norm": 14.172719955444336,
|
| 22 |
+
"learning_rate": 3.798449612403101e-06,
|
| 23 |
+
"loss": 3.6832,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.5836575875486382,
|
| 28 |
+
"grad_norm": 9.232404708862305,
|
| 29 |
+
"learning_rate": 5.736434108527133e-06,
|
| 30 |
+
"loss": 2.9012,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.7782101167315175,
|
| 35 |
+
"grad_norm": 9.204726219177246,
|
| 36 |
+
"learning_rate": 7.674418604651164e-06,
|
| 37 |
+
"loss": 2.6558,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.9727626459143969,
|
| 42 |
+
"grad_norm": 13.36579704284668,
|
| 43 |
+
"learning_rate": 9.612403100775196e-06,
|
| 44 |
+
"loss": 2.304,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 1.0,
|
| 49 |
+
"eval_accuracy": 0.8533916849015317,
|
| 50 |
+
"eval_loss": 0.4367648959159851,
|
| 51 |
+
"eval_macro_f1": 0.5793912034860027,
|
| 52 |
+
"eval_runtime": 1.2368,
|
| 53 |
+
"eval_samples_per_second": 369.498,
|
| 54 |
+
"eval_steps_per_second": 23.447,
|
| 55 |
+
"eval_weighted_f1": 0.8184832282800687,
|
| 56 |
+
"step": 129
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 1.1634241245136188,
|
| 60 |
+
"grad_norm": 8.186969757080078,
|
| 61 |
+
"learning_rate": 1.1550387596899227e-05,
|
| 62 |
+
"loss": 1.7317,
|
| 63 |
+
"step": 150
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 1.3579766536964981,
|
| 67 |
+
"grad_norm": 44.49769592285156,
|
| 68 |
+
"learning_rate": 1.3488372093023257e-05,
|
| 69 |
+
"loss": 1.5504,
|
| 70 |
+
"step": 175
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.5525291828793775,
|
| 74 |
+
"grad_norm": 15.99069595336914,
|
| 75 |
+
"learning_rate": 1.542635658914729e-05,
|
| 76 |
+
"loss": 1.49,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.7470817120622568,
|
| 81 |
+
"grad_norm": 15.231589317321777,
|
| 82 |
+
"learning_rate": 1.736434108527132e-05,
|
| 83 |
+
"loss": 1.2571,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.9416342412451362,
|
| 88 |
+
"grad_norm": 11.705756187438965,
|
| 89 |
+
"learning_rate": 1.9302325581395353e-05,
|
| 90 |
+
"loss": 1.0896,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 2.0,
|
| 95 |
+
"eval_accuracy": 0.9059080962800875,
|
| 96 |
+
"eval_loss": 0.29254111647605896,
|
| 97 |
+
"eval_macro_f1": 0.7893837634028514,
|
| 98 |
+
"eval_runtime": 0.9698,
|
| 99 |
+
"eval_samples_per_second": 471.252,
|
| 100 |
+
"eval_steps_per_second": 29.904,
|
| 101 |
+
"eval_weighted_f1": 0.8965275247843335,
|
| 102 |
+
"step": 258
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 2.132295719844358,
|
| 106 |
+
"grad_norm": 7.289089679718018,
|
| 107 |
+
"learning_rate": 1.9862187769164516e-05,
|
| 108 |
+
"loss": 1.0811,
|
| 109 |
+
"step": 275
|
| 110 |
+
},
|
| 111 |
+
{
|
| 112 |
+
"epoch": 2.3268482490272375,
|
| 113 |
+
"grad_norm": 19.28726577758789,
|
| 114 |
+
"learning_rate": 1.9646856158484065e-05,
|
| 115 |
+
"loss": 0.938,
|
| 116 |
+
"step": 300
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 2.5214007782101167,
|
| 120 |
+
"grad_norm": 14.579724311828613,
|
| 121 |
+
"learning_rate": 1.943152454780362e-05,
|
| 122 |
+
"loss": 0.8457,
|
| 123 |
+
"step": 325
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 2.7159533073929962,
|
| 127 |
+
"grad_norm": 26.952871322631836,
|
| 128 |
+
"learning_rate": 1.921619293712317e-05,
|
| 129 |
+
"loss": 1.0485,
|
| 130 |
+
"step": 350
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.9105058365758754,
|
| 134 |
+
"grad_norm": 11.434037208557129,
|
| 135 |
+
"learning_rate": 1.9000861326442723e-05,
|
| 136 |
+
"loss": 1.1113,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 3.0,
|
| 141 |
+
"eval_accuracy": 0.9146608315098468,
|
| 142 |
+
"eval_loss": 0.24419832229614258,
|
| 143 |
+
"eval_macro_f1": 0.8422144131192929,
|
| 144 |
+
"eval_runtime": 1.0063,
|
| 145 |
+
"eval_samples_per_second": 454.118,
|
| 146 |
+
"eval_steps_per_second": 28.817,
|
| 147 |
+
"eval_weighted_f1": 0.918285359370874,
|
| 148 |
+
"step": 387
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 3.1011673151750974,
|
| 152 |
+
"grad_norm": 23.606746673583984,
|
| 153 |
+
"learning_rate": 1.8785529715762276e-05,
|
| 154 |
+
"loss": 0.8873,
|
| 155 |
+
"step": 400
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"epoch": 3.2957198443579765,
|
| 159 |
+
"grad_norm": 13.393070220947266,
|
| 160 |
+
"learning_rate": 1.857019810508183e-05,
|
| 161 |
+
"loss": 0.6112,
|
| 162 |
+
"step": 425
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"epoch": 3.490272373540856,
|
| 166 |
+
"grad_norm": 2.7602410316467285,
|
| 167 |
+
"learning_rate": 1.835486649440138e-05,
|
| 168 |
+
"loss": 0.7564,
|
| 169 |
+
"step": 450
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 3.6848249027237356,
|
| 173 |
+
"grad_norm": 38.35893630981445,
|
| 174 |
+
"learning_rate": 1.813953488372093e-05,
|
| 175 |
+
"loss": 0.8032,
|
| 176 |
+
"step": 475
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 3.8793774319066148,
|
| 180 |
+
"grad_norm": 26.61602783203125,
|
| 181 |
+
"learning_rate": 1.7924203273040483e-05,
|
| 182 |
+
"loss": 0.9455,
|
| 183 |
+
"step": 500
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 4.0,
|
| 187 |
+
"eval_accuracy": 0.925601750547046,
|
| 188 |
+
"eval_loss": 0.24463064968585968,
|
| 189 |
+
"eval_macro_f1": 0.8577216356628122,
|
| 190 |
+
"eval_runtime": 1.6336,
|
| 191 |
+
"eval_samples_per_second": 279.754,
|
| 192 |
+
"eval_steps_per_second": 17.752,
|
| 193 |
+
"eval_weighted_f1": 0.9254741614987465,
|
| 194 |
+
"step": 516
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 4.070038910505836,
|
| 198 |
+
"grad_norm": 39.84136199951172,
|
| 199 |
+
"learning_rate": 1.7708871662360035e-05,
|
| 200 |
+
"loss": 0.6896,
|
| 201 |
+
"step": 525
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 4.264591439688716,
|
| 205 |
+
"grad_norm": 1.3257018327713013,
|
| 206 |
+
"learning_rate": 1.7493540051679588e-05,
|
| 207 |
+
"loss": 0.5164,
|
| 208 |
+
"step": 550
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"epoch": 4.4591439688715955,
|
| 212 |
+
"grad_norm": 4.106621265411377,
|
| 213 |
+
"learning_rate": 1.727820844099914e-05,
|
| 214 |
+
"loss": 0.4498,
|
| 215 |
+
"step": 575
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 4.653696498054475,
|
| 219 |
+
"grad_norm": 27.42701530456543,
|
| 220 |
+
"learning_rate": 1.7062876830318693e-05,
|
| 221 |
+
"loss": 0.7688,
|
| 222 |
+
"step": 600
|
| 223 |
+
},
|
| 224 |
+
{
|
| 225 |
+
"epoch": 4.848249027237354,
|
| 226 |
+
"grad_norm": 8.981760025024414,
|
| 227 |
+
"learning_rate": 1.6847545219638243e-05,
|
| 228 |
+
"loss": 0.6506,
|
| 229 |
+
"step": 625
|
| 230 |
+
},
|
| 231 |
+
{
|
| 232 |
+
"epoch": 5.0,
|
| 233 |
+
"eval_accuracy": 0.9277899343544858,
|
| 234 |
+
"eval_loss": 0.28580012917518616,
|
| 235 |
+
"eval_macro_f1": 0.870879330908566,
|
| 236 |
+
"eval_runtime": 1.0274,
|
| 237 |
+
"eval_samples_per_second": 444.802,
|
| 238 |
+
"eval_steps_per_second": 28.226,
|
| 239 |
+
"eval_weighted_f1": 0.9302773652022999,
|
| 240 |
+
"step": 645
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 5.038910505836576,
|
| 244 |
+
"grad_norm": 46.58235549926758,
|
| 245 |
+
"learning_rate": 1.6632213608957795e-05,
|
| 246 |
+
"loss": 0.7474,
|
| 247 |
+
"step": 650
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 5.233463035019455,
|
| 251 |
+
"grad_norm": 2.2681214809417725,
|
| 252 |
+
"learning_rate": 1.6416881998277348e-05,
|
| 253 |
+
"loss": 0.4227,
|
| 254 |
+
"step": 675
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 5.428015564202335,
|
| 258 |
+
"grad_norm": 49.91563415527344,
|
| 259 |
+
"learning_rate": 1.62015503875969e-05,
|
| 260 |
+
"loss": 0.5143,
|
| 261 |
+
"step": 700
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 5.622568093385214,
|
| 265 |
+
"grad_norm": 41.00137710571289,
|
| 266 |
+
"learning_rate": 1.5986218776916453e-05,
|
| 267 |
+
"loss": 0.4182,
|
| 268 |
+
"step": 725
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 5.817120622568093,
|
| 272 |
+
"grad_norm": 9.387858390808105,
|
| 273 |
+
"learning_rate": 1.5770887166236006e-05,
|
| 274 |
+
"loss": 0.4013,
|
| 275 |
+
"step": 750
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 6.0,
|
| 279 |
+
"eval_accuracy": 0.9037199124726477,
|
| 280 |
+
"eval_loss": 0.3829730451107025,
|
| 281 |
+
"eval_macro_f1": 0.8311931591145209,
|
| 282 |
+
"eval_runtime": 0.9951,
|
| 283 |
+
"eval_samples_per_second": 459.272,
|
| 284 |
+
"eval_steps_per_second": 29.144,
|
| 285 |
+
"eval_weighted_f1": 0.9057689254816129,
|
| 286 |
+
"step": 774
|
| 287 |
+
},
|
| 288 |
+
{
|
| 289 |
+
"epoch": 6.007782101167315,
|
| 290 |
+
"grad_norm": 44.55717468261719,
|
| 291 |
+
"learning_rate": 1.555555555555556e-05,
|
| 292 |
+
"loss": 0.6504,
|
| 293 |
+
"step": 775
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 6.202334630350195,
|
| 297 |
+
"grad_norm": 94.3865966796875,
|
| 298 |
+
"learning_rate": 1.534022394487511e-05,
|
| 299 |
+
"loss": 0.412,
|
| 300 |
+
"step": 800
|
| 301 |
+
},
|
| 302 |
+
{
|
| 303 |
+
"epoch": 6.396887159533074,
|
| 304 |
+
"grad_norm": 9.678958892822266,
|
| 305 |
+
"learning_rate": 1.512489233419466e-05,
|
| 306 |
+
"loss": 0.4283,
|
| 307 |
+
"step": 825
|
| 308 |
+
},
|
| 309 |
+
{
|
| 310 |
+
"epoch": 6.591439688715953,
|
| 311 |
+
"grad_norm": 37.90494918823242,
|
| 312 |
+
"learning_rate": 1.4909560723514215e-05,
|
| 313 |
+
"loss": 0.4208,
|
| 314 |
+
"step": 850
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 6.785992217898833,
|
| 318 |
+
"grad_norm": 21.357500076293945,
|
| 319 |
+
"learning_rate": 1.4694229112833764e-05,
|
| 320 |
+
"loss": 0.5355,
|
| 321 |
+
"step": 875
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 6.980544747081712,
|
| 325 |
+
"grad_norm": 52.673919677734375,
|
| 326 |
+
"learning_rate": 1.4478897502153318e-05,
|
| 327 |
+
"loss": 0.3123,
|
| 328 |
+
"step": 900
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"epoch": 7.0,
|
| 332 |
+
"eval_accuracy": 0.9409190371991247,
|
| 333 |
+
"eval_loss": 0.2741667926311493,
|
| 334 |
+
"eval_macro_f1": 0.8910597249801194,
|
| 335 |
+
"eval_runtime": 1.3714,
|
| 336 |
+
"eval_samples_per_second": 333.247,
|
| 337 |
+
"eval_steps_per_second": 21.147,
|
| 338 |
+
"eval_weighted_f1": 0.9413592377918585,
|
| 339 |
+
"step": 903
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 7.171206225680934,
|
| 343 |
+
"grad_norm": 5.324210166931152,
|
| 344 |
+
"learning_rate": 1.426356589147287e-05,
|
| 345 |
+
"loss": 0.3897,
|
| 346 |
+
"step": 925
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 7.365758754863813,
|
| 350 |
+
"grad_norm": 21.306001663208008,
|
| 351 |
+
"learning_rate": 1.4048234280792422e-05,
|
| 352 |
+
"loss": 0.3909,
|
| 353 |
+
"step": 950
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 7.560311284046692,
|
| 357 |
+
"grad_norm": 1.1874115467071533,
|
| 358 |
+
"learning_rate": 1.3832902670111974e-05,
|
| 359 |
+
"loss": 0.4574,
|
| 360 |
+
"step": 975
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 7.754863813229572,
|
| 364 |
+
"grad_norm": 0.7569160461425781,
|
| 365 |
+
"learning_rate": 1.3617571059431525e-05,
|
| 366 |
+
"loss": 0.2395,
|
| 367 |
+
"step": 1000
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 7.9494163424124515,
|
| 371 |
+
"grad_norm": 0.5100299715995789,
|
| 372 |
+
"learning_rate": 1.3402239448751078e-05,
|
| 373 |
+
"loss": 0.2427,
|
| 374 |
+
"step": 1025
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 8.0,
|
| 378 |
+
"eval_accuracy": 0.9299781181619255,
|
| 379 |
+
"eval_loss": 0.3452863097190857,
|
| 380 |
+
"eval_macro_f1": 0.8730694907605123,
|
| 381 |
+
"eval_runtime": 0.957,
|
| 382 |
+
"eval_samples_per_second": 477.558,
|
| 383 |
+
"eval_steps_per_second": 30.305,
|
| 384 |
+
"eval_weighted_f1": 0.9314037441796743,
|
| 385 |
+
"step": 1032
|
| 386 |
+
},
|
| 387 |
+
{
|
| 388 |
+
"epoch": 8.140077821011673,
|
| 389 |
+
"grad_norm": 39.542598724365234,
|
| 390 |
+
"learning_rate": 1.318690783807063e-05,
|
| 391 |
+
"loss": 0.313,
|
| 392 |
+
"step": 1050
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 8.334630350194553,
|
| 396 |
+
"grad_norm": 1.1928895711898804,
|
| 397 |
+
"learning_rate": 1.2971576227390181e-05,
|
| 398 |
+
"loss": 0.2376,
|
| 399 |
+
"step": 1075
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 8.529182879377432,
|
| 403 |
+
"grad_norm": 10.282620429992676,
|
| 404 |
+
"learning_rate": 1.2756244616709734e-05,
|
| 405 |
+
"loss": 0.1612,
|
| 406 |
+
"step": 1100
|
| 407 |
+
},
|
| 408 |
+
{
|
| 409 |
+
"epoch": 8.72373540856031,
|
| 410 |
+
"grad_norm": 10.176831245422363,
|
| 411 |
+
"learning_rate": 1.2540913006029287e-05,
|
| 412 |
+
"loss": 0.2931,
|
| 413 |
+
"step": 1125
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 8.918287937743191,
|
| 417 |
+
"grad_norm": 31.286165237426758,
|
| 418 |
+
"learning_rate": 1.2325581395348838e-05,
|
| 419 |
+
"loss": 0.4766,
|
| 420 |
+
"step": 1150
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"epoch": 9.0,
|
| 424 |
+
"eval_accuracy": 0.9431072210065645,
|
| 425 |
+
"eval_loss": 0.28641560673713684,
|
| 426 |
+
"eval_macro_f1": 0.8947346271312796,
|
| 427 |
+
"eval_runtime": 1.0696,
|
| 428 |
+
"eval_samples_per_second": 427.28,
|
| 429 |
+
"eval_steps_per_second": 27.114,
|
| 430 |
+
"eval_weighted_f1": 0.9433439808842939,
|
| 431 |
+
"step": 1161
|
| 432 |
+
}
|
| 433 |
+
],
|
| 434 |
+
"logging_steps": 25,
|
| 435 |
+
"max_steps": 2580,
|
| 436 |
+
"num_input_tokens_seen": 0,
|
| 437 |
+
"num_train_epochs": 20,
|
| 438 |
+
"save_steps": 500,
|
| 439 |
+
"stateful_callbacks": {
|
| 440 |
+
"EarlyStoppingCallback": {
|
| 441 |
+
"args": {
|
| 442 |
+
"early_stopping_patience": 3,
|
| 443 |
+
"early_stopping_threshold": 0.0
|
| 444 |
+
},
|
| 445 |
+
"attributes": {
|
| 446 |
+
"early_stopping_patience_counter": 0
|
| 447 |
+
}
|
| 448 |
+
},
|
| 449 |
+
"TrainerControl": {
|
| 450 |
+
"args": {
|
| 451 |
+
"should_epoch_stop": false,
|
| 452 |
+
"should_evaluate": false,
|
| 453 |
+
"should_log": false,
|
| 454 |
+
"should_save": true,
|
| 455 |
+
"should_training_stop": false
|
| 456 |
+
},
|
| 457 |
+
"attributes": {}
|
| 458 |
+
}
|
| 459 |
+
},
|
| 460 |
+
"total_flos": 0.0,
|
| 461 |
+
"train_batch_size": 8,
|
| 462 |
+
"trial_name": null,
|
| 463 |
+
"trial_params": null
|
| 464 |
+
}
|
ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1161/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:794d3a3bf27619fcfed2cb297279887b95c7737bbdd16633f806566d53bd318d
|
| 3 |
+
size 5624
|
ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1161/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1548/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1548/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1548/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2a06a7c23ecc35aa6192e805f189b2095d07c9c10301d5d7cafc88b567b4eba2
|
| 3 |
+
size 540026436
|
ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1548/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:935e09f9aeec2860cd32cab2ee02294d0692621a28159ea0952a2978ac80186c
|
| 3 |
+
size 1075446330
|
ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1548/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ba03f014e2a55bc8d89e0e1f348af6952ce24bc547421052a366b8221febf0d9
|
| 3 |
+
size 14244
|
ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1548/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:44b008d352354fda2e0b318c815f01a1b6e3fc9255d05d2811f91c9fdb135eff
|
| 3 |
+
size 1064
|
ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1548/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 |
+
}
|
ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1548/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1548/trainer_state.json
ADDED
|
@@ -0,0 +1,602 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1161,
|
| 3 |
+
"best_metric": 0.8947346271312796,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1161",
|
| 5 |
+
"epoch": 12.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 1548,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.19455252918287938,
|
| 14 |
+
"grad_norm": 12.175275802612305,
|
| 15 |
+
"learning_rate": 1.86046511627907e-06,
|
| 16 |
+
"loss": 4.6741,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.38910505836575876,
|
| 21 |
+
"grad_norm": 14.172719955444336,
|
| 22 |
+
"learning_rate": 3.798449612403101e-06,
|
| 23 |
+
"loss": 3.6832,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.5836575875486382,
|
| 28 |
+
"grad_norm": 9.232404708862305,
|
| 29 |
+
"learning_rate": 5.736434108527133e-06,
|
| 30 |
+
"loss": 2.9012,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.7782101167315175,
|
| 35 |
+
"grad_norm": 9.204726219177246,
|
| 36 |
+
"learning_rate": 7.674418604651164e-06,
|
| 37 |
+
"loss": 2.6558,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.9727626459143969,
|
| 42 |
+
"grad_norm": 13.36579704284668,
|
| 43 |
+
"learning_rate": 9.612403100775196e-06,
|
| 44 |
+
"loss": 2.304,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 1.0,
|
| 49 |
+
"eval_accuracy": 0.8533916849015317,
|
| 50 |
+
"eval_loss": 0.4367648959159851,
|
| 51 |
+
"eval_macro_f1": 0.5793912034860027,
|
| 52 |
+
"eval_runtime": 1.2368,
|
| 53 |
+
"eval_samples_per_second": 369.498,
|
| 54 |
+
"eval_steps_per_second": 23.447,
|
| 55 |
+
"eval_weighted_f1": 0.8184832282800687,
|
| 56 |
+
"step": 129
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 1.1634241245136188,
|
| 60 |
+
"grad_norm": 8.186969757080078,
|
| 61 |
+
"learning_rate": 1.1550387596899227e-05,
|
| 62 |
+
"loss": 1.7317,
|
| 63 |
+
"step": 150
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 1.3579766536964981,
|
| 67 |
+
"grad_norm": 44.49769592285156,
|
| 68 |
+
"learning_rate": 1.3488372093023257e-05,
|
| 69 |
+
"loss": 1.5504,
|
| 70 |
+
"step": 175
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 1.5525291828793775,
|
| 74 |
+
"grad_norm": 15.99069595336914,
|
| 75 |
+
"learning_rate": 1.542635658914729e-05,
|
| 76 |
+
"loss": 1.49,
|
| 77 |
+
"step": 200
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"epoch": 1.7470817120622568,
|
| 81 |
+
"grad_norm": 15.231589317321777,
|
| 82 |
+
"learning_rate": 1.736434108527132e-05,
|
| 83 |
+
"loss": 1.2571,
|
| 84 |
+
"step": 225
|
| 85 |
+
},
|
| 86 |
+
{
|
| 87 |
+
"epoch": 1.9416342412451362,
|
| 88 |
+
"grad_norm": 11.705756187438965,
|
| 89 |
+
"learning_rate": 1.9302325581395353e-05,
|
| 90 |
+
"loss": 1.0896,
|
| 91 |
+
"step": 250
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"epoch": 2.0,
|
| 95 |
+
"eval_accuracy": 0.9059080962800875,
|
| 96 |
+
"eval_loss": 0.29254111647605896,
|
| 97 |
+
"eval_macro_f1": 0.7893837634028514,
|
| 98 |
+
"eval_runtime": 0.9698,
|
| 99 |
+
"eval_samples_per_second": 471.252,
|
| 100 |
+
"eval_steps_per_second": 29.904,
|
| 101 |
+
"eval_weighted_f1": 0.8965275247843335,
|
| 102 |
+
"step": 258
|
| 103 |
+
},
|
| 104 |
+
{
|
| 105 |
+
"epoch": 2.132295719844358,
|
| 106 |
+
"grad_norm": 7.289089679718018,
|
| 107 |
+
"learning_rate": 1.9862187769164516e-05,
|
| 108 |
+
"loss": 1.0811,
|
| 109 |
+
"step": 275
|
| 110 |
+
},
|
| 111 |
+
{
|
| 112 |
+
"epoch": 2.3268482490272375,
|
| 113 |
+
"grad_norm": 19.28726577758789,
|
| 114 |
+
"learning_rate": 1.9646856158484065e-05,
|
| 115 |
+
"loss": 0.938,
|
| 116 |
+
"step": 300
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 2.5214007782101167,
|
| 120 |
+
"grad_norm": 14.579724311828613,
|
| 121 |
+
"learning_rate": 1.943152454780362e-05,
|
| 122 |
+
"loss": 0.8457,
|
| 123 |
+
"step": 325
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 2.7159533073929962,
|
| 127 |
+
"grad_norm": 26.952871322631836,
|
| 128 |
+
"learning_rate": 1.921619293712317e-05,
|
| 129 |
+
"loss": 1.0485,
|
| 130 |
+
"step": 350
|
| 131 |
+
},
|
| 132 |
+
{
|
| 133 |
+
"epoch": 2.9105058365758754,
|
| 134 |
+
"grad_norm": 11.434037208557129,
|
| 135 |
+
"learning_rate": 1.9000861326442723e-05,
|
| 136 |
+
"loss": 1.1113,
|
| 137 |
+
"step": 375
|
| 138 |
+
},
|
| 139 |
+
{
|
| 140 |
+
"epoch": 3.0,
|
| 141 |
+
"eval_accuracy": 0.9146608315098468,
|
| 142 |
+
"eval_loss": 0.24419832229614258,
|
| 143 |
+
"eval_macro_f1": 0.8422144131192929,
|
| 144 |
+
"eval_runtime": 1.0063,
|
| 145 |
+
"eval_samples_per_second": 454.118,
|
| 146 |
+
"eval_steps_per_second": 28.817,
|
| 147 |
+
"eval_weighted_f1": 0.918285359370874,
|
| 148 |
+
"step": 387
|
| 149 |
+
},
|
| 150 |
+
{
|
| 151 |
+
"epoch": 3.1011673151750974,
|
| 152 |
+
"grad_norm": 23.606746673583984,
|
| 153 |
+
"learning_rate": 1.8785529715762276e-05,
|
| 154 |
+
"loss": 0.8873,
|
| 155 |
+
"step": 400
|
| 156 |
+
},
|
| 157 |
+
{
|
| 158 |
+
"epoch": 3.2957198443579765,
|
| 159 |
+
"grad_norm": 13.393070220947266,
|
| 160 |
+
"learning_rate": 1.857019810508183e-05,
|
| 161 |
+
"loss": 0.6112,
|
| 162 |
+
"step": 425
|
| 163 |
+
},
|
| 164 |
+
{
|
| 165 |
+
"epoch": 3.490272373540856,
|
| 166 |
+
"grad_norm": 2.7602410316467285,
|
| 167 |
+
"learning_rate": 1.835486649440138e-05,
|
| 168 |
+
"loss": 0.7564,
|
| 169 |
+
"step": 450
|
| 170 |
+
},
|
| 171 |
+
{
|
| 172 |
+
"epoch": 3.6848249027237356,
|
| 173 |
+
"grad_norm": 38.35893630981445,
|
| 174 |
+
"learning_rate": 1.813953488372093e-05,
|
| 175 |
+
"loss": 0.8032,
|
| 176 |
+
"step": 475
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 3.8793774319066148,
|
| 180 |
+
"grad_norm": 26.61602783203125,
|
| 181 |
+
"learning_rate": 1.7924203273040483e-05,
|
| 182 |
+
"loss": 0.9455,
|
| 183 |
+
"step": 500
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 4.0,
|
| 187 |
+
"eval_accuracy": 0.925601750547046,
|
| 188 |
+
"eval_loss": 0.24463064968585968,
|
| 189 |
+
"eval_macro_f1": 0.8577216356628122,
|
| 190 |
+
"eval_runtime": 1.6336,
|
| 191 |
+
"eval_samples_per_second": 279.754,
|
| 192 |
+
"eval_steps_per_second": 17.752,
|
| 193 |
+
"eval_weighted_f1": 0.9254741614987465,
|
| 194 |
+
"step": 516
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 4.070038910505836,
|
| 198 |
+
"grad_norm": 39.84136199951172,
|
| 199 |
+
"learning_rate": 1.7708871662360035e-05,
|
| 200 |
+
"loss": 0.6896,
|
| 201 |
+
"step": 525
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 4.264591439688716,
|
| 205 |
+
"grad_norm": 1.3257018327713013,
|
| 206 |
+
"learning_rate": 1.7493540051679588e-05,
|
| 207 |
+
"loss": 0.5164,
|
| 208 |
+
"step": 550
|
| 209 |
+
},
|
| 210 |
+
{
|
| 211 |
+
"epoch": 4.4591439688715955,
|
| 212 |
+
"grad_norm": 4.106621265411377,
|
| 213 |
+
"learning_rate": 1.727820844099914e-05,
|
| 214 |
+
"loss": 0.4498,
|
| 215 |
+
"step": 575
|
| 216 |
+
},
|
| 217 |
+
{
|
| 218 |
+
"epoch": 4.653696498054475,
|
| 219 |
+
"grad_norm": 27.42701530456543,
|
| 220 |
+
"learning_rate": 1.7062876830318693e-05,
|
| 221 |
+
"loss": 0.7688,
|
| 222 |
+
"step": 600
|
| 223 |
+
},
|
| 224 |
+
{
|
| 225 |
+
"epoch": 4.848249027237354,
|
| 226 |
+
"grad_norm": 8.981760025024414,
|
| 227 |
+
"learning_rate": 1.6847545219638243e-05,
|
| 228 |
+
"loss": 0.6506,
|
| 229 |
+
"step": 625
|
| 230 |
+
},
|
| 231 |
+
{
|
| 232 |
+
"epoch": 5.0,
|
| 233 |
+
"eval_accuracy": 0.9277899343544858,
|
| 234 |
+
"eval_loss": 0.28580012917518616,
|
| 235 |
+
"eval_macro_f1": 0.870879330908566,
|
| 236 |
+
"eval_runtime": 1.0274,
|
| 237 |
+
"eval_samples_per_second": 444.802,
|
| 238 |
+
"eval_steps_per_second": 28.226,
|
| 239 |
+
"eval_weighted_f1": 0.9302773652022999,
|
| 240 |
+
"step": 645
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 5.038910505836576,
|
| 244 |
+
"grad_norm": 46.58235549926758,
|
| 245 |
+
"learning_rate": 1.6632213608957795e-05,
|
| 246 |
+
"loss": 0.7474,
|
| 247 |
+
"step": 650
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 5.233463035019455,
|
| 251 |
+
"grad_norm": 2.2681214809417725,
|
| 252 |
+
"learning_rate": 1.6416881998277348e-05,
|
| 253 |
+
"loss": 0.4227,
|
| 254 |
+
"step": 675
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 5.428015564202335,
|
| 258 |
+
"grad_norm": 49.91563415527344,
|
| 259 |
+
"learning_rate": 1.62015503875969e-05,
|
| 260 |
+
"loss": 0.5143,
|
| 261 |
+
"step": 700
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 5.622568093385214,
|
| 265 |
+
"grad_norm": 41.00137710571289,
|
| 266 |
+
"learning_rate": 1.5986218776916453e-05,
|
| 267 |
+
"loss": 0.4182,
|
| 268 |
+
"step": 725
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 5.817120622568093,
|
| 272 |
+
"grad_norm": 9.387858390808105,
|
| 273 |
+
"learning_rate": 1.5770887166236006e-05,
|
| 274 |
+
"loss": 0.4013,
|
| 275 |
+
"step": 750
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 6.0,
|
| 279 |
+
"eval_accuracy": 0.9037199124726477,
|
| 280 |
+
"eval_loss": 0.3829730451107025,
|
| 281 |
+
"eval_macro_f1": 0.8311931591145209,
|
| 282 |
+
"eval_runtime": 0.9951,
|
| 283 |
+
"eval_samples_per_second": 459.272,
|
| 284 |
+
"eval_steps_per_second": 29.144,
|
| 285 |
+
"eval_weighted_f1": 0.9057689254816129,
|
| 286 |
+
"step": 774
|
| 287 |
+
},
|
| 288 |
+
{
|
| 289 |
+
"epoch": 6.007782101167315,
|
| 290 |
+
"grad_norm": 44.55717468261719,
|
| 291 |
+
"learning_rate": 1.555555555555556e-05,
|
| 292 |
+
"loss": 0.6504,
|
| 293 |
+
"step": 775
|
| 294 |
+
},
|
| 295 |
+
{
|
| 296 |
+
"epoch": 6.202334630350195,
|
| 297 |
+
"grad_norm": 94.3865966796875,
|
| 298 |
+
"learning_rate": 1.534022394487511e-05,
|
| 299 |
+
"loss": 0.412,
|
| 300 |
+
"step": 800
|
| 301 |
+
},
|
| 302 |
+
{
|
| 303 |
+
"epoch": 6.396887159533074,
|
| 304 |
+
"grad_norm": 9.678958892822266,
|
| 305 |
+
"learning_rate": 1.512489233419466e-05,
|
| 306 |
+
"loss": 0.4283,
|
| 307 |
+
"step": 825
|
| 308 |
+
},
|
| 309 |
+
{
|
| 310 |
+
"epoch": 6.591439688715953,
|
| 311 |
+
"grad_norm": 37.90494918823242,
|
| 312 |
+
"learning_rate": 1.4909560723514215e-05,
|
| 313 |
+
"loss": 0.4208,
|
| 314 |
+
"step": 850
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 6.785992217898833,
|
| 318 |
+
"grad_norm": 21.357500076293945,
|
| 319 |
+
"learning_rate": 1.4694229112833764e-05,
|
| 320 |
+
"loss": 0.5355,
|
| 321 |
+
"step": 875
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 6.980544747081712,
|
| 325 |
+
"grad_norm": 52.673919677734375,
|
| 326 |
+
"learning_rate": 1.4478897502153318e-05,
|
| 327 |
+
"loss": 0.3123,
|
| 328 |
+
"step": 900
|
| 329 |
+
},
|
| 330 |
+
{
|
| 331 |
+
"epoch": 7.0,
|
| 332 |
+
"eval_accuracy": 0.9409190371991247,
|
| 333 |
+
"eval_loss": 0.2741667926311493,
|
| 334 |
+
"eval_macro_f1": 0.8910597249801194,
|
| 335 |
+
"eval_runtime": 1.3714,
|
| 336 |
+
"eval_samples_per_second": 333.247,
|
| 337 |
+
"eval_steps_per_second": 21.147,
|
| 338 |
+
"eval_weighted_f1": 0.9413592377918585,
|
| 339 |
+
"step": 903
|
| 340 |
+
},
|
| 341 |
+
{
|
| 342 |
+
"epoch": 7.171206225680934,
|
| 343 |
+
"grad_norm": 5.324210166931152,
|
| 344 |
+
"learning_rate": 1.426356589147287e-05,
|
| 345 |
+
"loss": 0.3897,
|
| 346 |
+
"step": 925
|
| 347 |
+
},
|
| 348 |
+
{
|
| 349 |
+
"epoch": 7.365758754863813,
|
| 350 |
+
"grad_norm": 21.306001663208008,
|
| 351 |
+
"learning_rate": 1.4048234280792422e-05,
|
| 352 |
+
"loss": 0.3909,
|
| 353 |
+
"step": 950
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 7.560311284046692,
|
| 357 |
+
"grad_norm": 1.1874115467071533,
|
| 358 |
+
"learning_rate": 1.3832902670111974e-05,
|
| 359 |
+
"loss": 0.4574,
|
| 360 |
+
"step": 975
|
| 361 |
+
},
|
| 362 |
+
{
|
| 363 |
+
"epoch": 7.754863813229572,
|
| 364 |
+
"grad_norm": 0.7569160461425781,
|
| 365 |
+
"learning_rate": 1.3617571059431525e-05,
|
| 366 |
+
"loss": 0.2395,
|
| 367 |
+
"step": 1000
|
| 368 |
+
},
|
| 369 |
+
{
|
| 370 |
+
"epoch": 7.9494163424124515,
|
| 371 |
+
"grad_norm": 0.5100299715995789,
|
| 372 |
+
"learning_rate": 1.3402239448751078e-05,
|
| 373 |
+
"loss": 0.2427,
|
| 374 |
+
"step": 1025
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 8.0,
|
| 378 |
+
"eval_accuracy": 0.9299781181619255,
|
| 379 |
+
"eval_loss": 0.3452863097190857,
|
| 380 |
+
"eval_macro_f1": 0.8730694907605123,
|
| 381 |
+
"eval_runtime": 0.957,
|
| 382 |
+
"eval_samples_per_second": 477.558,
|
| 383 |
+
"eval_steps_per_second": 30.305,
|
| 384 |
+
"eval_weighted_f1": 0.9314037441796743,
|
| 385 |
+
"step": 1032
|
| 386 |
+
},
|
| 387 |
+
{
|
| 388 |
+
"epoch": 8.140077821011673,
|
| 389 |
+
"grad_norm": 39.542598724365234,
|
| 390 |
+
"learning_rate": 1.318690783807063e-05,
|
| 391 |
+
"loss": 0.313,
|
| 392 |
+
"step": 1050
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 8.334630350194553,
|
| 396 |
+
"grad_norm": 1.1928895711898804,
|
| 397 |
+
"learning_rate": 1.2971576227390181e-05,
|
| 398 |
+
"loss": 0.2376,
|
| 399 |
+
"step": 1075
|
| 400 |
+
},
|
| 401 |
+
{
|
| 402 |
+
"epoch": 8.529182879377432,
|
| 403 |
+
"grad_norm": 10.282620429992676,
|
| 404 |
+
"learning_rate": 1.2756244616709734e-05,
|
| 405 |
+
"loss": 0.1612,
|
| 406 |
+
"step": 1100
|
| 407 |
+
},
|
| 408 |
+
{
|
| 409 |
+
"epoch": 8.72373540856031,
|
| 410 |
+
"grad_norm": 10.176831245422363,
|
| 411 |
+
"learning_rate": 1.2540913006029287e-05,
|
| 412 |
+
"loss": 0.2931,
|
| 413 |
+
"step": 1125
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 8.918287937743191,
|
| 417 |
+
"grad_norm": 31.286165237426758,
|
| 418 |
+
"learning_rate": 1.2325581395348838e-05,
|
| 419 |
+
"loss": 0.4766,
|
| 420 |
+
"step": 1150
|
| 421 |
+
},
|
| 422 |
+
{
|
| 423 |
+
"epoch": 9.0,
|
| 424 |
+
"eval_accuracy": 0.9431072210065645,
|
| 425 |
+
"eval_loss": 0.28641560673713684,
|
| 426 |
+
"eval_macro_f1": 0.8947346271312796,
|
| 427 |
+
"eval_runtime": 1.0696,
|
| 428 |
+
"eval_samples_per_second": 427.28,
|
| 429 |
+
"eval_steps_per_second": 27.114,
|
| 430 |
+
"eval_weighted_f1": 0.9433439808842939,
|
| 431 |
+
"step": 1161
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 9.108949416342412,
|
| 435 |
+
"grad_norm": 1.6606191396713257,
|
| 436 |
+
"learning_rate": 1.211024978466839e-05,
|
| 437 |
+
"loss": 0.2784,
|
| 438 |
+
"step": 1175
|
| 439 |
+
},
|
| 440 |
+
{
|
| 441 |
+
"epoch": 9.303501945525293,
|
| 442 |
+
"grad_norm": 3.397531747817993,
|
| 443 |
+
"learning_rate": 1.1894918173987941e-05,
|
| 444 |
+
"loss": 0.2861,
|
| 445 |
+
"step": 1200
|
| 446 |
+
},
|
| 447 |
+
{
|
| 448 |
+
"epoch": 9.498054474708171,
|
| 449 |
+
"grad_norm": 25.98397445678711,
|
| 450 |
+
"learning_rate": 1.1679586563307494e-05,
|
| 451 |
+
"loss": 0.2072,
|
| 452 |
+
"step": 1225
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 9.69260700389105,
|
| 456 |
+
"grad_norm": 0.2903095483779907,
|
| 457 |
+
"learning_rate": 1.1464254952627048e-05,
|
| 458 |
+
"loss": 0.1014,
|
| 459 |
+
"step": 1250
|
| 460 |
+
},
|
| 461 |
+
{
|
| 462 |
+
"epoch": 9.88715953307393,
|
| 463 |
+
"grad_norm": 18.12164306640625,
|
| 464 |
+
"learning_rate": 1.1248923341946599e-05,
|
| 465 |
+
"loss": 0.2802,
|
| 466 |
+
"step": 1275
|
| 467 |
+
},
|
| 468 |
+
{
|
| 469 |
+
"epoch": 10.0,
|
| 470 |
+
"eval_accuracy": 0.9321663019693655,
|
| 471 |
+
"eval_loss": 0.370513916015625,
|
| 472 |
+
"eval_macro_f1": 0.876772197524931,
|
| 473 |
+
"eval_runtime": 0.9543,
|
| 474 |
+
"eval_samples_per_second": 478.866,
|
| 475 |
+
"eval_steps_per_second": 30.388,
|
| 476 |
+
"eval_weighted_f1": 0.9330585133442931,
|
| 477 |
+
"step": 1290
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 10.077821011673151,
|
| 481 |
+
"grad_norm": 0.9774921536445618,
|
| 482 |
+
"learning_rate": 1.1033591731266152e-05,
|
| 483 |
+
"loss": 0.1059,
|
| 484 |
+
"step": 1300
|
| 485 |
+
},
|
| 486 |
+
{
|
| 487 |
+
"epoch": 10.272373540856032,
|
| 488 |
+
"grad_norm": 1.1316591501235962,
|
| 489 |
+
"learning_rate": 1.0818260120585703e-05,
|
| 490 |
+
"loss": 0.1125,
|
| 491 |
+
"step": 1325
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 10.46692607003891,
|
| 495 |
+
"grad_norm": 44.349830627441406,
|
| 496 |
+
"learning_rate": 1.0602928509905255e-05,
|
| 497 |
+
"loss": 0.162,
|
| 498 |
+
"step": 1350
|
| 499 |
+
},
|
| 500 |
+
{
|
| 501 |
+
"epoch": 10.66147859922179,
|
| 502 |
+
"grad_norm": 5.914244174957275,
|
| 503 |
+
"learning_rate": 1.0387596899224808e-05,
|
| 504 |
+
"loss": 0.1907,
|
| 505 |
+
"step": 1375
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 10.85603112840467,
|
| 509 |
+
"grad_norm": 5.745565414428711,
|
| 510 |
+
"learning_rate": 1.0172265288544359e-05,
|
| 511 |
+
"loss": 0.1172,
|
| 512 |
+
"step": 1400
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 11.0,
|
| 516 |
+
"eval_accuracy": 0.8971553610503282,
|
| 517 |
+
"eval_loss": 0.5507068037986755,
|
| 518 |
+
"eval_macro_f1": 0.8354016092815985,
|
| 519 |
+
"eval_runtime": 1.0035,
|
| 520 |
+
"eval_samples_per_second": 455.387,
|
| 521 |
+
"eval_steps_per_second": 28.898,
|
| 522 |
+
"eval_weighted_f1": 0.9038802372596354,
|
| 523 |
+
"step": 1419
|
| 524 |
+
},
|
| 525 |
+
{
|
| 526 |
+
"epoch": 11.04669260700389,
|
| 527 |
+
"grad_norm": 2.2701375484466553,
|
| 528 |
+
"learning_rate": 9.956933677863911e-06,
|
| 529 |
+
"loss": 0.3909,
|
| 530 |
+
"step": 1425
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 11.24124513618677,
|
| 534 |
+
"grad_norm": 25.465295791625977,
|
| 535 |
+
"learning_rate": 9.741602067183464e-06,
|
| 536 |
+
"loss": 0.1124,
|
| 537 |
+
"step": 1450
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 11.43579766536965,
|
| 541 |
+
"grad_norm": 57.656097412109375,
|
| 542 |
+
"learning_rate": 9.526270456503015e-06,
|
| 543 |
+
"loss": 0.1804,
|
| 544 |
+
"step": 1475
|
| 545 |
+
},
|
| 546 |
+
{
|
| 547 |
+
"epoch": 11.630350194552529,
|
| 548 |
+
"grad_norm": 1.8088877201080322,
|
| 549 |
+
"learning_rate": 9.310938845822568e-06,
|
| 550 |
+
"loss": 0.1729,
|
| 551 |
+
"step": 1500
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 11.82490272373541,
|
| 555 |
+
"grad_norm": 5.960810661315918,
|
| 556 |
+
"learning_rate": 9.09560723514212e-06,
|
| 557 |
+
"loss": 0.1704,
|
| 558 |
+
"step": 1525
|
| 559 |
+
},
|
| 560 |
+
{
|
| 561 |
+
"epoch": 12.0,
|
| 562 |
+
"eval_accuracy": 0.9277899343544858,
|
| 563 |
+
"eval_loss": 0.38572177290916443,
|
| 564 |
+
"eval_macro_f1": 0.8734584557574303,
|
| 565 |
+
"eval_runtime": 1.0015,
|
| 566 |
+
"eval_samples_per_second": 456.297,
|
| 567 |
+
"eval_steps_per_second": 28.955,
|
| 568 |
+
"eval_weighted_f1": 0.9297279013051296,
|
| 569 |
+
"step": 1548
|
| 570 |
+
}
|
| 571 |
+
],
|
| 572 |
+
"logging_steps": 25,
|
| 573 |
+
"max_steps": 2580,
|
| 574 |
+
"num_input_tokens_seen": 0,
|
| 575 |
+
"num_train_epochs": 20,
|
| 576 |
+
"save_steps": 500,
|
| 577 |
+
"stateful_callbacks": {
|
| 578 |
+
"EarlyStoppingCallback": {
|
| 579 |
+
"args": {
|
| 580 |
+
"early_stopping_patience": 3,
|
| 581 |
+
"early_stopping_threshold": 0.0
|
| 582 |
+
},
|
| 583 |
+
"attributes": {
|
| 584 |
+
"early_stopping_patience_counter": 3
|
| 585 |
+
}
|
| 586 |
+
},
|
| 587 |
+
"TrainerControl": {
|
| 588 |
+
"args": {
|
| 589 |
+
"should_epoch_stop": false,
|
| 590 |
+
"should_evaluate": false,
|
| 591 |
+
"should_log": false,
|
| 592 |
+
"should_save": true,
|
| 593 |
+
"should_training_stop": true
|
| 594 |
+
},
|
| 595 |
+
"attributes": {}
|
| 596 |
+
}
|
| 597 |
+
},
|
| 598 |
+
"total_flos": 0.0,
|
| 599 |
+
"train_batch_size": 8,
|
| 600 |
+
"trial_name": null,
|
| 601 |
+
"trial_params": null
|
| 602 |
+
}
|
ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1548/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:794d3a3bf27619fcfed2cb297279887b95c7737bbdd16633f806566d53bd318d
|
| 3 |
+
size 5624
|
ood-uit_vsfc-phobert-topic-domain-01-s44-20260812T034649Z-6beb231d/trainer/checkpoint-1548/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ecba2c64fabdae604e61692714379e800e733a4ce6c1b885ee005b3be1d8b4fc
|
| 3 |
+
size 540026404
|
ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-2592/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-2592/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-2592/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cfab79fa256e61be49bf93dec1132df787f9d6b4e75b36fc193d4aee77754c17
|
| 3 |
+
size 540026436
|
ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-2592/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d157ddee5dd5c2b3745a6f61f13e90cb6c66937e025d55dcf46b90adbeab941f
|
| 3 |
+
size 1075446330
|
ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-2592/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cdc2b3aaedf7ea71a37d28d3a6387093cb6628fc5acfdeb6fb8512157a60483b
|
| 3 |
+
size 14244
|
ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-2592/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4fc7a4447ea930f07fbd416c15be11186b453b4bef034c6bb6ee4e199c0f2f06
|
| 3 |
+
size 1064
|
ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-2592/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 |
+
}
|
ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-2592/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-2592/trainer_state.json
ADDED
|
@@ -0,0 +1,830 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 2592,
|
| 3 |
+
"best_metric": 0.8739757580611789,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-2592",
|
| 5 |
+
"epoch": 6.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 2592,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.05787037037037037,
|
| 14 |
+
"grad_norm": 18.34950828552246,
|
| 15 |
+
"learning_rate": 5.555555555555555e-07,
|
| 16 |
+
"loss": 4.7892,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.11574074074074074,
|
| 21 |
+
"grad_norm": 17.702529907226562,
|
| 22 |
+
"learning_rate": 1.1342592592592594e-06,
|
| 23 |
+
"loss": 4.5178,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.1736111111111111,
|
| 28 |
+
"grad_norm": 12.634126663208008,
|
| 29 |
+
"learning_rate": 1.7129629629629632e-06,
|
| 30 |
+
"loss": 4.01,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.23148148148148148,
|
| 35 |
+
"grad_norm": 11.140002250671387,
|
| 36 |
+
"learning_rate": 2.2916666666666666e-06,
|
| 37 |
+
"loss": 3.4342,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.28935185185185186,
|
| 42 |
+
"grad_norm": 10.938908576965332,
|
| 43 |
+
"learning_rate": 2.8703703703703706e-06,
|
| 44 |
+
"loss": 2.9665,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.3472222222222222,
|
| 49 |
+
"grad_norm": 10.947237968444824,
|
| 50 |
+
"learning_rate": 3.449074074074074e-06,
|
| 51 |
+
"loss": 2.4821,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.4050925925925926,
|
| 56 |
+
"grad_norm": 16.171165466308594,
|
| 57 |
+
"learning_rate": 4.027777777777779e-06,
|
| 58 |
+
"loss": 1.9626,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.46296296296296297,
|
| 63 |
+
"grad_norm": 15.498010635375977,
|
| 64 |
+
"learning_rate": 4.606481481481481e-06,
|
| 65 |
+
"loss": 1.4921,
|
| 66 |
+
"step": 200
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 0.5208333333333334,
|
| 70 |
+
"grad_norm": 20.90071678161621,
|
| 71 |
+
"learning_rate": 5.185185185185185e-06,
|
| 72 |
+
"loss": 1.4152,
|
| 73 |
+
"step": 225
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.5787037037037037,
|
| 77 |
+
"grad_norm": 14.971171379089355,
|
| 78 |
+
"learning_rate": 5.7638888888888886e-06,
|
| 79 |
+
"loss": 1.1941,
|
| 80 |
+
"step": 250
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.6365740740740741,
|
| 84 |
+
"grad_norm": 15.980868339538574,
|
| 85 |
+
"learning_rate": 6.342592592592594e-06,
|
| 86 |
+
"loss": 1.0189,
|
| 87 |
+
"step": 275
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.6944444444444444,
|
| 91 |
+
"grad_norm": 5.475146770477295,
|
| 92 |
+
"learning_rate": 6.9212962962962974e-06,
|
| 93 |
+
"loss": 1.0733,
|
| 94 |
+
"step": 300
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.7523148148148148,
|
| 98 |
+
"grad_norm": 19.17060661315918,
|
| 99 |
+
"learning_rate": 7.500000000000001e-06,
|
| 100 |
+
"loss": 1.0232,
|
| 101 |
+
"step": 325
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 0.8101851851851852,
|
| 105 |
+
"grad_norm": 30.633872985839844,
|
| 106 |
+
"learning_rate": 8.078703703703705e-06,
|
| 107 |
+
"loss": 0.9152,
|
| 108 |
+
"step": 350
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 0.8680555555555556,
|
| 112 |
+
"grad_norm": 14.854605674743652,
|
| 113 |
+
"learning_rate": 8.657407407407408e-06,
|
| 114 |
+
"loss": 0.8892,
|
| 115 |
+
"step": 375
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.9259259259259259,
|
| 119 |
+
"grad_norm": 15.524602890014648,
|
| 120 |
+
"learning_rate": 9.236111111111112e-06,
|
| 121 |
+
"loss": 1.2148,
|
| 122 |
+
"step": 400
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.9837962962962963,
|
| 126 |
+
"grad_norm": 8.35659408569336,
|
| 127 |
+
"learning_rate": 9.814814814814815e-06,
|
| 128 |
+
"loss": 1.0522,
|
| 129 |
+
"step": 425
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 1.0,
|
| 133 |
+
"eval_accuracy": 0.9427083333333334,
|
| 134 |
+
"eval_loss": 0.19718892872333527,
|
| 135 |
+
"eval_macro_f1": 0.6518213119914867,
|
| 136 |
+
"eval_runtime": 3.6776,
|
| 137 |
+
"eval_samples_per_second": 417.662,
|
| 138 |
+
"eval_steps_per_second": 26.104,
|
| 139 |
+
"eval_weighted_f1": 0.9292832259638407,
|
| 140 |
+
"step": 432
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 1.0416666666666667,
|
| 144 |
+
"grad_norm": 9.545726776123047,
|
| 145 |
+
"learning_rate": 1.0393518518518519e-05,
|
| 146 |
+
"loss": 0.9327,
|
| 147 |
+
"step": 450
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 1.099537037037037,
|
| 151 |
+
"grad_norm": 28.5867977142334,
|
| 152 |
+
"learning_rate": 1.0972222222222224e-05,
|
| 153 |
+
"loss": 0.8578,
|
| 154 |
+
"step": 475
|
| 155 |
+
},
|
| 156 |
+
{
|
| 157 |
+
"epoch": 1.1574074074074074,
|
| 158 |
+
"grad_norm": 26.383472442626953,
|
| 159 |
+
"learning_rate": 1.1550925925925926e-05,
|
| 160 |
+
"loss": 0.615,
|
| 161 |
+
"step": 500
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"epoch": 1.2152777777777777,
|
| 165 |
+
"grad_norm": 7.6479668617248535,
|
| 166 |
+
"learning_rate": 1.2129629629629631e-05,
|
| 167 |
+
"loss": 0.8475,
|
| 168 |
+
"step": 525
|
| 169 |
+
},
|
| 170 |
+
{
|
| 171 |
+
"epoch": 1.2731481481481481,
|
| 172 |
+
"grad_norm": 9.217597961425781,
|
| 173 |
+
"learning_rate": 1.2708333333333333e-05,
|
| 174 |
+
"loss": 0.7591,
|
| 175 |
+
"step": 550
|
| 176 |
+
},
|
| 177 |
+
{
|
| 178 |
+
"epoch": 1.3310185185185186,
|
| 179 |
+
"grad_norm": 9.50477123260498,
|
| 180 |
+
"learning_rate": 1.3287037037037039e-05,
|
| 181 |
+
"loss": 0.8333,
|
| 182 |
+
"step": 575
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 1.3888888888888888,
|
| 186 |
+
"grad_norm": 282.826416015625,
|
| 187 |
+
"learning_rate": 1.3865740740740742e-05,
|
| 188 |
+
"loss": 0.6888,
|
| 189 |
+
"step": 600
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 1.4467592592592593,
|
| 193 |
+
"grad_norm": 8.852691650390625,
|
| 194 |
+
"learning_rate": 1.4444444444444446e-05,
|
| 195 |
+
"loss": 0.969,
|
| 196 |
+
"step": 625
|
| 197 |
+
},
|
| 198 |
+
{
|
| 199 |
+
"epoch": 1.5046296296296298,
|
| 200 |
+
"grad_norm": 10.682754516601562,
|
| 201 |
+
"learning_rate": 1.502314814814815e-05,
|
| 202 |
+
"loss": 0.7574,
|
| 203 |
+
"step": 650
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 1.5625,
|
| 207 |
+
"grad_norm": 8.411812782287598,
|
| 208 |
+
"learning_rate": 1.5601851851851853e-05,
|
| 209 |
+
"loss": 0.8515,
|
| 210 |
+
"step": 675
|
| 211 |
+
},
|
| 212 |
+
{
|
| 213 |
+
"epoch": 1.6203703703703702,
|
| 214 |
+
"grad_norm": 4.259770393371582,
|
| 215 |
+
"learning_rate": 1.6180555555555557e-05,
|
| 216 |
+
"loss": 0.7792,
|
| 217 |
+
"step": 700
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"epoch": 1.6782407407407407,
|
| 221 |
+
"grad_norm": 43.64582061767578,
|
| 222 |
+
"learning_rate": 1.675925925925926e-05,
|
| 223 |
+
"loss": 0.7105,
|
| 224 |
+
"step": 725
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 1.7361111111111112,
|
| 228 |
+
"grad_norm": 12.682132720947266,
|
| 229 |
+
"learning_rate": 1.7337962962962964e-05,
|
| 230 |
+
"loss": 0.8808,
|
| 231 |
+
"step": 750
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 1.7939814814814814,
|
| 235 |
+
"grad_norm": 0.8860751986503601,
|
| 236 |
+
"learning_rate": 1.7916666666666667e-05,
|
| 237 |
+
"loss": 0.6108,
|
| 238 |
+
"step": 775
|
| 239 |
+
},
|
| 240 |
+
{
|
| 241 |
+
"epoch": 1.8518518518518519,
|
| 242 |
+
"grad_norm": 49.53169250488281,
|
| 243 |
+
"learning_rate": 1.849537037037037e-05,
|
| 244 |
+
"loss": 0.8255,
|
| 245 |
+
"step": 800
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 1.9097222222222223,
|
| 249 |
+
"grad_norm": 29.884153366088867,
|
| 250 |
+
"learning_rate": 1.9074074074074075e-05,
|
| 251 |
+
"loss": 1.0157,
|
| 252 |
+
"step": 825
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 1.9675925925925926,
|
| 256 |
+
"grad_norm": 10.482592582702637,
|
| 257 |
+
"learning_rate": 1.9652777777777778e-05,
|
| 258 |
+
"loss": 0.6026,
|
| 259 |
+
"step": 850
|
| 260 |
+
},
|
| 261 |
+
{
|
| 262 |
+
"epoch": 2.0,
|
| 263 |
+
"eval_accuracy": 0.951171875,
|
| 264 |
+
"eval_loss": 0.1825077086687088,
|
| 265 |
+
"eval_macro_f1": 0.8026620268618748,
|
| 266 |
+
"eval_runtime": 2.5618,
|
| 267 |
+
"eval_samples_per_second": 599.589,
|
| 268 |
+
"eval_steps_per_second": 37.474,
|
| 269 |
+
"eval_weighted_f1": 0.948829490199814,
|
| 270 |
+
"step": 864
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 2.025462962962963,
|
| 274 |
+
"grad_norm": 31.20493507385254,
|
| 275 |
+
"learning_rate": 1.997427983539095e-05,
|
| 276 |
+
"loss": 0.6416,
|
| 277 |
+
"step": 875
|
| 278 |
+
},
|
| 279 |
+
{
|
| 280 |
+
"epoch": 2.0833333333333335,
|
| 281 |
+
"grad_norm": 7.576644420623779,
|
| 282 |
+
"learning_rate": 1.9909979423868314e-05,
|
| 283 |
+
"loss": 0.7873,
|
| 284 |
+
"step": 900
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 2.1412037037037037,
|
| 288 |
+
"grad_norm": 5.631798267364502,
|
| 289 |
+
"learning_rate": 1.984567901234568e-05,
|
| 290 |
+
"loss": 0.7779,
|
| 291 |
+
"step": 925
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 2.199074074074074,
|
| 295 |
+
"grad_norm": 19.126354217529297,
|
| 296 |
+
"learning_rate": 1.9781378600823048e-05,
|
| 297 |
+
"loss": 0.5556,
|
| 298 |
+
"step": 950
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 2.2569444444444446,
|
| 302 |
+
"grad_norm": 5.889681339263916,
|
| 303 |
+
"learning_rate": 1.9717078189300413e-05,
|
| 304 |
+
"loss": 0.5082,
|
| 305 |
+
"step": 975
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 2.314814814814815,
|
| 309 |
+
"grad_norm": 86.90234375,
|
| 310 |
+
"learning_rate": 1.9652777777777778e-05,
|
| 311 |
+
"loss": 0.8321,
|
| 312 |
+
"step": 1000
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 2.372685185185185,
|
| 316 |
+
"grad_norm": 37.04526138305664,
|
| 317 |
+
"learning_rate": 1.9588477366255147e-05,
|
| 318 |
+
"loss": 0.6905,
|
| 319 |
+
"step": 1025
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 2.4305555555555554,
|
| 323 |
+
"grad_norm": 26.49477195739746,
|
| 324 |
+
"learning_rate": 1.9524176954732512e-05,
|
| 325 |
+
"loss": 0.6623,
|
| 326 |
+
"step": 1050
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 2.488425925925926,
|
| 330 |
+
"grad_norm": 61.23967361450195,
|
| 331 |
+
"learning_rate": 1.945987654320988e-05,
|
| 332 |
+
"loss": 0.7448,
|
| 333 |
+
"step": 1075
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 2.5462962962962963,
|
| 337 |
+
"grad_norm": 16.973339080810547,
|
| 338 |
+
"learning_rate": 1.9395576131687242e-05,
|
| 339 |
+
"loss": 0.4993,
|
| 340 |
+
"step": 1100
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 2.6041666666666665,
|
| 344 |
+
"grad_norm": 20.362926483154297,
|
| 345 |
+
"learning_rate": 1.933127572016461e-05,
|
| 346 |
+
"loss": 0.6888,
|
| 347 |
+
"step": 1125
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 2.662037037037037,
|
| 351 |
+
"grad_norm": 2.426138401031494,
|
| 352 |
+
"learning_rate": 1.926697530864198e-05,
|
| 353 |
+
"loss": 0.7204,
|
| 354 |
+
"step": 1150
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 2.7199074074074074,
|
| 358 |
+
"grad_norm": 38.947898864746094,
|
| 359 |
+
"learning_rate": 1.920267489711934e-05,
|
| 360 |
+
"loss": 0.8354,
|
| 361 |
+
"step": 1175
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 2.7777777777777777,
|
| 365 |
+
"grad_norm": 29.757408142089844,
|
| 366 |
+
"learning_rate": 1.913837448559671e-05,
|
| 367 |
+
"loss": 0.735,
|
| 368 |
+
"step": 1200
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 2.835648148148148,
|
| 372 |
+
"grad_norm": 4.250671863555908,
|
| 373 |
+
"learning_rate": 1.9074074074074075e-05,
|
| 374 |
+
"loss": 0.5405,
|
| 375 |
+
"step": 1225
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 2.8935185185185186,
|
| 379 |
+
"grad_norm": 4.031290531158447,
|
| 380 |
+
"learning_rate": 1.9009773662551443e-05,
|
| 381 |
+
"loss": 0.7388,
|
| 382 |
+
"step": 1250
|
| 383 |
+
},
|
| 384 |
+
{
|
| 385 |
+
"epoch": 2.951388888888889,
|
| 386 |
+
"grad_norm": 11.694859504699707,
|
| 387 |
+
"learning_rate": 1.8945473251028808e-05,
|
| 388 |
+
"loss": 0.6715,
|
| 389 |
+
"step": 1275
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 3.0,
|
| 393 |
+
"eval_accuracy": 0.9498697916666666,
|
| 394 |
+
"eval_loss": 0.18208689987659454,
|
| 395 |
+
"eval_macro_f1": 0.8273221081884374,
|
| 396 |
+
"eval_runtime": 3.7353,
|
| 397 |
+
"eval_samples_per_second": 411.214,
|
| 398 |
+
"eval_steps_per_second": 25.701,
|
| 399 |
+
"eval_weighted_f1": 0.9502826564284111,
|
| 400 |
+
"step": 1296
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 3.009259259259259,
|
| 404 |
+
"grad_norm": 3.5965843200683594,
|
| 405 |
+
"learning_rate": 1.8881172839506173e-05,
|
| 406 |
+
"loss": 0.6595,
|
| 407 |
+
"step": 1300
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 3.0671296296296298,
|
| 411 |
+
"grad_norm": 23.776273727416992,
|
| 412 |
+
"learning_rate": 1.8816872427983542e-05,
|
| 413 |
+
"loss": 0.3789,
|
| 414 |
+
"step": 1325
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 3.125,
|
| 418 |
+
"grad_norm": 1.3928412199020386,
|
| 419 |
+
"learning_rate": 1.8752572016460907e-05,
|
| 420 |
+
"loss": 0.7149,
|
| 421 |
+
"step": 1350
|
| 422 |
+
},
|
| 423 |
+
{
|
| 424 |
+
"epoch": 3.1828703703703702,
|
| 425 |
+
"grad_norm": 0.39887893199920654,
|
| 426 |
+
"learning_rate": 1.8688271604938272e-05,
|
| 427 |
+
"loss": 0.4188,
|
| 428 |
+
"step": 1375
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 3.240740740740741,
|
| 432 |
+
"grad_norm": 37.950408935546875,
|
| 433 |
+
"learning_rate": 1.862397119341564e-05,
|
| 434 |
+
"loss": 0.5107,
|
| 435 |
+
"step": 1400
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 3.298611111111111,
|
| 439 |
+
"grad_norm": 2.259492874145508,
|
| 440 |
+
"learning_rate": 1.8559670781893006e-05,
|
| 441 |
+
"loss": 0.5709,
|
| 442 |
+
"step": 1425
|
| 443 |
+
},
|
| 444 |
+
{
|
| 445 |
+
"epoch": 3.3564814814814814,
|
| 446 |
+
"grad_norm": 4.004894733428955,
|
| 447 |
+
"learning_rate": 1.849537037037037e-05,
|
| 448 |
+
"loss": 0.6343,
|
| 449 |
+
"step": 1450
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 3.414351851851852,
|
| 453 |
+
"grad_norm": 47.317169189453125,
|
| 454 |
+
"learning_rate": 1.843106995884774e-05,
|
| 455 |
+
"loss": 0.4641,
|
| 456 |
+
"step": 1475
|
| 457 |
+
},
|
| 458 |
+
{
|
| 459 |
+
"epoch": 3.4722222222222223,
|
| 460 |
+
"grad_norm": 29.025117874145508,
|
| 461 |
+
"learning_rate": 1.8366769547325105e-05,
|
| 462 |
+
"loss": 0.523,
|
| 463 |
+
"step": 1500
|
| 464 |
+
},
|
| 465 |
+
{
|
| 466 |
+
"epoch": 3.5300925925925926,
|
| 467 |
+
"grad_norm": 1.798436164855957,
|
| 468 |
+
"learning_rate": 1.830246913580247e-05,
|
| 469 |
+
"loss": 0.6264,
|
| 470 |
+
"step": 1525
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 3.587962962962963,
|
| 474 |
+
"grad_norm": 2.9521026611328125,
|
| 475 |
+
"learning_rate": 1.8238168724279838e-05,
|
| 476 |
+
"loss": 0.6125,
|
| 477 |
+
"step": 1550
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 3.6458333333333335,
|
| 481 |
+
"grad_norm": 4.000890254974365,
|
| 482 |
+
"learning_rate": 1.8173868312757203e-05,
|
| 483 |
+
"loss": 0.6466,
|
| 484 |
+
"step": 1575
|
| 485 |
+
},
|
| 486 |
+
{
|
| 487 |
+
"epoch": 3.7037037037037037,
|
| 488 |
+
"grad_norm": 0.3495193421840668,
|
| 489 |
+
"learning_rate": 1.8109567901234572e-05,
|
| 490 |
+
"loss": 0.4195,
|
| 491 |
+
"step": 1600
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 3.761574074074074,
|
| 495 |
+
"grad_norm": 0.2938222885131836,
|
| 496 |
+
"learning_rate": 1.8045267489711934e-05,
|
| 497 |
+
"loss": 0.5228,
|
| 498 |
+
"step": 1625
|
| 499 |
+
},
|
| 500 |
+
{
|
| 501 |
+
"epoch": 3.8194444444444446,
|
| 502 |
+
"grad_norm": 3.754556894302368,
|
| 503 |
+
"learning_rate": 1.7980967078189302e-05,
|
| 504 |
+
"loss": 0.5778,
|
| 505 |
+
"step": 1650
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 3.877314814814815,
|
| 509 |
+
"grad_norm": 0.34463047981262207,
|
| 510 |
+
"learning_rate": 1.7916666666666667e-05,
|
| 511 |
+
"loss": 0.6117,
|
| 512 |
+
"step": 1675
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 3.935185185185185,
|
| 516 |
+
"grad_norm": 20.288467407226562,
|
| 517 |
+
"learning_rate": 1.7852366255144032e-05,
|
| 518 |
+
"loss": 0.5804,
|
| 519 |
+
"step": 1700
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 3.9930555555555554,
|
| 523 |
+
"grad_norm": 2.052157402038574,
|
| 524 |
+
"learning_rate": 1.77880658436214e-05,
|
| 525 |
+
"loss": 0.6001,
|
| 526 |
+
"step": 1725
|
| 527 |
+
},
|
| 528 |
+
{
|
| 529 |
+
"epoch": 4.0,
|
| 530 |
+
"eval_accuracy": 0.9563802083333334,
|
| 531 |
+
"eval_loss": 0.19122980535030365,
|
| 532 |
+
"eval_macro_f1": 0.8559615194005116,
|
| 533 |
+
"eval_runtime": 3.6274,
|
| 534 |
+
"eval_samples_per_second": 423.445,
|
| 535 |
+
"eval_steps_per_second": 26.465,
|
| 536 |
+
"eval_weighted_f1": 0.9557423234178614,
|
| 537 |
+
"step": 1728
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 4.050925925925926,
|
| 541 |
+
"grad_norm": 37.834266662597656,
|
| 542 |
+
"learning_rate": 1.7723765432098766e-05,
|
| 543 |
+
"loss": 0.5204,
|
| 544 |
+
"step": 1750
|
| 545 |
+
},
|
| 546 |
+
{
|
| 547 |
+
"epoch": 4.108796296296297,
|
| 548 |
+
"grad_norm": 11.706012725830078,
|
| 549 |
+
"learning_rate": 1.7659465020576135e-05,
|
| 550 |
+
"loss": 0.5053,
|
| 551 |
+
"step": 1775
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 4.166666666666667,
|
| 555 |
+
"grad_norm": 74.52820587158203,
|
| 556 |
+
"learning_rate": 1.75951646090535e-05,
|
| 557 |
+
"loss": 0.3455,
|
| 558 |
+
"step": 1800
|
| 559 |
+
},
|
| 560 |
+
{
|
| 561 |
+
"epoch": 4.224537037037037,
|
| 562 |
+
"grad_norm": 4.659731388092041,
|
| 563 |
+
"learning_rate": 1.7530864197530865e-05,
|
| 564 |
+
"loss": 0.3939,
|
| 565 |
+
"step": 1825
|
| 566 |
+
},
|
| 567 |
+
{
|
| 568 |
+
"epoch": 4.282407407407407,
|
| 569 |
+
"grad_norm": 1.2450939416885376,
|
| 570 |
+
"learning_rate": 1.7466563786008233e-05,
|
| 571 |
+
"loss": 0.4454,
|
| 572 |
+
"step": 1850
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 4.340277777777778,
|
| 576 |
+
"grad_norm": 2.68330454826355,
|
| 577 |
+
"learning_rate": 1.74022633744856e-05,
|
| 578 |
+
"loss": 0.4551,
|
| 579 |
+
"step": 1875
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 4.398148148148148,
|
| 583 |
+
"grad_norm": 69.6890869140625,
|
| 584 |
+
"learning_rate": 1.7337962962962964e-05,
|
| 585 |
+
"loss": 0.6496,
|
| 586 |
+
"step": 1900
|
| 587 |
+
},
|
| 588 |
+
{
|
| 589 |
+
"epoch": 4.456018518518518,
|
| 590 |
+
"grad_norm": 1.563501238822937,
|
| 591 |
+
"learning_rate": 1.7273662551440332e-05,
|
| 592 |
+
"loss": 0.4315,
|
| 593 |
+
"step": 1925
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"epoch": 4.513888888888889,
|
| 597 |
+
"grad_norm": 48.1222038269043,
|
| 598 |
+
"learning_rate": 1.7209362139917697e-05,
|
| 599 |
+
"loss": 0.5742,
|
| 600 |
+
"step": 1950
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"epoch": 4.5717592592592595,
|
| 604 |
+
"grad_norm": 21.362390518188477,
|
| 605 |
+
"learning_rate": 1.7145061728395063e-05,
|
| 606 |
+
"loss": 0.3452,
|
| 607 |
+
"step": 1975
|
| 608 |
+
},
|
| 609 |
+
{
|
| 610 |
+
"epoch": 4.62962962962963,
|
| 611 |
+
"grad_norm": 1.2060670852661133,
|
| 612 |
+
"learning_rate": 1.708076131687243e-05,
|
| 613 |
+
"loss": 0.4128,
|
| 614 |
+
"step": 2000
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 4.6875,
|
| 618 |
+
"grad_norm": 0.4610968828201294,
|
| 619 |
+
"learning_rate": 1.7016460905349796e-05,
|
| 620 |
+
"loss": 0.4159,
|
| 621 |
+
"step": 2025
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 4.74537037037037,
|
| 625 |
+
"grad_norm": 17.866178512573242,
|
| 626 |
+
"learning_rate": 1.695216049382716e-05,
|
| 627 |
+
"loss": 0.4982,
|
| 628 |
+
"step": 2050
|
| 629 |
+
},
|
| 630 |
+
{
|
| 631 |
+
"epoch": 4.8032407407407405,
|
| 632 |
+
"grad_norm": 25.51858139038086,
|
| 633 |
+
"learning_rate": 1.6887860082304527e-05,
|
| 634 |
+
"loss": 0.5368,
|
| 635 |
+
"step": 2075
|
| 636 |
+
},
|
| 637 |
+
{
|
| 638 |
+
"epoch": 4.861111111111111,
|
| 639 |
+
"grad_norm": 1.593138575553894,
|
| 640 |
+
"learning_rate": 1.6823559670781895e-05,
|
| 641 |
+
"loss": 0.4265,
|
| 642 |
+
"step": 2100
|
| 643 |
+
},
|
| 644 |
+
{
|
| 645 |
+
"epoch": 4.918981481481482,
|
| 646 |
+
"grad_norm": 0.2837437391281128,
|
| 647 |
+
"learning_rate": 1.675925925925926e-05,
|
| 648 |
+
"loss": 0.4619,
|
| 649 |
+
"step": 2125
|
| 650 |
+
},
|
| 651 |
+
{
|
| 652 |
+
"epoch": 4.976851851851852,
|
| 653 |
+
"grad_norm": 20.54825782775879,
|
| 654 |
+
"learning_rate": 1.6694958847736625e-05,
|
| 655 |
+
"loss": 0.454,
|
| 656 |
+
"step": 2150
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 5.0,
|
| 660 |
+
"eval_accuracy": 0.9537760416666666,
|
| 661 |
+
"eval_loss": 0.2081468105316162,
|
| 662 |
+
"eval_macro_f1": 0.8459268835739424,
|
| 663 |
+
"eval_runtime": 3.5561,
|
| 664 |
+
"eval_samples_per_second": 431.934,
|
| 665 |
+
"eval_steps_per_second": 26.996,
|
| 666 |
+
"eval_weighted_f1": 0.9516449736946061,
|
| 667 |
+
"step": 2160
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 5.034722222222222,
|
| 671 |
+
"grad_norm": 1.7740075588226318,
|
| 672 |
+
"learning_rate": 1.6630658436213994e-05,
|
| 673 |
+
"loss": 0.4355,
|
| 674 |
+
"step": 2175
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 5.092592592592593,
|
| 678 |
+
"grad_norm": 0.23488830029964447,
|
| 679 |
+
"learning_rate": 1.656635802469136e-05,
|
| 680 |
+
"loss": 0.2468,
|
| 681 |
+
"step": 2200
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 5.150462962962963,
|
| 685 |
+
"grad_norm": 0.2840358018875122,
|
| 686 |
+
"learning_rate": 1.6502057613168724e-05,
|
| 687 |
+
"loss": 0.1017,
|
| 688 |
+
"step": 2225
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 5.208333333333333,
|
| 692 |
+
"grad_norm": 3.759443521499634,
|
| 693 |
+
"learning_rate": 1.6437757201646093e-05,
|
| 694 |
+
"loss": 0.3733,
|
| 695 |
+
"step": 2250
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 5.266203703703704,
|
| 699 |
+
"grad_norm": 0.17505097389221191,
|
| 700 |
+
"learning_rate": 1.6373456790123458e-05,
|
| 701 |
+
"loss": 0.2434,
|
| 702 |
+
"step": 2275
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 5.324074074074074,
|
| 706 |
+
"grad_norm": 1.583662509918213,
|
| 707 |
+
"learning_rate": 1.6309156378600823e-05,
|
| 708 |
+
"loss": 0.2803,
|
| 709 |
+
"step": 2300
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 5.381944444444445,
|
| 713 |
+
"grad_norm": 0.16322031617164612,
|
| 714 |
+
"learning_rate": 1.624485596707819e-05,
|
| 715 |
+
"loss": 0.2419,
|
| 716 |
+
"step": 2325
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 5.439814814814815,
|
| 720 |
+
"grad_norm": 1.7690539360046387,
|
| 721 |
+
"learning_rate": 1.6180555555555557e-05,
|
| 722 |
+
"loss": 0.5026,
|
| 723 |
+
"step": 2350
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 5.497685185185185,
|
| 727 |
+
"grad_norm": 0.19385340809822083,
|
| 728 |
+
"learning_rate": 1.6116255144032925e-05,
|
| 729 |
+
"loss": 0.3747,
|
| 730 |
+
"step": 2375
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 5.555555555555555,
|
| 734 |
+
"grad_norm": 25.727092742919922,
|
| 735 |
+
"learning_rate": 1.605195473251029e-05,
|
| 736 |
+
"loss": 0.5075,
|
| 737 |
+
"step": 2400
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 5.613425925925926,
|
| 741 |
+
"grad_norm": 3.826075553894043,
|
| 742 |
+
"learning_rate": 1.5987654320987655e-05,
|
| 743 |
+
"loss": 0.4436,
|
| 744 |
+
"step": 2425
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 5.671296296296296,
|
| 748 |
+
"grad_norm": 9.009202003479004,
|
| 749 |
+
"learning_rate": 1.5923353909465024e-05,
|
| 750 |
+
"loss": 0.5325,
|
| 751 |
+
"step": 2450
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 5.729166666666667,
|
| 755 |
+
"grad_norm": 3.7121341228485107,
|
| 756 |
+
"learning_rate": 1.5859053497942386e-05,
|
| 757 |
+
"loss": 0.46,
|
| 758 |
+
"step": 2475
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 5.787037037037037,
|
| 762 |
+
"grad_norm": 0.35388895869255066,
|
| 763 |
+
"learning_rate": 1.5794753086419754e-05,
|
| 764 |
+
"loss": 0.4543,
|
| 765 |
+
"step": 2500
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 5.844907407407407,
|
| 769 |
+
"grad_norm": 2.879859685897827,
|
| 770 |
+
"learning_rate": 1.573045267489712e-05,
|
| 771 |
+
"loss": 0.5029,
|
| 772 |
+
"step": 2525
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 5.902777777777778,
|
| 776 |
+
"grad_norm": 4.517524719238281,
|
| 777 |
+
"learning_rate": 1.5666152263374488e-05,
|
| 778 |
+
"loss": 0.49,
|
| 779 |
+
"step": 2550
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 5.960648148148148,
|
| 783 |
+
"grad_norm": 35.183074951171875,
|
| 784 |
+
"learning_rate": 1.5601851851851853e-05,
|
| 785 |
+
"loss": 0.3209,
|
| 786 |
+
"step": 2575
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 6.0,
|
| 790 |
+
"eval_accuracy": 0.9583333333333334,
|
| 791 |
+
"eval_loss": 0.19548334181308746,
|
| 792 |
+
"eval_macro_f1": 0.8739757580611789,
|
| 793 |
+
"eval_runtime": 2.5743,
|
| 794 |
+
"eval_samples_per_second": 596.663,
|
| 795 |
+
"eval_steps_per_second": 37.291,
|
| 796 |
+
"eval_weighted_f1": 0.9579647631340399,
|
| 797 |
+
"step": 2592
|
| 798 |
+
}
|
| 799 |
+
],
|
| 800 |
+
"logging_steps": 25,
|
| 801 |
+
"max_steps": 8640,
|
| 802 |
+
"num_input_tokens_seen": 0,
|
| 803 |
+
"num_train_epochs": 20,
|
| 804 |
+
"save_steps": 500,
|
| 805 |
+
"stateful_callbacks": {
|
| 806 |
+
"EarlyStoppingCallback": {
|
| 807 |
+
"args": {
|
| 808 |
+
"early_stopping_patience": 3,
|
| 809 |
+
"early_stopping_threshold": 0.0
|
| 810 |
+
},
|
| 811 |
+
"attributes": {
|
| 812 |
+
"early_stopping_patience_counter": 0
|
| 813 |
+
}
|
| 814 |
+
},
|
| 815 |
+
"TrainerControl": {
|
| 816 |
+
"args": {
|
| 817 |
+
"should_epoch_stop": false,
|
| 818 |
+
"should_evaluate": false,
|
| 819 |
+
"should_log": false,
|
| 820 |
+
"should_save": true,
|
| 821 |
+
"should_training_stop": false
|
| 822 |
+
},
|
| 823 |
+
"attributes": {}
|
| 824 |
+
}
|
| 825 |
+
},
|
| 826 |
+
"total_flos": 0.0,
|
| 827 |
+
"train_batch_size": 8,
|
| 828 |
+
"trial_name": null,
|
| 829 |
+
"trial_params": null
|
| 830 |
+
}
|
ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-2592/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3c3407bdf0f213cd25e377b83459c541b3e7805ba365d4b8290866b369c921e2
|
| 3 |
+
size 5624
|
ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-2592/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-3888/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-3888/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-3888/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1e394e06dd1a7c3e2047b278310b7a0d6f7af424a833cceae8d1fb5165dfffab
|
| 3 |
+
size 540026436
|
ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-3888/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f3be39976e8873df12de3f5e08d3eb19b08b1a150b845d6d3211299bdb2ccc10
|
| 3 |
+
size 1075446330
|
ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-3888/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:04bef5130452101227bbeace512f2c6f44183bc3f946985633f1046013a14dbe
|
| 3 |
+
size 14244
|
ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-3888/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9d6dd98ce2e9ba7fea97959cfff523d12d0d4fbb7cb464d5bbf82cfe250d2404
|
| 3 |
+
size 1064
|
ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-3888/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 |
+
}
|
ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-3888/tokenizer_config.json
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"eos_token": "</s>",
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "<mask>",
|
| 50 |
+
"model_max_length": 256,
|
| 51 |
+
"pad_token": "<pad>",
|
| 52 |
+
"sep_token": "</s>",
|
| 53 |
+
"tokenizer_class": "PhobertTokenizer",
|
| 54 |
+
"unk_token": "<unk>"
|
| 55 |
+
}
|
ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-3888/trainer_state.json
ADDED
|
@@ -0,0 +1,1227 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 2592,
|
| 3 |
+
"best_metric": 0.8739757580611789,
|
| 4 |
+
"best_model_checkpoint": "/root/VNEncoder/results/checkpoints/ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-2592",
|
| 5 |
+
"epoch": 9.0,
|
| 6 |
+
"eval_steps": 500,
|
| 7 |
+
"global_step": 3888,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.05787037037037037,
|
| 14 |
+
"grad_norm": 18.34950828552246,
|
| 15 |
+
"learning_rate": 5.555555555555555e-07,
|
| 16 |
+
"loss": 4.7892,
|
| 17 |
+
"step": 25
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.11574074074074074,
|
| 21 |
+
"grad_norm": 17.702529907226562,
|
| 22 |
+
"learning_rate": 1.1342592592592594e-06,
|
| 23 |
+
"loss": 4.5178,
|
| 24 |
+
"step": 50
|
| 25 |
+
},
|
| 26 |
+
{
|
| 27 |
+
"epoch": 0.1736111111111111,
|
| 28 |
+
"grad_norm": 12.634126663208008,
|
| 29 |
+
"learning_rate": 1.7129629629629632e-06,
|
| 30 |
+
"loss": 4.01,
|
| 31 |
+
"step": 75
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"epoch": 0.23148148148148148,
|
| 35 |
+
"grad_norm": 11.140002250671387,
|
| 36 |
+
"learning_rate": 2.2916666666666666e-06,
|
| 37 |
+
"loss": 3.4342,
|
| 38 |
+
"step": 100
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.28935185185185186,
|
| 42 |
+
"grad_norm": 10.938908576965332,
|
| 43 |
+
"learning_rate": 2.8703703703703706e-06,
|
| 44 |
+
"loss": 2.9665,
|
| 45 |
+
"step": 125
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.3472222222222222,
|
| 49 |
+
"grad_norm": 10.947237968444824,
|
| 50 |
+
"learning_rate": 3.449074074074074e-06,
|
| 51 |
+
"loss": 2.4821,
|
| 52 |
+
"step": 150
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"epoch": 0.4050925925925926,
|
| 56 |
+
"grad_norm": 16.171165466308594,
|
| 57 |
+
"learning_rate": 4.027777777777779e-06,
|
| 58 |
+
"loss": 1.9626,
|
| 59 |
+
"step": 175
|
| 60 |
+
},
|
| 61 |
+
{
|
| 62 |
+
"epoch": 0.46296296296296297,
|
| 63 |
+
"grad_norm": 15.498010635375977,
|
| 64 |
+
"learning_rate": 4.606481481481481e-06,
|
| 65 |
+
"loss": 1.4921,
|
| 66 |
+
"step": 200
|
| 67 |
+
},
|
| 68 |
+
{
|
| 69 |
+
"epoch": 0.5208333333333334,
|
| 70 |
+
"grad_norm": 20.90071678161621,
|
| 71 |
+
"learning_rate": 5.185185185185185e-06,
|
| 72 |
+
"loss": 1.4152,
|
| 73 |
+
"step": 225
|
| 74 |
+
},
|
| 75 |
+
{
|
| 76 |
+
"epoch": 0.5787037037037037,
|
| 77 |
+
"grad_norm": 14.971171379089355,
|
| 78 |
+
"learning_rate": 5.7638888888888886e-06,
|
| 79 |
+
"loss": 1.1941,
|
| 80 |
+
"step": 250
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.6365740740740741,
|
| 84 |
+
"grad_norm": 15.980868339538574,
|
| 85 |
+
"learning_rate": 6.342592592592594e-06,
|
| 86 |
+
"loss": 1.0189,
|
| 87 |
+
"step": 275
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.6944444444444444,
|
| 91 |
+
"grad_norm": 5.475146770477295,
|
| 92 |
+
"learning_rate": 6.9212962962962974e-06,
|
| 93 |
+
"loss": 1.0733,
|
| 94 |
+
"step": 300
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"epoch": 0.7523148148148148,
|
| 98 |
+
"grad_norm": 19.17060661315918,
|
| 99 |
+
"learning_rate": 7.500000000000001e-06,
|
| 100 |
+
"loss": 1.0232,
|
| 101 |
+
"step": 325
|
| 102 |
+
},
|
| 103 |
+
{
|
| 104 |
+
"epoch": 0.8101851851851852,
|
| 105 |
+
"grad_norm": 30.633872985839844,
|
| 106 |
+
"learning_rate": 8.078703703703705e-06,
|
| 107 |
+
"loss": 0.9152,
|
| 108 |
+
"step": 350
|
| 109 |
+
},
|
| 110 |
+
{
|
| 111 |
+
"epoch": 0.8680555555555556,
|
| 112 |
+
"grad_norm": 14.854605674743652,
|
| 113 |
+
"learning_rate": 8.657407407407408e-06,
|
| 114 |
+
"loss": 0.8892,
|
| 115 |
+
"step": 375
|
| 116 |
+
},
|
| 117 |
+
{
|
| 118 |
+
"epoch": 0.9259259259259259,
|
| 119 |
+
"grad_norm": 15.524602890014648,
|
| 120 |
+
"learning_rate": 9.236111111111112e-06,
|
| 121 |
+
"loss": 1.2148,
|
| 122 |
+
"step": 400
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.9837962962962963,
|
| 126 |
+
"grad_norm": 8.35659408569336,
|
| 127 |
+
"learning_rate": 9.814814814814815e-06,
|
| 128 |
+
"loss": 1.0522,
|
| 129 |
+
"step": 425
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 1.0,
|
| 133 |
+
"eval_accuracy": 0.9427083333333334,
|
| 134 |
+
"eval_loss": 0.19718892872333527,
|
| 135 |
+
"eval_macro_f1": 0.6518213119914867,
|
| 136 |
+
"eval_runtime": 3.6776,
|
| 137 |
+
"eval_samples_per_second": 417.662,
|
| 138 |
+
"eval_steps_per_second": 26.104,
|
| 139 |
+
"eval_weighted_f1": 0.9292832259638407,
|
| 140 |
+
"step": 432
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 1.0416666666666667,
|
| 144 |
+
"grad_norm": 9.545726776123047,
|
| 145 |
+
"learning_rate": 1.0393518518518519e-05,
|
| 146 |
+
"loss": 0.9327,
|
| 147 |
+
"step": 450
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 1.099537037037037,
|
| 151 |
+
"grad_norm": 28.5867977142334,
|
| 152 |
+
"learning_rate": 1.0972222222222224e-05,
|
| 153 |
+
"loss": 0.8578,
|
| 154 |
+
"step": 475
|
| 155 |
+
},
|
| 156 |
+
{
|
| 157 |
+
"epoch": 1.1574074074074074,
|
| 158 |
+
"grad_norm": 26.383472442626953,
|
| 159 |
+
"learning_rate": 1.1550925925925926e-05,
|
| 160 |
+
"loss": 0.615,
|
| 161 |
+
"step": 500
|
| 162 |
+
},
|
| 163 |
+
{
|
| 164 |
+
"epoch": 1.2152777777777777,
|
| 165 |
+
"grad_norm": 7.6479668617248535,
|
| 166 |
+
"learning_rate": 1.2129629629629631e-05,
|
| 167 |
+
"loss": 0.8475,
|
| 168 |
+
"step": 525
|
| 169 |
+
},
|
| 170 |
+
{
|
| 171 |
+
"epoch": 1.2731481481481481,
|
| 172 |
+
"grad_norm": 9.217597961425781,
|
| 173 |
+
"learning_rate": 1.2708333333333333e-05,
|
| 174 |
+
"loss": 0.7591,
|
| 175 |
+
"step": 550
|
| 176 |
+
},
|
| 177 |
+
{
|
| 178 |
+
"epoch": 1.3310185185185186,
|
| 179 |
+
"grad_norm": 9.50477123260498,
|
| 180 |
+
"learning_rate": 1.3287037037037039e-05,
|
| 181 |
+
"loss": 0.8333,
|
| 182 |
+
"step": 575
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 1.3888888888888888,
|
| 186 |
+
"grad_norm": 282.826416015625,
|
| 187 |
+
"learning_rate": 1.3865740740740742e-05,
|
| 188 |
+
"loss": 0.6888,
|
| 189 |
+
"step": 600
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 1.4467592592592593,
|
| 193 |
+
"grad_norm": 8.852691650390625,
|
| 194 |
+
"learning_rate": 1.4444444444444446e-05,
|
| 195 |
+
"loss": 0.969,
|
| 196 |
+
"step": 625
|
| 197 |
+
},
|
| 198 |
+
{
|
| 199 |
+
"epoch": 1.5046296296296298,
|
| 200 |
+
"grad_norm": 10.682754516601562,
|
| 201 |
+
"learning_rate": 1.502314814814815e-05,
|
| 202 |
+
"loss": 0.7574,
|
| 203 |
+
"step": 650
|
| 204 |
+
},
|
| 205 |
+
{
|
| 206 |
+
"epoch": 1.5625,
|
| 207 |
+
"grad_norm": 8.411812782287598,
|
| 208 |
+
"learning_rate": 1.5601851851851853e-05,
|
| 209 |
+
"loss": 0.8515,
|
| 210 |
+
"step": 675
|
| 211 |
+
},
|
| 212 |
+
{
|
| 213 |
+
"epoch": 1.6203703703703702,
|
| 214 |
+
"grad_norm": 4.259770393371582,
|
| 215 |
+
"learning_rate": 1.6180555555555557e-05,
|
| 216 |
+
"loss": 0.7792,
|
| 217 |
+
"step": 700
|
| 218 |
+
},
|
| 219 |
+
{
|
| 220 |
+
"epoch": 1.6782407407407407,
|
| 221 |
+
"grad_norm": 43.64582061767578,
|
| 222 |
+
"learning_rate": 1.675925925925926e-05,
|
| 223 |
+
"loss": 0.7105,
|
| 224 |
+
"step": 725
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 1.7361111111111112,
|
| 228 |
+
"grad_norm": 12.682132720947266,
|
| 229 |
+
"learning_rate": 1.7337962962962964e-05,
|
| 230 |
+
"loss": 0.8808,
|
| 231 |
+
"step": 750
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 1.7939814814814814,
|
| 235 |
+
"grad_norm": 0.8860751986503601,
|
| 236 |
+
"learning_rate": 1.7916666666666667e-05,
|
| 237 |
+
"loss": 0.6108,
|
| 238 |
+
"step": 775
|
| 239 |
+
},
|
| 240 |
+
{
|
| 241 |
+
"epoch": 1.8518518518518519,
|
| 242 |
+
"grad_norm": 49.53169250488281,
|
| 243 |
+
"learning_rate": 1.849537037037037e-05,
|
| 244 |
+
"loss": 0.8255,
|
| 245 |
+
"step": 800
|
| 246 |
+
},
|
| 247 |
+
{
|
| 248 |
+
"epoch": 1.9097222222222223,
|
| 249 |
+
"grad_norm": 29.884153366088867,
|
| 250 |
+
"learning_rate": 1.9074074074074075e-05,
|
| 251 |
+
"loss": 1.0157,
|
| 252 |
+
"step": 825
|
| 253 |
+
},
|
| 254 |
+
{
|
| 255 |
+
"epoch": 1.9675925925925926,
|
| 256 |
+
"grad_norm": 10.482592582702637,
|
| 257 |
+
"learning_rate": 1.9652777777777778e-05,
|
| 258 |
+
"loss": 0.6026,
|
| 259 |
+
"step": 850
|
| 260 |
+
},
|
| 261 |
+
{
|
| 262 |
+
"epoch": 2.0,
|
| 263 |
+
"eval_accuracy": 0.951171875,
|
| 264 |
+
"eval_loss": 0.1825077086687088,
|
| 265 |
+
"eval_macro_f1": 0.8026620268618748,
|
| 266 |
+
"eval_runtime": 2.5618,
|
| 267 |
+
"eval_samples_per_second": 599.589,
|
| 268 |
+
"eval_steps_per_second": 37.474,
|
| 269 |
+
"eval_weighted_f1": 0.948829490199814,
|
| 270 |
+
"step": 864
|
| 271 |
+
},
|
| 272 |
+
{
|
| 273 |
+
"epoch": 2.025462962962963,
|
| 274 |
+
"grad_norm": 31.20493507385254,
|
| 275 |
+
"learning_rate": 1.997427983539095e-05,
|
| 276 |
+
"loss": 0.6416,
|
| 277 |
+
"step": 875
|
| 278 |
+
},
|
| 279 |
+
{
|
| 280 |
+
"epoch": 2.0833333333333335,
|
| 281 |
+
"grad_norm": 7.576644420623779,
|
| 282 |
+
"learning_rate": 1.9909979423868314e-05,
|
| 283 |
+
"loss": 0.7873,
|
| 284 |
+
"step": 900
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 2.1412037037037037,
|
| 288 |
+
"grad_norm": 5.631798267364502,
|
| 289 |
+
"learning_rate": 1.984567901234568e-05,
|
| 290 |
+
"loss": 0.7779,
|
| 291 |
+
"step": 925
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 2.199074074074074,
|
| 295 |
+
"grad_norm": 19.126354217529297,
|
| 296 |
+
"learning_rate": 1.9781378600823048e-05,
|
| 297 |
+
"loss": 0.5556,
|
| 298 |
+
"step": 950
|
| 299 |
+
},
|
| 300 |
+
{
|
| 301 |
+
"epoch": 2.2569444444444446,
|
| 302 |
+
"grad_norm": 5.889681339263916,
|
| 303 |
+
"learning_rate": 1.9717078189300413e-05,
|
| 304 |
+
"loss": 0.5082,
|
| 305 |
+
"step": 975
|
| 306 |
+
},
|
| 307 |
+
{
|
| 308 |
+
"epoch": 2.314814814814815,
|
| 309 |
+
"grad_norm": 86.90234375,
|
| 310 |
+
"learning_rate": 1.9652777777777778e-05,
|
| 311 |
+
"loss": 0.8321,
|
| 312 |
+
"step": 1000
|
| 313 |
+
},
|
| 314 |
+
{
|
| 315 |
+
"epoch": 2.372685185185185,
|
| 316 |
+
"grad_norm": 37.04526138305664,
|
| 317 |
+
"learning_rate": 1.9588477366255147e-05,
|
| 318 |
+
"loss": 0.6905,
|
| 319 |
+
"step": 1025
|
| 320 |
+
},
|
| 321 |
+
{
|
| 322 |
+
"epoch": 2.4305555555555554,
|
| 323 |
+
"grad_norm": 26.49477195739746,
|
| 324 |
+
"learning_rate": 1.9524176954732512e-05,
|
| 325 |
+
"loss": 0.6623,
|
| 326 |
+
"step": 1050
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 2.488425925925926,
|
| 330 |
+
"grad_norm": 61.23967361450195,
|
| 331 |
+
"learning_rate": 1.945987654320988e-05,
|
| 332 |
+
"loss": 0.7448,
|
| 333 |
+
"step": 1075
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 2.5462962962962963,
|
| 337 |
+
"grad_norm": 16.973339080810547,
|
| 338 |
+
"learning_rate": 1.9395576131687242e-05,
|
| 339 |
+
"loss": 0.4993,
|
| 340 |
+
"step": 1100
|
| 341 |
+
},
|
| 342 |
+
{
|
| 343 |
+
"epoch": 2.6041666666666665,
|
| 344 |
+
"grad_norm": 20.362926483154297,
|
| 345 |
+
"learning_rate": 1.933127572016461e-05,
|
| 346 |
+
"loss": 0.6888,
|
| 347 |
+
"step": 1125
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 2.662037037037037,
|
| 351 |
+
"grad_norm": 2.426138401031494,
|
| 352 |
+
"learning_rate": 1.926697530864198e-05,
|
| 353 |
+
"loss": 0.7204,
|
| 354 |
+
"step": 1150
|
| 355 |
+
},
|
| 356 |
+
{
|
| 357 |
+
"epoch": 2.7199074074074074,
|
| 358 |
+
"grad_norm": 38.947898864746094,
|
| 359 |
+
"learning_rate": 1.920267489711934e-05,
|
| 360 |
+
"loss": 0.8354,
|
| 361 |
+
"step": 1175
|
| 362 |
+
},
|
| 363 |
+
{
|
| 364 |
+
"epoch": 2.7777777777777777,
|
| 365 |
+
"grad_norm": 29.757408142089844,
|
| 366 |
+
"learning_rate": 1.913837448559671e-05,
|
| 367 |
+
"loss": 0.735,
|
| 368 |
+
"step": 1200
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 2.835648148148148,
|
| 372 |
+
"grad_norm": 4.250671863555908,
|
| 373 |
+
"learning_rate": 1.9074074074074075e-05,
|
| 374 |
+
"loss": 0.5405,
|
| 375 |
+
"step": 1225
|
| 376 |
+
},
|
| 377 |
+
{
|
| 378 |
+
"epoch": 2.8935185185185186,
|
| 379 |
+
"grad_norm": 4.031290531158447,
|
| 380 |
+
"learning_rate": 1.9009773662551443e-05,
|
| 381 |
+
"loss": 0.7388,
|
| 382 |
+
"step": 1250
|
| 383 |
+
},
|
| 384 |
+
{
|
| 385 |
+
"epoch": 2.951388888888889,
|
| 386 |
+
"grad_norm": 11.694859504699707,
|
| 387 |
+
"learning_rate": 1.8945473251028808e-05,
|
| 388 |
+
"loss": 0.6715,
|
| 389 |
+
"step": 1275
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 3.0,
|
| 393 |
+
"eval_accuracy": 0.9498697916666666,
|
| 394 |
+
"eval_loss": 0.18208689987659454,
|
| 395 |
+
"eval_macro_f1": 0.8273221081884374,
|
| 396 |
+
"eval_runtime": 3.7353,
|
| 397 |
+
"eval_samples_per_second": 411.214,
|
| 398 |
+
"eval_steps_per_second": 25.701,
|
| 399 |
+
"eval_weighted_f1": 0.9502826564284111,
|
| 400 |
+
"step": 1296
|
| 401 |
+
},
|
| 402 |
+
{
|
| 403 |
+
"epoch": 3.009259259259259,
|
| 404 |
+
"grad_norm": 3.5965843200683594,
|
| 405 |
+
"learning_rate": 1.8881172839506173e-05,
|
| 406 |
+
"loss": 0.6595,
|
| 407 |
+
"step": 1300
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 3.0671296296296298,
|
| 411 |
+
"grad_norm": 23.776273727416992,
|
| 412 |
+
"learning_rate": 1.8816872427983542e-05,
|
| 413 |
+
"loss": 0.3789,
|
| 414 |
+
"step": 1325
|
| 415 |
+
},
|
| 416 |
+
{
|
| 417 |
+
"epoch": 3.125,
|
| 418 |
+
"grad_norm": 1.3928412199020386,
|
| 419 |
+
"learning_rate": 1.8752572016460907e-05,
|
| 420 |
+
"loss": 0.7149,
|
| 421 |
+
"step": 1350
|
| 422 |
+
},
|
| 423 |
+
{
|
| 424 |
+
"epoch": 3.1828703703703702,
|
| 425 |
+
"grad_norm": 0.39887893199920654,
|
| 426 |
+
"learning_rate": 1.8688271604938272e-05,
|
| 427 |
+
"loss": 0.4188,
|
| 428 |
+
"step": 1375
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 3.240740740740741,
|
| 432 |
+
"grad_norm": 37.950408935546875,
|
| 433 |
+
"learning_rate": 1.862397119341564e-05,
|
| 434 |
+
"loss": 0.5107,
|
| 435 |
+
"step": 1400
|
| 436 |
+
},
|
| 437 |
+
{
|
| 438 |
+
"epoch": 3.298611111111111,
|
| 439 |
+
"grad_norm": 2.259492874145508,
|
| 440 |
+
"learning_rate": 1.8559670781893006e-05,
|
| 441 |
+
"loss": 0.5709,
|
| 442 |
+
"step": 1425
|
| 443 |
+
},
|
| 444 |
+
{
|
| 445 |
+
"epoch": 3.3564814814814814,
|
| 446 |
+
"grad_norm": 4.004894733428955,
|
| 447 |
+
"learning_rate": 1.849537037037037e-05,
|
| 448 |
+
"loss": 0.6343,
|
| 449 |
+
"step": 1450
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 3.414351851851852,
|
| 453 |
+
"grad_norm": 47.317169189453125,
|
| 454 |
+
"learning_rate": 1.843106995884774e-05,
|
| 455 |
+
"loss": 0.4641,
|
| 456 |
+
"step": 1475
|
| 457 |
+
},
|
| 458 |
+
{
|
| 459 |
+
"epoch": 3.4722222222222223,
|
| 460 |
+
"grad_norm": 29.025117874145508,
|
| 461 |
+
"learning_rate": 1.8366769547325105e-05,
|
| 462 |
+
"loss": 0.523,
|
| 463 |
+
"step": 1500
|
| 464 |
+
},
|
| 465 |
+
{
|
| 466 |
+
"epoch": 3.5300925925925926,
|
| 467 |
+
"grad_norm": 1.798436164855957,
|
| 468 |
+
"learning_rate": 1.830246913580247e-05,
|
| 469 |
+
"loss": 0.6264,
|
| 470 |
+
"step": 1525
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 3.587962962962963,
|
| 474 |
+
"grad_norm": 2.9521026611328125,
|
| 475 |
+
"learning_rate": 1.8238168724279838e-05,
|
| 476 |
+
"loss": 0.6125,
|
| 477 |
+
"step": 1550
|
| 478 |
+
},
|
| 479 |
+
{
|
| 480 |
+
"epoch": 3.6458333333333335,
|
| 481 |
+
"grad_norm": 4.000890254974365,
|
| 482 |
+
"learning_rate": 1.8173868312757203e-05,
|
| 483 |
+
"loss": 0.6466,
|
| 484 |
+
"step": 1575
|
| 485 |
+
},
|
| 486 |
+
{
|
| 487 |
+
"epoch": 3.7037037037037037,
|
| 488 |
+
"grad_norm": 0.3495193421840668,
|
| 489 |
+
"learning_rate": 1.8109567901234572e-05,
|
| 490 |
+
"loss": 0.4195,
|
| 491 |
+
"step": 1600
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 3.761574074074074,
|
| 495 |
+
"grad_norm": 0.2938222885131836,
|
| 496 |
+
"learning_rate": 1.8045267489711934e-05,
|
| 497 |
+
"loss": 0.5228,
|
| 498 |
+
"step": 1625
|
| 499 |
+
},
|
| 500 |
+
{
|
| 501 |
+
"epoch": 3.8194444444444446,
|
| 502 |
+
"grad_norm": 3.754556894302368,
|
| 503 |
+
"learning_rate": 1.7980967078189302e-05,
|
| 504 |
+
"loss": 0.5778,
|
| 505 |
+
"step": 1650
|
| 506 |
+
},
|
| 507 |
+
{
|
| 508 |
+
"epoch": 3.877314814814815,
|
| 509 |
+
"grad_norm": 0.34463047981262207,
|
| 510 |
+
"learning_rate": 1.7916666666666667e-05,
|
| 511 |
+
"loss": 0.6117,
|
| 512 |
+
"step": 1675
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 3.935185185185185,
|
| 516 |
+
"grad_norm": 20.288467407226562,
|
| 517 |
+
"learning_rate": 1.7852366255144032e-05,
|
| 518 |
+
"loss": 0.5804,
|
| 519 |
+
"step": 1700
|
| 520 |
+
},
|
| 521 |
+
{
|
| 522 |
+
"epoch": 3.9930555555555554,
|
| 523 |
+
"grad_norm": 2.052157402038574,
|
| 524 |
+
"learning_rate": 1.77880658436214e-05,
|
| 525 |
+
"loss": 0.6001,
|
| 526 |
+
"step": 1725
|
| 527 |
+
},
|
| 528 |
+
{
|
| 529 |
+
"epoch": 4.0,
|
| 530 |
+
"eval_accuracy": 0.9563802083333334,
|
| 531 |
+
"eval_loss": 0.19122980535030365,
|
| 532 |
+
"eval_macro_f1": 0.8559615194005116,
|
| 533 |
+
"eval_runtime": 3.6274,
|
| 534 |
+
"eval_samples_per_second": 423.445,
|
| 535 |
+
"eval_steps_per_second": 26.465,
|
| 536 |
+
"eval_weighted_f1": 0.9557423234178614,
|
| 537 |
+
"step": 1728
|
| 538 |
+
},
|
| 539 |
+
{
|
| 540 |
+
"epoch": 4.050925925925926,
|
| 541 |
+
"grad_norm": 37.834266662597656,
|
| 542 |
+
"learning_rate": 1.7723765432098766e-05,
|
| 543 |
+
"loss": 0.5204,
|
| 544 |
+
"step": 1750
|
| 545 |
+
},
|
| 546 |
+
{
|
| 547 |
+
"epoch": 4.108796296296297,
|
| 548 |
+
"grad_norm": 11.706012725830078,
|
| 549 |
+
"learning_rate": 1.7659465020576135e-05,
|
| 550 |
+
"loss": 0.5053,
|
| 551 |
+
"step": 1775
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 4.166666666666667,
|
| 555 |
+
"grad_norm": 74.52820587158203,
|
| 556 |
+
"learning_rate": 1.75951646090535e-05,
|
| 557 |
+
"loss": 0.3455,
|
| 558 |
+
"step": 1800
|
| 559 |
+
},
|
| 560 |
+
{
|
| 561 |
+
"epoch": 4.224537037037037,
|
| 562 |
+
"grad_norm": 4.659731388092041,
|
| 563 |
+
"learning_rate": 1.7530864197530865e-05,
|
| 564 |
+
"loss": 0.3939,
|
| 565 |
+
"step": 1825
|
| 566 |
+
},
|
| 567 |
+
{
|
| 568 |
+
"epoch": 4.282407407407407,
|
| 569 |
+
"grad_norm": 1.2450939416885376,
|
| 570 |
+
"learning_rate": 1.7466563786008233e-05,
|
| 571 |
+
"loss": 0.4454,
|
| 572 |
+
"step": 1850
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 4.340277777777778,
|
| 576 |
+
"grad_norm": 2.68330454826355,
|
| 577 |
+
"learning_rate": 1.74022633744856e-05,
|
| 578 |
+
"loss": 0.4551,
|
| 579 |
+
"step": 1875
|
| 580 |
+
},
|
| 581 |
+
{
|
| 582 |
+
"epoch": 4.398148148148148,
|
| 583 |
+
"grad_norm": 69.6890869140625,
|
| 584 |
+
"learning_rate": 1.7337962962962964e-05,
|
| 585 |
+
"loss": 0.6496,
|
| 586 |
+
"step": 1900
|
| 587 |
+
},
|
| 588 |
+
{
|
| 589 |
+
"epoch": 4.456018518518518,
|
| 590 |
+
"grad_norm": 1.563501238822937,
|
| 591 |
+
"learning_rate": 1.7273662551440332e-05,
|
| 592 |
+
"loss": 0.4315,
|
| 593 |
+
"step": 1925
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"epoch": 4.513888888888889,
|
| 597 |
+
"grad_norm": 48.1222038269043,
|
| 598 |
+
"learning_rate": 1.7209362139917697e-05,
|
| 599 |
+
"loss": 0.5742,
|
| 600 |
+
"step": 1950
|
| 601 |
+
},
|
| 602 |
+
{
|
| 603 |
+
"epoch": 4.5717592592592595,
|
| 604 |
+
"grad_norm": 21.362390518188477,
|
| 605 |
+
"learning_rate": 1.7145061728395063e-05,
|
| 606 |
+
"loss": 0.3452,
|
| 607 |
+
"step": 1975
|
| 608 |
+
},
|
| 609 |
+
{
|
| 610 |
+
"epoch": 4.62962962962963,
|
| 611 |
+
"grad_norm": 1.2060670852661133,
|
| 612 |
+
"learning_rate": 1.708076131687243e-05,
|
| 613 |
+
"loss": 0.4128,
|
| 614 |
+
"step": 2000
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 4.6875,
|
| 618 |
+
"grad_norm": 0.4610968828201294,
|
| 619 |
+
"learning_rate": 1.7016460905349796e-05,
|
| 620 |
+
"loss": 0.4159,
|
| 621 |
+
"step": 2025
|
| 622 |
+
},
|
| 623 |
+
{
|
| 624 |
+
"epoch": 4.74537037037037,
|
| 625 |
+
"grad_norm": 17.866178512573242,
|
| 626 |
+
"learning_rate": 1.695216049382716e-05,
|
| 627 |
+
"loss": 0.4982,
|
| 628 |
+
"step": 2050
|
| 629 |
+
},
|
| 630 |
+
{
|
| 631 |
+
"epoch": 4.8032407407407405,
|
| 632 |
+
"grad_norm": 25.51858139038086,
|
| 633 |
+
"learning_rate": 1.6887860082304527e-05,
|
| 634 |
+
"loss": 0.5368,
|
| 635 |
+
"step": 2075
|
| 636 |
+
},
|
| 637 |
+
{
|
| 638 |
+
"epoch": 4.861111111111111,
|
| 639 |
+
"grad_norm": 1.593138575553894,
|
| 640 |
+
"learning_rate": 1.6823559670781895e-05,
|
| 641 |
+
"loss": 0.4265,
|
| 642 |
+
"step": 2100
|
| 643 |
+
},
|
| 644 |
+
{
|
| 645 |
+
"epoch": 4.918981481481482,
|
| 646 |
+
"grad_norm": 0.2837437391281128,
|
| 647 |
+
"learning_rate": 1.675925925925926e-05,
|
| 648 |
+
"loss": 0.4619,
|
| 649 |
+
"step": 2125
|
| 650 |
+
},
|
| 651 |
+
{
|
| 652 |
+
"epoch": 4.976851851851852,
|
| 653 |
+
"grad_norm": 20.54825782775879,
|
| 654 |
+
"learning_rate": 1.6694958847736625e-05,
|
| 655 |
+
"loss": 0.454,
|
| 656 |
+
"step": 2150
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 5.0,
|
| 660 |
+
"eval_accuracy": 0.9537760416666666,
|
| 661 |
+
"eval_loss": 0.2081468105316162,
|
| 662 |
+
"eval_macro_f1": 0.8459268835739424,
|
| 663 |
+
"eval_runtime": 3.5561,
|
| 664 |
+
"eval_samples_per_second": 431.934,
|
| 665 |
+
"eval_steps_per_second": 26.996,
|
| 666 |
+
"eval_weighted_f1": 0.9516449736946061,
|
| 667 |
+
"step": 2160
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 5.034722222222222,
|
| 671 |
+
"grad_norm": 1.7740075588226318,
|
| 672 |
+
"learning_rate": 1.6630658436213994e-05,
|
| 673 |
+
"loss": 0.4355,
|
| 674 |
+
"step": 2175
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 5.092592592592593,
|
| 678 |
+
"grad_norm": 0.23488830029964447,
|
| 679 |
+
"learning_rate": 1.656635802469136e-05,
|
| 680 |
+
"loss": 0.2468,
|
| 681 |
+
"step": 2200
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 5.150462962962963,
|
| 685 |
+
"grad_norm": 0.2840358018875122,
|
| 686 |
+
"learning_rate": 1.6502057613168724e-05,
|
| 687 |
+
"loss": 0.1017,
|
| 688 |
+
"step": 2225
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 5.208333333333333,
|
| 692 |
+
"grad_norm": 3.759443521499634,
|
| 693 |
+
"learning_rate": 1.6437757201646093e-05,
|
| 694 |
+
"loss": 0.3733,
|
| 695 |
+
"step": 2250
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 5.266203703703704,
|
| 699 |
+
"grad_norm": 0.17505097389221191,
|
| 700 |
+
"learning_rate": 1.6373456790123458e-05,
|
| 701 |
+
"loss": 0.2434,
|
| 702 |
+
"step": 2275
|
| 703 |
+
},
|
| 704 |
+
{
|
| 705 |
+
"epoch": 5.324074074074074,
|
| 706 |
+
"grad_norm": 1.583662509918213,
|
| 707 |
+
"learning_rate": 1.6309156378600823e-05,
|
| 708 |
+
"loss": 0.2803,
|
| 709 |
+
"step": 2300
|
| 710 |
+
},
|
| 711 |
+
{
|
| 712 |
+
"epoch": 5.381944444444445,
|
| 713 |
+
"grad_norm": 0.16322031617164612,
|
| 714 |
+
"learning_rate": 1.624485596707819e-05,
|
| 715 |
+
"loss": 0.2419,
|
| 716 |
+
"step": 2325
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 5.439814814814815,
|
| 720 |
+
"grad_norm": 1.7690539360046387,
|
| 721 |
+
"learning_rate": 1.6180555555555557e-05,
|
| 722 |
+
"loss": 0.5026,
|
| 723 |
+
"step": 2350
|
| 724 |
+
},
|
| 725 |
+
{
|
| 726 |
+
"epoch": 5.497685185185185,
|
| 727 |
+
"grad_norm": 0.19385340809822083,
|
| 728 |
+
"learning_rate": 1.6116255144032925e-05,
|
| 729 |
+
"loss": 0.3747,
|
| 730 |
+
"step": 2375
|
| 731 |
+
},
|
| 732 |
+
{
|
| 733 |
+
"epoch": 5.555555555555555,
|
| 734 |
+
"grad_norm": 25.727092742919922,
|
| 735 |
+
"learning_rate": 1.605195473251029e-05,
|
| 736 |
+
"loss": 0.5075,
|
| 737 |
+
"step": 2400
|
| 738 |
+
},
|
| 739 |
+
{
|
| 740 |
+
"epoch": 5.613425925925926,
|
| 741 |
+
"grad_norm": 3.826075553894043,
|
| 742 |
+
"learning_rate": 1.5987654320987655e-05,
|
| 743 |
+
"loss": 0.4436,
|
| 744 |
+
"step": 2425
|
| 745 |
+
},
|
| 746 |
+
{
|
| 747 |
+
"epoch": 5.671296296296296,
|
| 748 |
+
"grad_norm": 9.009202003479004,
|
| 749 |
+
"learning_rate": 1.5923353909465024e-05,
|
| 750 |
+
"loss": 0.5325,
|
| 751 |
+
"step": 2450
|
| 752 |
+
},
|
| 753 |
+
{
|
| 754 |
+
"epoch": 5.729166666666667,
|
| 755 |
+
"grad_norm": 3.7121341228485107,
|
| 756 |
+
"learning_rate": 1.5859053497942386e-05,
|
| 757 |
+
"loss": 0.46,
|
| 758 |
+
"step": 2475
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 5.787037037037037,
|
| 762 |
+
"grad_norm": 0.35388895869255066,
|
| 763 |
+
"learning_rate": 1.5794753086419754e-05,
|
| 764 |
+
"loss": 0.4543,
|
| 765 |
+
"step": 2500
|
| 766 |
+
},
|
| 767 |
+
{
|
| 768 |
+
"epoch": 5.844907407407407,
|
| 769 |
+
"grad_norm": 2.879859685897827,
|
| 770 |
+
"learning_rate": 1.573045267489712e-05,
|
| 771 |
+
"loss": 0.5029,
|
| 772 |
+
"step": 2525
|
| 773 |
+
},
|
| 774 |
+
{
|
| 775 |
+
"epoch": 5.902777777777778,
|
| 776 |
+
"grad_norm": 4.517524719238281,
|
| 777 |
+
"learning_rate": 1.5666152263374488e-05,
|
| 778 |
+
"loss": 0.49,
|
| 779 |
+
"step": 2550
|
| 780 |
+
},
|
| 781 |
+
{
|
| 782 |
+
"epoch": 5.960648148148148,
|
| 783 |
+
"grad_norm": 35.183074951171875,
|
| 784 |
+
"learning_rate": 1.5601851851851853e-05,
|
| 785 |
+
"loss": 0.3209,
|
| 786 |
+
"step": 2575
|
| 787 |
+
},
|
| 788 |
+
{
|
| 789 |
+
"epoch": 6.0,
|
| 790 |
+
"eval_accuracy": 0.9583333333333334,
|
| 791 |
+
"eval_loss": 0.19548334181308746,
|
| 792 |
+
"eval_macro_f1": 0.8739757580611789,
|
| 793 |
+
"eval_runtime": 2.5743,
|
| 794 |
+
"eval_samples_per_second": 596.663,
|
| 795 |
+
"eval_steps_per_second": 37.291,
|
| 796 |
+
"eval_weighted_f1": 0.9579647631340399,
|
| 797 |
+
"step": 2592
|
| 798 |
+
},
|
| 799 |
+
{
|
| 800 |
+
"epoch": 6.018518518518518,
|
| 801 |
+
"grad_norm": 12.735112190246582,
|
| 802 |
+
"learning_rate": 1.5537551440329218e-05,
|
| 803 |
+
"loss": 0.3838,
|
| 804 |
+
"step": 2600
|
| 805 |
+
},
|
| 806 |
+
{
|
| 807 |
+
"epoch": 6.076388888888889,
|
| 808 |
+
"grad_norm": 17.395652770996094,
|
| 809 |
+
"learning_rate": 1.5473251028806587e-05,
|
| 810 |
+
"loss": 0.32,
|
| 811 |
+
"step": 2625
|
| 812 |
+
},
|
| 813 |
+
{
|
| 814 |
+
"epoch": 6.1342592592592595,
|
| 815 |
+
"grad_norm": 0.2329498827457428,
|
| 816 |
+
"learning_rate": 1.5408950617283952e-05,
|
| 817 |
+
"loss": 0.4262,
|
| 818 |
+
"step": 2650
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 6.19212962962963,
|
| 822 |
+
"grad_norm": 2.936939001083374,
|
| 823 |
+
"learning_rate": 1.5344650205761317e-05,
|
| 824 |
+
"loss": 0.4,
|
| 825 |
+
"step": 2675
|
| 826 |
+
},
|
| 827 |
+
{
|
| 828 |
+
"epoch": 6.25,
|
| 829 |
+
"grad_norm": 0.6509174108505249,
|
| 830 |
+
"learning_rate": 1.5280349794238685e-05,
|
| 831 |
+
"loss": 0.3224,
|
| 832 |
+
"step": 2700
|
| 833 |
+
},
|
| 834 |
+
{
|
| 835 |
+
"epoch": 6.30787037037037,
|
| 836 |
+
"grad_norm": 1.9225895404815674,
|
| 837 |
+
"learning_rate": 1.5216049382716052e-05,
|
| 838 |
+
"loss": 0.2663,
|
| 839 |
+
"step": 2725
|
| 840 |
+
},
|
| 841 |
+
{
|
| 842 |
+
"epoch": 6.3657407407407405,
|
| 843 |
+
"grad_norm": 105.22135162353516,
|
| 844 |
+
"learning_rate": 1.5151748971193416e-05,
|
| 845 |
+
"loss": 0.3169,
|
| 846 |
+
"step": 2750
|
| 847 |
+
},
|
| 848 |
+
{
|
| 849 |
+
"epoch": 6.423611111111111,
|
| 850 |
+
"grad_norm": 0.26478296518325806,
|
| 851 |
+
"learning_rate": 1.5087448559670783e-05,
|
| 852 |
+
"loss": 0.4674,
|
| 853 |
+
"step": 2775
|
| 854 |
+
},
|
| 855 |
+
{
|
| 856 |
+
"epoch": 6.481481481481482,
|
| 857 |
+
"grad_norm": 63.378543853759766,
|
| 858 |
+
"learning_rate": 1.502314814814815e-05,
|
| 859 |
+
"loss": 0.2673,
|
| 860 |
+
"step": 2800
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 6.539351851851852,
|
| 864 |
+
"grad_norm": 78.96966552734375,
|
| 865 |
+
"learning_rate": 1.4958847736625515e-05,
|
| 866 |
+
"loss": 0.2042,
|
| 867 |
+
"step": 2825
|
| 868 |
+
},
|
| 869 |
+
{
|
| 870 |
+
"epoch": 6.597222222222222,
|
| 871 |
+
"grad_norm": 6.753354072570801,
|
| 872 |
+
"learning_rate": 1.4894547325102881e-05,
|
| 873 |
+
"loss": 0.2953,
|
| 874 |
+
"step": 2850
|
| 875 |
+
},
|
| 876 |
+
{
|
| 877 |
+
"epoch": 6.655092592592593,
|
| 878 |
+
"grad_norm": 68.41000366210938,
|
| 879 |
+
"learning_rate": 1.4830246913580248e-05,
|
| 880 |
+
"loss": 0.1997,
|
| 881 |
+
"step": 2875
|
| 882 |
+
},
|
| 883 |
+
{
|
| 884 |
+
"epoch": 6.712962962962963,
|
| 885 |
+
"grad_norm": 2.088919162750244,
|
| 886 |
+
"learning_rate": 1.4765946502057615e-05,
|
| 887 |
+
"loss": 0.2788,
|
| 888 |
+
"step": 2900
|
| 889 |
+
},
|
| 890 |
+
{
|
| 891 |
+
"epoch": 6.770833333333333,
|
| 892 |
+
"grad_norm": 83.2401123046875,
|
| 893 |
+
"learning_rate": 1.470164609053498e-05,
|
| 894 |
+
"loss": 0.2982,
|
| 895 |
+
"step": 2925
|
| 896 |
+
},
|
| 897 |
+
{
|
| 898 |
+
"epoch": 6.828703703703704,
|
| 899 |
+
"grad_norm": 0.2836779057979584,
|
| 900 |
+
"learning_rate": 1.4637345679012347e-05,
|
| 901 |
+
"loss": 0.3611,
|
| 902 |
+
"step": 2950
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 6.886574074074074,
|
| 906 |
+
"grad_norm": 0.5471726655960083,
|
| 907 |
+
"learning_rate": 1.4573045267489714e-05,
|
| 908 |
+
"loss": 0.3308,
|
| 909 |
+
"step": 2975
|
| 910 |
+
},
|
| 911 |
+
{
|
| 912 |
+
"epoch": 6.944444444444445,
|
| 913 |
+
"grad_norm": 78.7370834350586,
|
| 914 |
+
"learning_rate": 1.4508744855967079e-05,
|
| 915 |
+
"loss": 0.3321,
|
| 916 |
+
"step": 3000
|
| 917 |
+
},
|
| 918 |
+
{
|
| 919 |
+
"epoch": 7.0,
|
| 920 |
+
"eval_accuracy": 0.9498697916666666,
|
| 921 |
+
"eval_loss": 0.238317608833313,
|
| 922 |
+
"eval_macro_f1": 0.8321721266873331,
|
| 923 |
+
"eval_runtime": 2.6412,
|
| 924 |
+
"eval_samples_per_second": 581.544,
|
| 925 |
+
"eval_steps_per_second": 36.346,
|
| 926 |
+
"eval_weighted_f1": 0.9518076387137152,
|
| 927 |
+
"step": 3024
|
| 928 |
+
},
|
| 929 |
+
{
|
| 930 |
+
"epoch": 7.002314814814815,
|
| 931 |
+
"grad_norm": 0.22757171094417572,
|
| 932 |
+
"learning_rate": 1.4444444444444446e-05,
|
| 933 |
+
"loss": 0.3372,
|
| 934 |
+
"step": 3025
|
| 935 |
+
},
|
| 936 |
+
{
|
| 937 |
+
"epoch": 7.060185185185185,
|
| 938 |
+
"grad_norm": 0.19115474820137024,
|
| 939 |
+
"learning_rate": 1.4380144032921813e-05,
|
| 940 |
+
"loss": 0.2498,
|
| 941 |
+
"step": 3050
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 7.118055555555555,
|
| 945 |
+
"grad_norm": 1.7725087404251099,
|
| 946 |
+
"learning_rate": 1.431584362139918e-05,
|
| 947 |
+
"loss": 0.2753,
|
| 948 |
+
"step": 3075
|
| 949 |
+
},
|
| 950 |
+
{
|
| 951 |
+
"epoch": 7.175925925925926,
|
| 952 |
+
"grad_norm": 0.20514951646327972,
|
| 953 |
+
"learning_rate": 1.4251543209876545e-05,
|
| 954 |
+
"loss": 0.3777,
|
| 955 |
+
"step": 3100
|
| 956 |
+
},
|
| 957 |
+
{
|
| 958 |
+
"epoch": 7.233796296296297,
|
| 959 |
+
"grad_norm": 29.983375549316406,
|
| 960 |
+
"learning_rate": 1.4187242798353911e-05,
|
| 961 |
+
"loss": 0.2641,
|
| 962 |
+
"step": 3125
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 7.291666666666667,
|
| 966 |
+
"grad_norm": 0.38168174028396606,
|
| 967 |
+
"learning_rate": 1.4122942386831278e-05,
|
| 968 |
+
"loss": 0.3696,
|
| 969 |
+
"step": 3150
|
| 970 |
+
},
|
| 971 |
+
{
|
| 972 |
+
"epoch": 7.349537037037037,
|
| 973 |
+
"grad_norm": 151.57350158691406,
|
| 974 |
+
"learning_rate": 1.4058641975308642e-05,
|
| 975 |
+
"loss": 0.2191,
|
| 976 |
+
"step": 3175
|
| 977 |
+
},
|
| 978 |
+
{
|
| 979 |
+
"epoch": 7.407407407407407,
|
| 980 |
+
"grad_norm": 0.11410918086767197,
|
| 981 |
+
"learning_rate": 1.3994341563786009e-05,
|
| 982 |
+
"loss": 0.1134,
|
| 983 |
+
"step": 3200
|
| 984 |
+
},
|
| 985 |
+
{
|
| 986 |
+
"epoch": 7.465277777777778,
|
| 987 |
+
"grad_norm": 0.13736143708229065,
|
| 988 |
+
"learning_rate": 1.3930041152263375e-05,
|
| 989 |
+
"loss": 0.256,
|
| 990 |
+
"step": 3225
|
| 991 |
+
},
|
| 992 |
+
{
|
| 993 |
+
"epoch": 7.523148148148148,
|
| 994 |
+
"grad_norm": 0.16516904532909393,
|
| 995 |
+
"learning_rate": 1.3865740740740742e-05,
|
| 996 |
+
"loss": 0.3258,
|
| 997 |
+
"step": 3250
|
| 998 |
+
},
|
| 999 |
+
{
|
| 1000 |
+
"epoch": 7.581018518518518,
|
| 1001 |
+
"grad_norm": 0.17313998937606812,
|
| 1002 |
+
"learning_rate": 1.3801440329218107e-05,
|
| 1003 |
+
"loss": 0.2503,
|
| 1004 |
+
"step": 3275
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 7.638888888888889,
|
| 1008 |
+
"grad_norm": 0.13650918006896973,
|
| 1009 |
+
"learning_rate": 1.3737139917695474e-05,
|
| 1010 |
+
"loss": 0.4083,
|
| 1011 |
+
"step": 3300
|
| 1012 |
+
},
|
| 1013 |
+
{
|
| 1014 |
+
"epoch": 7.6967592592592595,
|
| 1015 |
+
"grad_norm": 0.9357269406318665,
|
| 1016 |
+
"learning_rate": 1.3672839506172841e-05,
|
| 1017 |
+
"loss": 0.394,
|
| 1018 |
+
"step": 3325
|
| 1019 |
+
},
|
| 1020 |
+
{
|
| 1021 |
+
"epoch": 7.75462962962963,
|
| 1022 |
+
"grad_norm": 0.12254858016967773,
|
| 1023 |
+
"learning_rate": 1.3608539094650206e-05,
|
| 1024 |
+
"loss": 0.3048,
|
| 1025 |
+
"step": 3350
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 7.8125,
|
| 1029 |
+
"grad_norm": 46.00386428833008,
|
| 1030 |
+
"learning_rate": 1.3544238683127573e-05,
|
| 1031 |
+
"loss": 0.2064,
|
| 1032 |
+
"step": 3375
|
| 1033 |
+
},
|
| 1034 |
+
{
|
| 1035 |
+
"epoch": 7.87037037037037,
|
| 1036 |
+
"grad_norm": 0.3903789222240448,
|
| 1037 |
+
"learning_rate": 1.347993827160494e-05,
|
| 1038 |
+
"loss": 0.3405,
|
| 1039 |
+
"step": 3400
|
| 1040 |
+
},
|
| 1041 |
+
{
|
| 1042 |
+
"epoch": 7.9282407407407405,
|
| 1043 |
+
"grad_norm": 0.11856985837221146,
|
| 1044 |
+
"learning_rate": 1.3415637860082307e-05,
|
| 1045 |
+
"loss": 0.2315,
|
| 1046 |
+
"step": 3425
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 7.986111111111111,
|
| 1050 |
+
"grad_norm": 0.1913604438304901,
|
| 1051 |
+
"learning_rate": 1.3351337448559672e-05,
|
| 1052 |
+
"loss": 0.2185,
|
| 1053 |
+
"step": 3450
|
| 1054 |
+
},
|
| 1055 |
+
{
|
| 1056 |
+
"epoch": 8.0,
|
| 1057 |
+
"eval_accuracy": 0.953125,
|
| 1058 |
+
"eval_loss": 0.23835718631744385,
|
| 1059 |
+
"eval_macro_f1": 0.8462214643630374,
|
| 1060 |
+
"eval_runtime": 2.5588,
|
| 1061 |
+
"eval_samples_per_second": 600.29,
|
| 1062 |
+
"eval_steps_per_second": 37.518,
|
| 1063 |
+
"eval_weighted_f1": 0.9532382823255752,
|
| 1064 |
+
"step": 3456
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 8.043981481481481,
|
| 1068 |
+
"grad_norm": 28.561185836791992,
|
| 1069 |
+
"learning_rate": 1.3287037037037039e-05,
|
| 1070 |
+
"loss": 0.1885,
|
| 1071 |
+
"step": 3475
|
| 1072 |
+
},
|
| 1073 |
+
{
|
| 1074 |
+
"epoch": 8.101851851851851,
|
| 1075 |
+
"grad_norm": 0.10930147767066956,
|
| 1076 |
+
"learning_rate": 1.3222736625514405e-05,
|
| 1077 |
+
"loss": 0.2552,
|
| 1078 |
+
"step": 3500
|
| 1079 |
+
},
|
| 1080 |
+
{
|
| 1081 |
+
"epoch": 8.159722222222221,
|
| 1082 |
+
"grad_norm": 0.1793520152568817,
|
| 1083 |
+
"learning_rate": 1.315843621399177e-05,
|
| 1084 |
+
"loss": 0.1973,
|
| 1085 |
+
"step": 3525
|
| 1086 |
+
},
|
| 1087 |
+
{
|
| 1088 |
+
"epoch": 8.217592592592593,
|
| 1089 |
+
"grad_norm": 0.14172351360321045,
|
| 1090 |
+
"learning_rate": 1.3094135802469137e-05,
|
| 1091 |
+
"loss": 0.2239,
|
| 1092 |
+
"step": 3550
|
| 1093 |
+
},
|
| 1094 |
+
{
|
| 1095 |
+
"epoch": 8.275462962962964,
|
| 1096 |
+
"grad_norm": 0.7105873227119446,
|
| 1097 |
+
"learning_rate": 1.3029835390946504e-05,
|
| 1098 |
+
"loss": 0.2119,
|
| 1099 |
+
"step": 3575
|
| 1100 |
+
},
|
| 1101 |
+
{
|
| 1102 |
+
"epoch": 8.333333333333334,
|
| 1103 |
+
"grad_norm": 0.12962496280670166,
|
| 1104 |
+
"learning_rate": 1.2965534979423868e-05,
|
| 1105 |
+
"loss": 0.2278,
|
| 1106 |
+
"step": 3600
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 8.391203703703704,
|
| 1110 |
+
"grad_norm": 5.1139607429504395,
|
| 1111 |
+
"learning_rate": 1.2901234567901235e-05,
|
| 1112 |
+
"loss": 0.2189,
|
| 1113 |
+
"step": 3625
|
| 1114 |
+
},
|
| 1115 |
+
{
|
| 1116 |
+
"epoch": 8.449074074074074,
|
| 1117 |
+
"grad_norm": 0.09980235248804092,
|
| 1118 |
+
"learning_rate": 1.2836934156378601e-05,
|
| 1119 |
+
"loss": 0.2259,
|
| 1120 |
+
"step": 3650
|
| 1121 |
+
},
|
| 1122 |
+
{
|
| 1123 |
+
"epoch": 8.506944444444445,
|
| 1124 |
+
"grad_norm": 5.740118980407715,
|
| 1125 |
+
"learning_rate": 1.2772633744855968e-05,
|
| 1126 |
+
"loss": 0.2289,
|
| 1127 |
+
"step": 3675
|
| 1128 |
+
},
|
| 1129 |
+
{
|
| 1130 |
+
"epoch": 8.564814814814815,
|
| 1131 |
+
"grad_norm": 16.739770889282227,
|
| 1132 |
+
"learning_rate": 1.2708333333333333e-05,
|
| 1133 |
+
"loss": 0.2798,
|
| 1134 |
+
"step": 3700
|
| 1135 |
+
},
|
| 1136 |
+
{
|
| 1137 |
+
"epoch": 8.622685185185185,
|
| 1138 |
+
"grad_norm": 96.03089904785156,
|
| 1139 |
+
"learning_rate": 1.26440329218107e-05,
|
| 1140 |
+
"loss": 0.2687,
|
| 1141 |
+
"step": 3725
|
| 1142 |
+
},
|
| 1143 |
+
{
|
| 1144 |
+
"epoch": 8.680555555555555,
|
| 1145 |
+
"grad_norm": 2.6519460678100586,
|
| 1146 |
+
"learning_rate": 1.2579732510288067e-05,
|
| 1147 |
+
"loss": 0.345,
|
| 1148 |
+
"step": 3750
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 8.738425925925926,
|
| 1152 |
+
"grad_norm": 0.2936369478702545,
|
| 1153 |
+
"learning_rate": 1.2515432098765432e-05,
|
| 1154 |
+
"loss": 0.2355,
|
| 1155 |
+
"step": 3775
|
| 1156 |
+
},
|
| 1157 |
+
{
|
| 1158 |
+
"epoch": 8.796296296296296,
|
| 1159 |
+
"grad_norm": 0.13087953627109528,
|
| 1160 |
+
"learning_rate": 1.2451131687242799e-05,
|
| 1161 |
+
"loss": 0.2985,
|
| 1162 |
+
"step": 3800
|
| 1163 |
+
},
|
| 1164 |
+
{
|
| 1165 |
+
"epoch": 8.854166666666666,
|
| 1166 |
+
"grad_norm": 3.2063426971435547,
|
| 1167 |
+
"learning_rate": 1.2386831275720166e-05,
|
| 1168 |
+
"loss": 0.2435,
|
| 1169 |
+
"step": 3825
|
| 1170 |
+
},
|
| 1171 |
+
{
|
| 1172 |
+
"epoch": 8.912037037037036,
|
| 1173 |
+
"grad_norm": 67.11962127685547,
|
| 1174 |
+
"learning_rate": 1.2322530864197533e-05,
|
| 1175 |
+
"loss": 0.2028,
|
| 1176 |
+
"step": 3850
|
| 1177 |
+
},
|
| 1178 |
+
{
|
| 1179 |
+
"epoch": 8.969907407407408,
|
| 1180 |
+
"grad_norm": 147.0647430419922,
|
| 1181 |
+
"learning_rate": 1.2258230452674898e-05,
|
| 1182 |
+
"loss": 0.378,
|
| 1183 |
+
"step": 3875
|
| 1184 |
+
},
|
| 1185 |
+
{
|
| 1186 |
+
"epoch": 9.0,
|
| 1187 |
+
"eval_accuracy": 0.9576822916666666,
|
| 1188 |
+
"eval_loss": 0.2343606948852539,
|
| 1189 |
+
"eval_macro_f1": 0.848293850150288,
|
| 1190 |
+
"eval_runtime": 2.5408,
|
| 1191 |
+
"eval_samples_per_second": 604.545,
|
| 1192 |
+
"eval_steps_per_second": 37.784,
|
| 1193 |
+
"eval_weighted_f1": 0.9582705979114049,
|
| 1194 |
+
"step": 3888
|
| 1195 |
+
}
|
| 1196 |
+
],
|
| 1197 |
+
"logging_steps": 25,
|
| 1198 |
+
"max_steps": 8640,
|
| 1199 |
+
"num_input_tokens_seen": 0,
|
| 1200 |
+
"num_train_epochs": 20,
|
| 1201 |
+
"save_steps": 500,
|
| 1202 |
+
"stateful_callbacks": {
|
| 1203 |
+
"EarlyStoppingCallback": {
|
| 1204 |
+
"args": {
|
| 1205 |
+
"early_stopping_patience": 3,
|
| 1206 |
+
"early_stopping_threshold": 0.0
|
| 1207 |
+
},
|
| 1208 |
+
"attributes": {
|
| 1209 |
+
"early_stopping_patience_counter": 3
|
| 1210 |
+
}
|
| 1211 |
+
},
|
| 1212 |
+
"TrainerControl": {
|
| 1213 |
+
"args": {
|
| 1214 |
+
"should_epoch_stop": false,
|
| 1215 |
+
"should_evaluate": false,
|
| 1216 |
+
"should_log": false,
|
| 1217 |
+
"should_save": true,
|
| 1218 |
+
"should_training_stop": true
|
| 1219 |
+
},
|
| 1220 |
+
"attributes": {}
|
| 1221 |
+
}
|
| 1222 |
+
},
|
| 1223 |
+
"total_flos": 0.0,
|
| 1224 |
+
"train_batch_size": 8,
|
| 1225 |
+
"trial_name": null,
|
| 1226 |
+
"trial_params": null
|
| 1227 |
+
}
|
ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-3888/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3c3407bdf0f213cd25e377b83459c541b3e7805ba365d4b8290866b369c921e2
|
| 3 |
+
size 5624
|
ood-uit_vsfc-phobert-topic-domain-02-s44-20260812T035847Z-6beb231d/trainer/checkpoint-3888/vocab.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-uit_vsfc-phobert-topic-domain-03-s44-20260812T044906Z-6beb231d/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:77e19bc28f749c7dd54bc814b334657a7900384fd8cdbb29cc74a7087a8d1011
|
| 3 |
+
size 540026404
|
ood-uit_vsfc-phobert-topic-domain-03-s44-20260812T044906Z-6beb231d/trainer/checkpoint-1850/added_tokens.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<mask>": 64000
|
| 3 |
+
}
|
ood-uit_vsfc-phobert-topic-domain-03-s44-20260812T044906Z-6beb231d/trainer/checkpoint-1850/bpe.codes
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
ood-uit_vsfc-phobert-topic-domain-03-s44-20260812T044906Z-6beb231d/trainer/checkpoint-1850/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f3c0c6d39863584906ad49f744e6bded394063b445051506cd548ca3343beaeb
|
| 3 |
+
size 540026436
|